Conjugacy classes and finite p-groups(2005)--Adan-Bante
数学符号公式缩写的英文发音
数学符号公式缩写的英文发音数学缩写列表this article is a listing of abbreviated names of mathematical functions, function-like operators and other mathematical terminology.这篇文章是一个数学函数,类似于函数的操作符和其他的数学术语的缩写名列表。
this list is limited to abbreviations of two or more letters. the capitalization of some of these abbreviations is not standardized – different authors use different capitalizations.this list is inplete; you can help by expanding it.这个列表受限于两个或更多字母的缩略语。
其中,一些缩略语字母大写并不是标准的 - 不同的使用不同的大写形式。
ac – axiom of choice.[1] 选择公理a.c. – absolutely continuous. 绝对连续的acrd – inverse chord function. 逆弦函数adj – adjugate of a matrix. 矩阵的伴随矩阵a.e. – almost everywhere. 殆遍,几乎处处ai – airy function. 艾里函数al – action limit. 处置界限alt – alternating group (alt(n) is also written as an.) 交错群a.m. – arithmetic mean. 算数平均数arccos – inverse cosine function. 反余弦函数arccosec – inverse cosecant function. (also written as arccsc.) 反余割函数arccot – inverse cotangent function. 反余切函数arccsc – inverse cosecant function. (also written as arccosec.) 反余割函数arcexc – inverse excosecant function. (also written as arcexcsc, arcexcosec.) 反外余割函数arcexcosec – inverse excosecant function. (also written as arcexcsc, arcexc.) 反外余割函数arcexcsc – inverse excosecant function. (also written as arcexcosec, arcexc.) 反外余割函数arcexs – inverse exsecant function. (also written as arcexsec.) 反外正割函数arcexsec – inverse exsecant function. (also written as arcexs.) 反外正割函数arcosech – inverse hyperbolic cosecant function. (also written as arcsch.) 反双曲余割函数arcosh – inverse hyperbolic cosine function. 反双曲余弦函数arcoth – inverse hyperbolic cotangent function. 反双曲余切函数arcsch – inverse hyperbolic cosecant function. (also written as arcosech.) 反双曲余割函数arcsec – inverse secant function. 反正割函数arcsin – inverse sine function. 反正弦函数arctan – inverse tangent function. 反正切函数arctan2 – inverse tangent function with two arguments. (also written as atan2.) 带有2个参数的反正切函数arg – argument of a plex number.[2] 复数的参数arg max – argument of the maximum. 最大值时的参数arg min – argument of the minimum. 最小值时的参数arsech – inverse hyperbolic secant function. 反双曲正割函数arsinh – inverse hyperbolic sine function. 反双曲正弦函数artanh – inverse hyperbolic tangent function. 反双曲正切函数a.s. – almost surely. 殆必,几乎必然atan2 – inverse tangent function with two arguments. (also written as arctan2.) 同 arctan2,带有两个参数的反正切函数a.p. – arithmetic progression. 等差数列aut – automorphism group. 自同构群bd – boundary. 边界(拓扑学)bi – airy function of the second kind. 第二类艾里函数bias – bias of an estimator 估计器偏置card – cardinality of a set.[3] (card(x) is also written #x, ♯x or |x|.) 集合的势cdf – cumulative distribution function. 累积分布函数c.f. – cumulative frequency. 累积频率char – characteristic of a ring. 环的特征chi – hyperbolic cosine integral function. 双曲余弦积分函数ci – cosine integral function. 余弦积分函数cis – cos + i sin function. 欧拉公式函数cl – conjugacy class. 共轭类cl – topological closure. 拓扑学闭包cod, codom – codomain. 到达域cok, coker – cokernel. 上核,余核cor – corollary. 推论,余定理corr – correlation. 相关cos – cosine function. 余弦函数割函数cosech – hyperbolic cosecant function. (also written as csch.) 双曲余割函数cosh – hyperbolic cosine function. 双曲余弦函数cosiv – coversine function. (also written as cover, covers, cvs.) 余矢函数cot – cotangent function. (also written as ctg.) 余切函数coth – hyperbolic cotangent function. 双曲余切函数cov – covariance of a pair of random variables. 协方差cover – coversine function. (also written as covers, cvs, cosiv.) 余矢函数covercos – covercosine function. (also written as cvc.) 正矢函数covers – coversine function. (also written as cover, cvs, cosiv.) 余矢函数crd – chord function. 弦(几何)函数csc – cosecant function. (also written as cosec.) 余割函数csch – hyperbolic cosecant function. (also written as cosech.) 双曲余割函数函数curl – curl of a vector field. (also written as rot.) 向量场的旋度cvc – covercosine function. (also written as covercos.) 余余矢函数cvs – coversine function. (also written as cover, covers, cosiv.) 正余矢函数def – define or definition. 定义deg – degree of a polynomial. (also written as ∂.)多项式的次数del – del, a differential operator. (also writtenas .) 微分运算符det – determinant of a matrix or linear transformation. 矩阵或线性变换的行列式dim – dimension of a vector space. 向量空间的维度div – divergence of a vector field. 向量场的散度dkl – decalitre 公斗。
Discovering Accurate and Interesting Classification Rules Using Genetic Algorithm
Discovering Accurate and Interesting Classification Rules Using GeneticAlgorithmJanaki Gopalan Reda Alhajj Ken BarkerAbstractDiscovering accurate and interesting classification rulesis a significant task in the post-processing stage of a datamining(DM)process.Therefore,an optimization problemexists between the accuracy and the interesting metrics forpost-processing rule sets.To achieve a balance,in thispaper,we propose two major post-processing tasks.Inthefirst task,we use a genetic algorithm(GA)tofind thebest combination of rules that maximizes the predictiveaccuracy on the sample training set.Thus we obtain themaximized accuracy.In the second task,we rank the rulesby assigning objective rule interestingness(RI)measures(or weights)for the rules in the rule set.Henceforth,we propose a pruning strategy using a GA tofind thebest combination of interesting rules with the maximized(or greater)accuracy.We tested our implementation onthree data sets.The results are very encouraging;theydemonstrate the applicability and effectiveness of ourapproach.Keywords:post-processing,data mining,classificationrules,rule interestingness,genetic algorithms.1IntroductionData mining is generally defined as the process ofextracting previously unknown knowledge from a givendatabase.A DM process is divided into three stagesnamely,the pre-processing,mining,and the post-processingstages[1,20].The post-processing stage of the DM pro-cess involves interpretation of the discovered knowledge orsome post-processing of this knowledge.An example ofjective boosted hypothesis rule[4].Therefore,the accu-racy of the obtained results are biased by the accuracy with which these weights are obtained.Moreover,these weightsare based on one metric which is the classification accuracy of the classifier.In this paper,we propose a pruning strat-egy by extending the idea proposed by Thompson.In our strategy,we use a GA with objective rule interestingness measures(based on Freitas[7])tofind the most interest-ing subset with the performance accuracy of atleast on the sample set(problem space).These measures are based on several objective metrics(including the accuracy metric) to derive interesting as well as accurate rules.Therefore, the resulting rule set from the solution of our GA method is the best combination of accurate interesting classification rules.These rules are then tested for their accuracy on the unknown validation set(the solution space).The rest of the paper is organized as follows.Section2 describes the related work in rule-set refinement for classi-fication rules.Section3discusses the implementation using a GA for this problem.In Section4,we give the experimen-tal results using the GA method.Section5is conclusions and future work.2Related WorkIn this section,we discuss the related work in rule-set refinement for classification rules.They are:1)The rule interestingness(RI)principles proposed for classification rules;and2)finding the best set(or subset)of rules from the discovered rule-set.The task of assigning a RI measure is discussedfirst,followed by the discussion of deriving the best combination of accurate rules.Methods for the selection of interesting classification rules can be divided into subjective and objective methods. Subjective methods are user-driven and domain-dependent. By contrast,objective methods are data-driven and domain-independent.A comprehensive review about subjective as-pects of RI is available[10].Piatetsky-Shapiro[11],Major and Mangano[12],and Kamber and Shinghal[13]propose objective principles for RI to include the rule quality factors of coverage,completeness,and a confidence factor.Freitas (1999)[7]extended the objective RI principles[11,12,13] to include additional factors such as the disjunct size,im-balance of class distributions,attribute interestingness,mis-classification costs,and the asymmetric nature of classifica-tion rules.We consider the next problem of pruning rule sets.Pro-dromidis et al.[14]present methods for pruning classifiers in a distributed meta-learning system.A pre-training prun-ing is used to select a subset of classifiers from an ensem-ble which are then combined by a meta-learned combiner. Margineantu and Deitterich[2]use a backfitting algorithm for pruning classifier sets.This involves choosing an ad-ditional classifier to add to a set of classifiers by a greedy search and then checking that each of the other classifiers in the set cannot be replaced by another to produce a better en-semble.Thompson[4]proposes a GA to prune a classifier ensemble tofind the right combination of classifiers with-out over-fitting the training set.The proposed GA uses a real-valued encoding.Each chromosome has real-valued genes,where is the number of classifiers in the ensemble. Each gene represents the voting weight of its corresponding classifier calculated using a boosted hypothesis(WVBH) rule.Thefitness function consists of measuring the pre-dictive accuracy of the classifier ensemble with the weights proposed by the chromosomes on a hold-out set(different from the training set).Two major conclusions are drawn:1) In a majority of the experiments that were performed with the classifier sets,it was found that a subset of classifiers from the original classifier ensemble had a better classifica-tion accuracy.2)The GA method was very efficient infind-ing the right set of pruned classifiers.Moreover,the pruned classifier sets from the GA method have a better classifica-tion accuracy over the pruned classifiers sets from earlier work[4].In this paper,we propose a pruning strategy using a GA tofind the best set of interesting and accurate rules by ex-tending the idea proposed by Thompson.In the rest of the paper,wefirst discuss our proposed approach that employs GA to address this problem.Finally,we present the experi-mental validations along with future work.3Post-processing Rule Sets Using Genetic AlgorithmsGA’s were introduced by Holland[8],as a general model of an adaptive processes,but subsequently widely exploited as optimizers[8].Basically,a GA can be used for solving problems for which it is possible to construct an objective function(also known asfitness function)to estimate how a given representative(solution)fits the considered environ-ment(problem).In general,the main motivation for using GAs in any data mining process is that they perform a global search and cope better with interaction than the greedy rule induction algorithms often used in data mining.Genetic algorithms can be used in the post-processing stage of the DM process.Very little work has been reported in the literature in this area.As reviewed in earlier sections, Thompson[4]proposes a GA to prune a classifier ensemble efficiently.We implemented GAKPER,a GA based Knowledge Discovery algorithm for deriving Efficient Rules to achieve our goal.In our implementation,the original dataset is di-vided into a sample set(to train)and a validation set(to test).This task is divided into two parts.In thefirst part, a binary encoded GA is used tofind the most accurate sub-set of rules with the best classification accuracy on the sample set(problem space).In the second part,a binary en-coded GA is used tofind the most interesting subset with accuracy of at least on the sample set.Finally,the derived accurate interesting rules are tested on the unknown valida-tion set(solution space)for their accuracy.Each of these parts are described in the subsections below.3.1GAKPER ALGORITHM-PART1A binary encoded GA is used to search for the best com-bination of accurate rules.Each chromosome in the popu-lation is a subset of the classification rules.The length of the chromosome is the number of rules in the rule set;in that,each gene represents the corresponding classification rule.For example,if there are rules in the original rule set,then“1001100111”is a possible chromosome,where thefirst,fourth,fifth,eight,ninth,and tenth rules from the rule set are chosen to represent the set of accurate rules.A solution in the phenotype space is only represented by a single chromosome and all possible chromosomes are valid. Hence a1-1mapping exists between the genotype and phe-notype spaces.Thefitness functionfirst measures the predictive accu-racy of the rules(represented by the chromosome)on the entire sample set.This is achieved as follows.The true class values for all instances in the sample set is stored prior to running the GA.The classes predicted by the rule set repre-senting the chromosome are known.Therefore,to classify the test instances from the sample set,thefitness function takes a vote of the rules from the rule set.Thus,to calculate the class of a single test instance,the class predicted by the rules representing the chromosome(based on majority vote) is matched with the original stored class value of the test in-stance.This process is repeated for all the test instances in the sample set.Finally,the best combination of rules that maximizes the predictive accuracy on the sample set(i.e., the number of correctly predicted test instances)is obtained. The maximum accuracy is denoted as.If represents the number of correctly predicted instances by the chromo-some,thefitness value of the chromosome is defined as:(1) All these aspects are precisely encoded and implemented into the GA and all the chromosomes(potential solutions) should be awarded or punished according to the criteria stated above during the process of evolution.The outcome of several evolutions modeled by this GA generates the right set of accurate rules.The GAKPER algorithm(Part I)is presented below: ALGORITHM:GAKPER-IInput:Classification rule set Output:Set of accurate rulesMethod:1)Search for the right set of accuraterules using the GA Method as follows: 1a)Randomly create an initial populationof potential accurate rules.1b)Iteratively perform the following substeps on the population until thetermination criterion in the GA issatisfied:a)FITNESS:Evaluate fitness f(x)of eachchromosome in the populationb)NEW POPULATIONb1)SELECTION:Based on f(x)b2)RECOMBINATION:2-point Cross-over ofchromosomesb3)MUTATION:Mutate chromosomesb4)ACCEPTATION:Reject or accept new onec)REPLACE:Replace old with newpopulation in new generationd)TEST:Test problem criterium(Number of generations is>100)2)After the termination criterion issatisfied,the best chromosome in thepopulation produced during the run isdesignated as the required combinationof the accurate rules.3.2GAKPER ALGORITHM-PART IIA binary encoded GA is used to search for the best com-bination of interesting accurate rules.The representation for the chromosome is the same as described in Part1.Each chromosome in the population is a subset of the classifica-tion rules.The length of the chromosome is the number of rules in the rule set,in that,each gene represents the corre-sponding classification rule.The maximum predictive accuracy is known from the result of the GAKPER algorithm-Part I.The RI measure (or weight)proposed by Freitas is assigned to all the rules. Thefitness function optimizes the weights tofind the best combination of interesting rules with a classification accu-racy of at least.It is important to note that accuracy of the rules when tested on the test instances(from the sample set) are based on the weighted majority vote.Thus,a rule with a higher RI measure ranks higher in classification over a rule with a lower RI measure.The weights of the rules rep-resenting a chromosome are denoted as. If the accuracy of the rules is greater than(or equal)to, then thefitness value of the chromosome is the sum of the weights,otherwise the chromosome is given a defaultfit-ness value.The intuitive idea of choosing this default value is:1)the value has to be greater than zero to continue the GA runs in subsequent generations,and2)since thefitness criteria is not satisfied by the respective chromosome,an ar-bitrary value is chosen to punish the leastfit chromosome. Thefitness value of the chromosome is defined as:If(2)All these aspects are precisely encoded and implemented into the GA and all the chromosomes(potential solutions) should be awarded or punished according to the criteria stated above during the process of evolution.The outcome of several evolutions modeled by this GA generates the right set of accurate interesting rules.The GAKPER algorithm(Part II)is presented below:ALGORITHM:GAKPER-IIInput:Classification rule setOutput:Set of accurate interesting rules Method: (1)Assign weights for the rules in the input ruleset.(2)Search for the right set of accurateinteresting rules using the GA Method asfollows:2a)Randomly create an initial population of potential accurate interesting rules.2b)Iteratively perform the following sub steps on the population until the terminationcriterion in the GA is satisfied:a)FITNESS:Evaluate fitness f(x)of eachchromosome in the populationb)NEW POPULATIONb1)SELECTION:Based on f(x)b2)RECOMBINATION:2-point Cross-over ofchromosomesb3)MUTATION:Mutate chromosomesb4)ACCEPTATION:Reject or accept new onec)REPLACE:Replace old with new population:the new generationd)TEST:Test problem criterium(Number of generations is>100) (3)After the termination criterion is satisfied,the best chromosome in the population produced during the run is designated as the rightcombination of the accurate interesting rules. The approach is implemented as a standard GA writ-ten in C,similar to Grefenstette’s GENESIS program; Baker’s SUS selection algorithm[21]is employed;2-point crossover is maintained at60%and mutation is very low; and selection is based on proportionalfitness.It is impor-tant to note that this approach optimizes the predictive ac-curacy and the interesting measures of the rule set on the entire sample set,which is the problem space.4Results and DiscussionsAll the tests have been conducted using a single Proces-sor,Intel(R)Xeon(TM)UNIX machine with CPU power of 2.80GHz and cache size of512KB.The GAKPER algorithm implementation is tested onfive datasets.The data-splitting,that is,dividing the dataset into the sample set and validation set is performed using ran-dom sampling technique.The classification rules are ob-tained using the sample set.Recall that,to prune the rule set to derive the set of interesting accurate rules,the fol-lowing tests are performed.The GAKPER(part I)is used to derive the subset of rules that maximizes the classifica-tion accuracy on the sample set.To derive the most inter-esting subset,a RI measure or weight(based on Freitas)is assigned to the rules.GAKPER(part II)is used again,this time,to maximize the interestingness measure of the rules whose classification accuracy on the sample set is at least .The accuracy of the derived interesting accurate rules (on the validation set)from this approach is compared with: 1)pruning the rule set using GA without assigning initial weights for the rules,and2)without any pruning methods, using the entire rule set.The results for thefive datasets are presented below.Data Set1:Breast Cancer Data SetThis is a real data set obtained from Tom Baker Cancer Cen-ter,Calgary,Alberta,Canada.The original dataset consists of records and attributes.Each record represents follow-up data for one breast cancer case.Breast cancer “recurred”in some of these patients after the initial occur-rence.Hence,each patient is classified as“recurrent”or “non-recurrent”,depending on his or her status.With re-spect to classification of the dataset,there are2classes: 1)Recurrent Patients,and2)Non Recurrent Patients.The original dataset is divided into a sample set and a validation set using the random sampling technique. The sample set has instances whereas the validation set has instances.The GA based approaches,i.e., GAKPER-I and GAKPER-II,is used(on the sample set) tofind the best combination of accurate interesting rules. These rules are tested on the unknown validation set for ac-curacy.The GA parameters are:,-,,,and the is.30 experiments were performed with the GA approaches and every GA experiment was run for generations in both approaches.The post-processing results on the sample set and the validation sets are presented in Table1and Table2 respectively.Table1.Post-processing Results using Dif-ferent Approaches on the Sample Set for Dataset1WithoutWeightsCorrectly Predicted(in%)80Incorrectly Predicted(in%)10Unknown(in%)10Table2.Post-processing Results using Dif-ferent Approaches on the Validation Set for Dataset1WithoutWeightsCorrectly Predicted(in%)77Incorrectly Predicted(in%)8Unknown(in%)15With Weights Entire Rule Set94940066Table4.Post-processing Results using Dif-ferent Approaches on the Validation Set for Dataset2WithoutWeightsCorrectly Predicted(in%)91Incorrectly Predicted(in%)3Unknown(in%)6using the different approaches on the sample set(the prob-lem space)for Dataset2is presented.Table4presents the accuracy results of the discovered rule sets(using the differ-ent approaches)when tested on the validation set(the solu-tion space)for Dataset2.For this dataset,we found that the accuracy of the result on the unknown validation set using different approaches is the same(as presented in thefirst, second and third columns in Table4).This is also depicted graphically in Fig2.Data Set3:US-CENSUS-DATASETThis data is the USCensus1990raw data set obtained from the UCI repository[19].The data was collected as part of the1990census.The original dataset consisted of cate-gorical attributes and instances.The dataset used here consists of instances and attributes derived from the original USCensus1990rawdataset.It contains classes namely,iClass=0,iClass=5,and iClass=1.Each class refers to the native country of the candidate under con-sideration.The original dataset is divided into a sample set and a validation set using the random sampling technique.The sample set has instances while the validation set has instances.The GA based approaches,i.e.,GAKPER-Iand GAKPER-II,is used(on the sample set)tofind the best combination of accurate interesting rules.These rules are tested on the unknown validation set for accuracy.Finally, the same GA parameters enumerated earlier for Dataset1 is used here.The post-processing results are presented in Table5and Table6.In Table5,the accuracy of the rule sets,while pruning, using the different approaches on the sample set(the prob-lem space)for Dataset3is presented.Table6presents the accuracy results of the discovered rule sets(using the dif-ferent approaches)when tested on the validation set(the solution space)for Dataset3.It is very important to ob-serve that,for this dataset,the accuracy of the result on the unknown validation set using our approach(as presented in thefirst column in Table6)is much greater than the accu-racy of the results using the traditional approaches(as pre-sented in second and third columns in Table6).This is also depicted graphically in Fig3.Table5.Post-processing Results using Differ-ent Approaches on the Sample the US census dataWithoutWeightsCorrectly Predicted(in%)66Incorrectly Predicted(in%)34Unknown(in%)0WithWeightsEntireRule Set7560254000From the results,it can be observed that:1)in majority of the tests performed,a subset of rules pruned from the original set has a better performance accuracy;and2)it is possible to derive the most interesting subset with a higher classification accuracy as compared to the original rule set. Therefore,the GA with its inherent robust search strategies is most suitable for the post-processing problem considered in this paper.5Conclusions and Future WorkIn this paper,we propose and implement a GA based methodology to derive interesting and accurate classifica-tion rules from a dataset.The fundamental goal of any data mining model is to derive interesting rules.At the same time,accuracy is a key issue.Therefore,in the post-processing component,the problem of deriving interesting accurate rules is addressed. The earlier works in this area addressed the following two problems independently:1)the problem offinding the ac-curate set(or subset)of rules using pruning strategies to theoriginal rule set[4];and2)the problem of assigning sub-jective or objective RI measures to the rules to determine their interestingness[7,11,12,13].In our work,a new methodology is proposed byfirst assigning an objective RI measure based on Freitas[7]to the rules and using pruning strategies with a GA to search for the right set of interesting accurate rules.An alternative approach worth investigating is tofind the interesting accurate rules using multi-objective genetic al-gorithms.The goal is to optimize two parameters,namely, the interesting and the accuracy metrics for the classifica-tion rules simultaneously.References[1]Dorain Pyle,“Data Preparation For Data Mining”,Morgan Kaufmann,1999.[2]D.D.Margineantu and T.G.Dietterich,“PruningAdaptive Boosting”.Proceedings of the14Inter-national Conference on Machine Learning,San Fran-cisco,CA,pp.211-218,1997.[3]J.R.Quinlan,“Boosting First Order Learning”.Pro-ceedings of the14International Conference on Ma-chine Learning,1997.[4]S.Thompson,“Genetic Algorithms as Postprocessorsfor Data Mining”.Data Mining with Evolutionary Algorithms:Research Directions-Papers from the AAAI Workshop”,pp18-22,1999.[5]R.E.Schapire and Y.Freund and P.Bartlett and W.S.Lee,“Boosting the Margin:a New Explanation for the Effectiveness of V oting Methods”.Machine Learning:Proceedings of the14International Con-ference,pp.322-330,1997.[6]P.Domingos,“Knowledge Acquisition from Exam-ples via Multiple Models”.Machine Learning:Pro-ceedings of the14International Conference,pp98-106,1997.[7]A.A.Freitas,“On Rule Interestingness Measures”.Advances in Evolutionary Computation,Knowledge-Based Systems,12,1999.[8]D.E.Goldberg,“Genetic Algorithms in Search,Op-timization and Machine Learning”.Addison Wesley, Longman Publishing Co.,Inc.,Boston,MA,1989. [9]A.A.Freitas,“A Survey of Evolutionary Algo-rithms for Data Mining and Knowledge Discovery”.Advances in Evolutionary Computation,Springer-Verlag,2001.[10]B.Liu,W.Hsu,and Yiming Ma,“Integrating Classifi-cation and Association Rule Mining”.Proc.of KDD, pp.80-86,1998.[11]G.Piatetsky-Shapiro,“Discovery,Analysis,and Pre-sentation of Strong Rules”.Knowledge Discovery in Databases,pp.229-248,1991.[12]J.A.Major and J.J.Mangano,“Selecting AmongRules Induced from a Hurricane Database”.Proceed-ings of AAI-93,Workshop on Knowledge Discovery in Databases,pp.28-44,1993.[13]M.Kamber and R.Shinghal,“Evaluating the Inter-estingness of Characteristic Rules”.Proceedings of the2International Conference on KDD”,pp.28-44, 1993.[14]A.L.Prodromidis and S.Stolfo,“Pruning Classifiersin a Distributed Meta-Learning System”.Proceedings of the KDD”,pp.151-160,1998.[15]C.M.Fonseca and P.J.Fleming,“Genetic Algo-rithms for Multi-Objective Optimization:Formula-tion,Discussion and Generalization”.Proceedings of the Fifth International Conference on Genetic Algo-rithms,pp.93-100,1985.[16]R.J.Bayardo,“Brute-Force Mining of High-Confidence Classification Rules”.Proc.of KDD, pp.123-126,1997.C.M.Fonseca and P.J.Fleming [17]Han,“CMAR:Accurate and Efficient ClassificationBased on Multiple Class-Association Rules”.Proc.of IEEE-ICDM,pp.369-376,2001.[18]R.Agrawal,and R.Srikant,“Fast algorithms for min-ing association rules,”Proc.of VLDB,Santiago,Chile, September1994.[19]C.L.Blake,and C.J.Merz,“UCI Repository of ma-chine learning databases”.University of California, Department of Information and Computer Science, Irvine,CA,1998.[20]I.H.Witten and E.Frank,“Data Mining:Practical Ma-chine Learning Tools and Techniques with Java Imple-mentations”.Morgan Kaufmann,October1999. [21]J.E.Baker,Reducing Bias&Inefficiency in the Selec-tion Algorithm.Proc.of the International Conference on Genetic Algorithms,pp.14-21,1987.。
classification
classificationClassification is a fundamental task in machine learning and data analysis. It involves categorizing data into predefined classes or categories based on their features or characteristics. The goal of classification is to build a model that can accurately predict the class of new, unseen instances.In this document, we will explore the concept of classification, different types of classification algorithms, and their applications in various domains. We will also discuss the process of building and evaluating a classification model.I. Introduction to ClassificationA. Definition and Importance of ClassificationClassification is the process of assigning predefined labels or classes to instances based on their relevant features. It plays a vital role in numerous fields, including finance, healthcare, marketing, and customer service. By classifying data, organizations can make informed decisions, automate processes, and enhance efficiency.B. Types of Classification Problems1. Binary Classification: In binary classification, instances are classified into one of two classes. For example, spam detection, fraud detection, and sentiment analysis are binary classification problems.2. Multi-class Classification: In multi-class classification, instances are classified into more than two classes. Examples of multi-class classification problems include document categorization, image recognition, and disease diagnosis.II. Classification AlgorithmsA. Decision TreesDecision trees are widely used for classification tasks. They provide a clear and interpretable way to classify instances by creating a tree-like model. Decision trees use a set of rules based on features to make decisions, leading down different branches until a leaf node (class label) is reached. Some popular decision tree algorithms include C4.5, CART, and Random Forest.B. Naive BayesNaive Bayes is a probabilistic classification algorithm based on Bayes' theorem. It assumes that the features are statistically independent of each other, despite the simplifying assumption, which often doesn't hold in the realworld. Naive Bayes is known for its simplicity and efficiency and works well in text classification and spam filtering.C. Support Vector MachinesSupport Vector Machines (SVMs) are powerful classification algorithms that find the optimal hyperplane in high-dimensional space to separate instances into different classes. SVMs are good at dealing with linear and non-linear classification problems. They have applications in image recognition, hand-written digit recognition, and text categorization.D. K-Nearest Neighbors (KNN)K-Nearest Neighbors is a simple yet effective classification algorithm. It classifies an instance based on its k nearest neighbors in the training set. KNN is a non-parametric algorithm, meaning it does not assume any specific distribution of the data. It has applications in recommendation systems and pattern recognition.E. Artificial Neural Networks (ANN)Artificial Neural Networks are inspired by the biological structure of the human brain. They consist of interconnected nodes (neurons) organized in layers. ANN algorithms, such asMultilayer Perceptron and Convolutional Neural Networks, have achieved remarkable success in various classification tasks, including image recognition, speech recognition, and natural language processing.III. Building a Classification ModelA. Data PreprocessingBefore implementing a classification algorithm, data preprocessing is necessary. This step involves cleaning the data, handling missing values, and encoding categorical variables. It may also include feature scaling and dimensionality reduction techniques like Principal Component Analysis (PCA).B. Training and TestingTo build a classification model, a labeled dataset is divided into a training set and a testing set. The training set is used to fit the model on the data, while the testing set is used to evaluate the performance of the model. Cross-validation techniques like k-fold cross-validation can be used to obtain more accurate estimates of the model's performance.C. Evaluation MetricsSeveral metrics can be used to evaluate the performance of a classification model. Accuracy, precision, recall, and F1-score are commonly used metrics. Additionally, ROC curves and AUC (Area Under Curve) can assess the model's performance across different probability thresholds.IV. Applications of ClassificationA. Spam DetectionClassification algorithms can be used to detect spam emails accurately. By training a model on a dataset of labeled spam and non-spam emails, it can learn to classify incoming emails as either spam or legitimate.B. Fraud DetectionClassification algorithms are essential in fraud detection systems. By analyzing features such as account activity, transaction patterns, and user behavior, a model can identify potentially fraudulent transactions or activities.C. Disease DiagnosisClassification algorithms can assist in disease diagnosis by analyzing patient data, including symptoms, medical history, and test results. By comparing the patient's data againsthistorical data, the model can predict the likelihood of a specific disease.D. Image RecognitionClassification algorithms, particularly deep learning algorithms like Convolutional Neural Networks (CNNs), have revolutionized image recognition tasks. They can accurately identify objects or scenes in images, enabling applications like facial recognition and autonomous driving.V. ConclusionClassification is a vital task in machine learning and data analysis. It enables us to categorize instances into different classes based on their features. By understanding different classification algorithms and their applications, organizations can make better decisions, automate processes, and gain valuable insights from their data.。
国外Advanced Cognitive Psychology课程课件-Concepts and Categories
People abstract common elements of a formed category and use a common representation to stand for that category How is the category updated? Family Resemblance
Doesn’t take into account category size or variability in examples Context
What may be more typical in one setting may not be elsewhere
Correlations among attributes
Similarity based retrieval
Since the exemplar approach retains more information about the category itself it gets around some of the problems faced by the prototype theory (e.g. context effects), but also how a prototype could be recognized at test when wasn’t presented previously
Overlap of common attributes Classification is made based on overlap between prototype and exemplar
Prototype
The prototype view can explain both typicality effects and the fact that prototypes that had not been previously presented are correctly classified (even more accurately) Problems with prototype explanation
非正规子群是Sylow子群的有限群
非正规子群是Sylow子群的有限群褚智伟【期刊名称】《《南通大学学报(自然科学版)》》【年(卷),期】2019(018)002【总页数】4页(P87-90)【关键词】有限群; 非正规子群; 共轭; Sylow子群【作者】褚智伟【作者单位】南通师范高等专科学校学前教育第一学院江苏南通 226006【正文语种】中文【中图分类】O152.1研究子群的正规性与有限群结构的关系是有限群的重要课题之一。
著名的Dedekind 群就是每个子群都正规的有限群。
这里我们讨论其对偶问题:非正规子群的性质对有限群结构的影响。
Sylow 子群是有限群中最重要的子群,它的正规性影响到群的幂零性,又是群性质和数量性质沟通的桥梁。
文献[1-2]研究了子群的性质对有限群结构的影响;文献[3]给出了恰有p 个相互共轭的非正规子群的有限群结构;文献[4]给出了非正规子群的共轭类类数为4的有限幂零群;文献[5-10]给出了恰有2,4,5,6,7 个非正规子群的有限群及有限NN-群的非正规子群的有限群结构。
本文将研究非正规子群为Sylow 子群的群结构,从数量性质出发探讨群结构的存在性。
1 预备知识和相关引理本文中涉及的群均为有限群。
Pr表示群G 的Sylow r-子群;nr表示群G 的Sylow r-子群的个数;表示群G 的非正规子群的共轭类类数;τ(G)表示群G 中非正规子群的个数;π(G)表示群G 的阶所含全体素因子的集合;表示与H 共轭的子群的个数。
首先,我们介绍几个有用的引理。
引理1[11]若G 为有限群,则如下命题等价:1)。
2)是一个非交换可裂扩张,其中N 为素数阶循环群,P 为素数幂阶循环群,,即,其中:p=2 时,n ≥3;p ≥3 时,n ≥2。
引理2[12]若G 为有限非幂零群,,P为G 的非正规的Sylow p-子群,则G 中除Sylow p-子群外,其余Sylow 子群都正规于G。
当P < NG(P)时,,其中是一非交换可裂扩张,。
时俭益男一九四八年三月三十一日生于浙江省
时俭益时俭益,男,一九四八年三月三十一日生于浙江省慈城镇(现名慈溪市),留英博士,华东师范大学数学系教授,博士生导师,终身教授;南开大学特聘讲座教授。
一九八七年加入民盟, 一九九二年加入中共。
任民盟华东师范大学委员会副主任,第八、九、十届上海市政协委员。
曾任第八届民盟上海市委常委和高教委副主任。
时俭益于一九六七年高中毕业, 一九六九年到安徽省来安县插队落户, 一九七四年返沪, 同年十一月到华东师范大学历史系培训班学习, 一九七六年初毕业后任华东师范大学图书馆职员。
自学修完数学本科课程, 于一九七八年考取华东师范大学数学系代数群专业研究生, 一九八一年六月毕业并获理学硕士学位。
随即留校任教, 并公派留学英国瓦瑞克大学数学系, 一九八四年十月在该校获Ph. D. 学位。
一九八五年一月回华东师范大学数学系任教至今。
时俭益专攻代数群表示理论及其相关的组合数学。
出版专著与教科书各一部,发表论文三十五篇(其中三十一篇在国际SCI 核心刊物上,四篇在全国性杂志或论文集上)。
先后担任过十一名博士生与十八名硕士生的学位论文指导工作,主讲研究生、本科生与留学生课程十余门。
时俭益研究代数群与黑克代数的卡茨当--罗斯蒂克表示理论,在该理论的核心课题---- 仿射外尔群的胞腔理论方面取得了具有国际领先水平的突出成就。
他圆满地解决了型$\widetilde{A}$ 仿射外尔群的胞腔分解问题,以该成果为主要内容写成的专著《某些仿射外尔群的卡茨当--罗斯蒂克胞腔》在德国出版。
在卡茨当-罗斯蒂克表示理论研究中该书是被引用次数最多的基本参考文献之一。
美国MIT 著名的代数学家佛根教授在为该书作评论时写道: “这是非常漂亮的数学成果,内容叙述得清楚而完备,那些希望应用或推广其成果的数学家将会由衷地感谢作者”。
时俭益把在对称群上著名的鲁滨逊-宣斯坦特算法推广到型A仿射外尔群上。
从而深刻地揭示了该族仿射外尔群左胞腔的性质。
同时也为刻画其它典型仿射外尔群的胞腔提供了组合论模式。
非平凡循环子群共轭类类数较小的有限非可解群
非平凡循环子群共轭类类数较小的有限非可解群史江涛;张翠【摘要】本文完全刻画非平凡循环子群共轭类类数不大于2的有限群的结构,证明了非平凡循环子群共轭类类数不大于4的有限非可解群仅有 PSL2(r ),其中 r =5,7,8,9.%The structures of finite groups having at most two conjugacy classes of non-trivial cyclic sub-groups are completely characterized.It is proven that a finite non-solvable group G having at most four conjugacy classes of non-trivial cyclic subgroups must be isomorphic to PSL2(r ),where r =5,7,8,9.【期刊名称】《广西师范大学学报(自然科学版)》【年(卷),期】2014(000)003【总页数】5页(P52-56)【关键词】有限群;循环子群;非可解群【作者】史江涛;张翠【作者单位】烟台大学数学与信息科学学院,山东烟台 264005;烟台大学数学与信息科学学院,山东烟台 264005【正文语种】中文【中图分类】O152.1分类某些特殊子群具有较小共轭类类数的有限群是现代群论研究的一个重要课题。
比较早期和经典的结果是研究极大子群具有较小共轭类类数对有限群可解性的影响,见文献[1-2]。
之后,一些群论学者开始转向研究其他特殊子群的共轭类类数对有限群结构的影响。
非循环子群即生成元个数大于等于2的子群。
作为子群共轭类类数的进一步研究,李世荣等在文献[3]中分类了非循环真子群的共轭类类数等于1的有限群。
孟伟等在文献[4]中分类了非循环真子群的共轭类类数等于2的有限群。
设G为有限群,以v(G)表示G的非循环真子群的共轭类类数。
群的概念教学中几个有限生成群的例子
群的概念教学中几个有限生成群的例子霍丽君(重庆理工大学理学院重庆400054)摘要:群的概念是抽象代数中的最基本的概念之一,在抽象代数课程的教学环节中融入一些有趣的群例,借助于这些较为具体的群例来解释抽象的群理论,对于激发学生的学习兴趣以及锻炼学生的数学思维能力等方面都会起到一定的积极作用。
该文介绍了一种利用英文字母表在一定的规则下构造的有限生成自由群的例子,即该自由群的同音商,称为英语同音群。
此外,该文结合线性代数中的矩阵相关知识,给出了有限生成群SL2(Z )以及若于有限生成特殊射影线性群的例子。
关键词:有限生成群英语同音群一般线性群特殊射影线性群中图分类号:O151.2文献标识码:A文章编号:1672-3791(2022)03(b)-0165-04Several Examples of Finitely Generated Groups in the ConceptTeaching of GroupsHUO Lijun(School of Science,Chongqing University of Technology,Chongqing,400054China)Abstract:The concept of group is one of the most basic concepts in abstract algebra.Integrating some interesting group examples into the teaching of abstract algebra course and explaining the abstract group theory with the help of these more specific group examples will play a positive role in stimulating students'learning interest and training students'mathematical thinking ability.In this paper,we introduce an example of finitely generated free group by using the English alphabet under some certain rules,which is called homophonic quotients of free groups,or briefly called English homophonic group.In addition,combined with the theory of matrix in linear algebra,we give some examples of about finitely generated group SL_2(Z)and finitely generated special projective linear groups.Key Words:Group;Finitely generated group,English homophonic group;General linear group;Special projective linear group1引言及准备知识群是代数学中一个最基本的代数结构,群的概念已有悠久的历史,最早起源于19世纪初叶人们对代数方程的研究,它是阿贝尔、伽罗瓦等著名数学家对高次代数方程有无公式解问题进行探索的结果,有关群的理论被公认为是19世纪最杰出的数学成就之一[1-2]。
丘成桐大学生数学竞赛数学专业大纲(英汉对照版)
S.T.YAU College Student Mathematics Contests
Algebra, Number Theory and Combinatorics (second draft)
代数、数论和组合(修改稿)
Linear Algebra Abstract vector spaces; subspaces; dimension; matrices and linear transformations; matrix algebras and groups; determinants and traces; eigenvectors and eigenvalues, characteristic and minimal polynomials; diagonalization and triangularization of operators; invariant subspaces and canonical forms; inner products and orthogonal bases; reduction of quadratic forms; hermitian and unitary operators, bilinear forms; dual spaces; adjoints. tensor products and tensor algebras; 线性代数 抽象向 量空间 ;子 空间; 维度; 矩阵 和线性 变换 ;矩阵 代数 和群; 行列式 和迹 ;特 征向量 和特征 值; 特征多 项式和 最小 多项式 ;对 角化和 三角 化;不 变子空 间和 典范 形式; 内积和 正交 基;二 次型的 化简 ;埃尔 米特 算子和 酉算 子;双 线性型 ;对 偶空 间;伴随矩阵,张量积和张量代数; Integers and polynomials Integers, Euclidean algorithm, unique decomposition; congruence and the Chinese Remainder theorem; Quadratic reciprocity ; Indeterminate Equations. Polynomials, Euclidean algorithm, uniqueness decomposition, zeros; The fundamental theorem of algebra; P olynomials of integer coefficients, the Gauss lemma and the Eisenstein criterion; Polynomials of several variables, homogenous and symmetric polynomials, the fundamental theorem of symmetric polynomials. 整数和多项式 整数 , 欧 几里 得 算法 , 唯 一分 解 ; 同余 和 中国 剩 余 定理 ; 二次 互 反 性; 不 定方 程. 多项式 ,欧几 里得 算法, 唯一分 解, 零点; 代数 基本定 理; 整系数 多项式 ,高 斯引 理和爱 森斯坦 判别 法;多 元多项 式, 齐次多 项式 和对称 多项 式;对 称多项 式基 本定 理. Group Groups and homomorphisms, Sylow theorem, finitely generated abelian groups. Examples: permutation groups, cyclic groups, dihedral groups, matrix groups, simple groups, Jordan-Holder theorem, linear groups (GL(n, F) and its subgroups), p-groups, solvable and nilpotent groups, group extensions, semi-direct products, free groups, amalgamated products and group presentations. 群 群和同态, 西罗定 理,有限 生成阿 贝尔群. 例子:置 换群、循 环群、二 面体群、 矩阵 群、单群、若尔当-赫尔德定理、线性群(GL(n,F)及其子群)、p群、可解群和 幂零群,群扩张,半直积,自由群,融合积和群表示. Ring Basic properties of rings, units, ideals, homomorphisms, quotient rings, prime and maximal ideals,
国台湾师范大学九十三学博士班考试入学招生试题
-2-
σ ∈G
(a) Give an example that N K / F is not surjective. (b) Suppose that F is a finite field. Show that G is cyclic by describing G explicitly. (c) Suppose that F is a finite field. Show that N K / F is surjective.國立臺灣師範源自學九十三學年度博士班考試入學招生試題
代數科試題 (數學研究所用,本試題共 2 頁第 1 頁)
注意: 依次序作答,只要標明題號,不必抄題。 答案必須寫在答案卷上,否則不予計分。
總分 100 分,每題 20 分 Notations: Z : the ring of integers Q : the field of rational numbers C : the field of complex numbers 1. Let G be a finite group and let H be a subgroup of G. Recall that the normalizer of H in G is defined by N(H)={a ∈ G | aHa −1 =H}. (a) Show that N(H) is the largest subgroup of G in which H is normal. (b) Show that the number of distinct conjugates xHx −1 of H in G is equal to [G : N(H)]. (c) Suppose that G is of order 48. Show that G is not simple. 2. Consider the permutation group S 4 and the alternating group A 4 . (a) Describe all the conjugacy classes of S 4 and write down its class equation. (b) Show that there exists a normal subgroup K of A 4 which is isomorphic to the Klein four group. (c) Show that S 4 is solvable. (d) Does A 4 have a subgroup of order 6 ? 3. An element α of C is said to be integral over Z if there exists a nonzero monic polynomial f(x) over Z such that f( α )=0. (a) Show that the following statements are equivalent : (i) α is integral over Z. (ii) There exists a nonzero finitely generated Z-submodule M of C such that α M ⊂ M. (b) Determine all the elements of the quadratic field Q( − 3 ) which are integral over Z. 4. For a finite Galois extension K/F with Galois group G , the norm map N K / F : K → F is defined by N K / F (a)= ∏ σ (a) , where a ∈ K.
基于E-DP和DLP一种无限非交换群上的密钥交换协议
2022年 5月 May 2022Digital Technology &Application 第40卷 第5期Vol.40 No.5数字技术与应用中图分类号:TP309 文献标识码:A 文章编号:1007-9416(2022)05-0228-03DOI:10.19695/12-1369.2022.05.70基于E-DP和DLP一种无限非交换群上的密钥交换协议东莞职业技术学院现代工业创新实践中心 张静首先,本文提出与分解问题等价的一个困难问题:等价分解问题。
其次,基于等价分解问题(E-DP)和离散对数问题(DLP),提出了一种无限非交换群上的密钥交换协议。
在协议中,通过半直积的运算法则,使共享密钥同时包含两个困难问题。
这两个困难问题共同保证了密钥交换协议的抗攻击性。
最后利用代数攻击和暴力攻击进行分析,证明了协议具有较高的安全性。
Sidelnikov等人提出在无限非交换群和半群上建立公钥密码系统的思想[1]。
非交换群上的公钥密码系统主要是应用困难问题隐藏因子,公开的困难问题有共轭搜索问题、分解问题、子群元素搜索问题、离散对数问题、同态搜索问题等。
许多研究已经提出了一些非交换代数结构上基于困难问题的密钥交换协议。
在文献[2]中,作者提出了无限非交换群上基于分解问题的密钥交换协议[2]。
在文献[3]中,作者提出了辫群上基于共轭搜索问题的一种新的密码系统[3],其中辫群是一种无限非交换群,并具有很好的密码特征。
随着量子算法和分解算法的发展[4],只应用一个困难问题已经不能保证密码系统的安全性。
而且单独应用共轭搜索问题构建的密码系统也已经被证明对于安全性是不充分的[5]。
为了提高安全性需要同时使用几个困难问题,2007年,Eligijus等人同时使用共轭搜索问题和离散对数问题在非交换群上构建了一种密码协议,并阐明了同时使用两个困难问题足够保证密码交换协议的安全性[6]。
2013年,Maggie Habeeb等人第一次将群的半直积用到密码系统中,并提出了一种基于半直积运算的密码交换协议[7]。
有限非交换单群与S3-猜想
离散单群 M 11 M 12 M 22 M 23 M 24 J1 J2 J3 J4
共轭类长 720 8640 40320
443520 10644480
9340 40320 3684840 1314781379486023680
离散单群 Co1 Co2 Co3 Fi22 Fi23 Fi24 HS
McL He
569
理论数学
罗雨玥,刘燕俊
则:① 当 CSn (σ ) 含有一个奇置换时, Kσ 是 An 的一个共轭类;② 当 CSn (σ ) 不含有奇置换时, Kσ 在 An
中分裂为以下两个长度相同的共轭类:
{ } { } = Kσ′ τστ −1 τ ∈ Sn ,τ 是偶置换 = ; Kσ′′ τστ −1 τ ∈ Sn ,τ 是奇置换 .
别共轭 n, n, 2n, 2n +1, 2n, 2n 个半单元素的幂。
证明:这是([7],推论 3.4)。
□
引理 2.6:设 σ ∈ An ,Kσ 是 An 中所有与 σ 有相同类型置换的集合,考虑 σ 在 Sn 中的中心化子 CSn (σ ) ,
DOI: 10.12677/pm.2019.95076
罗雨玥,刘燕俊
Copyright © 2019 by author(s) and Hans Publishers Inc. This work is licensed under the Creative Commons Attribution International License (CC BY). /licenses/by/4.0/
2F4 (2)′
共轭类长
5031936000
21349785600 14865016320 6825772800000
抽象代数
一、课程目的与教学基本要求本课程是在学生已学习大学一年级“几何与代数”必修课的基础上,进一步学习群、环、域三个基本的抽象的代数结构。
要求学生牢固掌握关于这三种抽象的代数结构的基本事实、结果、例子。
对这三种代数结构在别的相关学科,如数论、物理学等的应用有一般了解。
二、课程内容第1章准备知识(Things Familiar and Less Familiar)10课时复习集合论、集合间映射及数学归纳法知识,通过学习集合间映射为继续学习群论打基础。
1、几个注记(A Few Preliminary Remarks)2、集论(Set Theory)3、映射(Mappings)4、A(S)(The Set of 1-1 Mappings of S onto Itself)5、整数(The Integers)6、数学归纳法(Mathematical Induction)7、复数(Complex Numbers)第2章群(Groups) 22课时建立关于群、子群、商群及直积的基本概念及基本性质;通过实例帮助建立抽象概念,掌握群同态定理及其应用;了解有限阿贝尔群的结构。
1、群的定义和例子(Definitions and Examples of Groups)2、一些简单注记(Some Simple Remarks)3、子群(Subgroups)4、拉格朗日定理(Lagrange’s Theorem)5、同态与正规子群(Homomorphisms and Normal Subgroups)6、商群(Factor Groups)7、同态定理(The Homomorphism Theorems)8、柯西定理(Cauchy’s Theorem)9、直积(Direct Products)10、有限阿贝尔群(Finite Abelian Groups) (选讲)11、共轭与西罗定理(Conjugacy and Sylow’s Theorem)(选讲)第3章对称群(The Symmetric Group) 8课时掌握对称群的结构定理,了解单群的概念及例子。
特征类内类间 指标
特征类内类间指标English:The feature within-class and between-class indices are measures used in pattern recognition and machine learning to evaluate the quality of feature extraction or selection methods. These indices allow us to assess how well a set of features discriminates between different classes or categories.The within-class scatter matrix (Sw) is a measure of the spread or variation within each class. It quantifies how much the data points of the same class deviate from their centroid or mean. A smaller within-class scatter matrix indicates that the data points within each class are closer to their respective mean, suggesting better separability between classes.On the other hand, the between-class scatter matrix (Sb) measures the separation between different classes. It quantifies how much the centroids of different classes deviate from each other. A largerbetween-class scatter matrix suggests a greater separation between classes, indicating better discriminative power of the chosen features.Based on these matrices, feature evaluation indices can be computed. One commonly used index is the Fisher's discriminant ratio (FDR), which is defined as the ratio of the between-class scatter matrix to the within-class scatter matrix. A higher FDR value indicates better feature discriminability.Another commonly used index is the Bhattacharyya distance, which calculates the overlap or similarity between the class distributions based on their means and covariances. A smaller Bhattacharyya distance indicates better class separability.These feature evaluation indices are important in feature selection and extraction tasks, as they help us determine the most informative and discriminative set of features for a given classification problem. They provide a quantitative measure of the quality of the selected or extracted features, allowing us to compare different feature sets or algorithms and select the most appropriate ones.中文翻译:特征类内类间指标是用于模式识别和机器学习中评估特征提取或选择方法质量的指标。
具有给定共轭类长的有限群
具有给定共轭类长的有限群邵长国;蒋琴会【摘要】通过有限群G的共轭类长集合cs(G)来刻画有限群A6和S6,得到如下结论:如果cs(G)=cs(G)={1,p3·r,p·q2·r,p3·q2,q2·r},则G≌A6;如果cs(G)={1,q·r,p3·r,q2·r,p·q2·r,p3 ·q·r,p4·q2},则G≌S6.%This paper characterizes the groups A6 and S6 by the set of size of conjugacy classes, and gets the following results: If cs(G) = cs(G) = {1, p3 · R,p · Q2 · R,p3 · Q2,q2 · R} , then G=A6; If cs(G) = {1 ,q · R,p3 · T,q2 · R,p · Q2 · R,p3 · Q · R,p4 · Q2 } , thenG≌S6.【期刊名称】《上海大学学报(自然科学版)》【年(卷),期】2011(017)005【总页数】3页(P600-602)【关键词】有限群;共轭类;单K3-群;素图【作者】邵长国;蒋琴会【作者单位】上海大学理学院,上海200444;上海大学理学院,上海200444【正文语种】中文【中图分类】O152.1共轭类的某些数量性质与有限群结构之间的关系是有限群论研究的重要课题之一.在共轭类的数量性质中,有关共轭类长的研究非常活跃.如何通过类长来刻画有限群的某些性质是人们比较感兴趣的课题[1-3].本工作将通过有限群G的共轭类长集合cs(G)来刻画有限群A6和S6,并得到如下定理.定理1 设G为一有限群,且Z(G)=1,如果cs(G)={1,p3·r,p·q2·r,p3·q2,q2·r},则G≅A6.定理2 设G为一有限群,且Z(G)=1,如果cs(G)={1,q·r,p3·r,q2·r,p·q2·r,p3·q·r,p4·q2},则G≅S6.本工作用π(G)表示有限群G阶的素因子集合,πe(G)为G的元素阶的集合,πc(G)={p|p|n,n∈cs(G)}.若G的元素的阶为素数幂,则称群G为质幂元群;若|π(G)|=3,则有限单群G称为单K3-群.群G的素图Γ(G)定义如下:图的顶点集为群G的阶的所有素因子组成的集合,2个顶点p和q邻接,当且仅当pq∈πe(G).Γ(G)的连通分支数用t(G)表示,连通分支用πi表示,i=1,2,….若G 为偶数阶群,则总假定2∈π1.1 预备知识引理1[4]设G为一有限群,且Z(G)=1,则π(G)=πc(G).引理2[5]如果G有2个元素x和y,使得xy= yx,且(o(x),o(y))=1,则CG(xy)=CG(x)∩CG(y).引理3[6]若G为一有限可解质幂元群,则|π(G)|≤2.引理4[1]若G为一单K3-群,则G≅A5(22·3· 5),A6(23·32·5),L2(7)(23·3·7),L2(8)(23· 32·7),L2(17)(24·32·17),L3(3)(24·33·13),U3(3)(25·33·7)或U4(2)(26·34·5).引理 5[7]设 G为偶阶 2-Frobenius群,则t(G)=2,且G有正规列1HKG,使π(K/H)= π2,π(H)∪π(G/K)=π1,G/K和K/H均为循环群.特别地,|G/K|<|K/H|,G可解.引理6[8]设G为一有限群,其素图分支个数大于1,则G有下列情形之一.(1)G为Frobenius群或2-Frobenius群.(2)G有如下的正规列:1HKG,H和G/K为π1-群,H为幂零群,K/H为一单群,且|G/K|| |Aut(K/H)|.2 定理证明定理1 设G为一有限群,且Z(G)=1,如果cs(G)={1,p3·r,p·q2·r,p3·q2,q2·r},则G≅A6.证明由引理1,π(G)={p,q,r}.不妨设|G|=pa·qb·rc.因cs(G)={1,p3·r,p·q2·r,p3·q2,q2·r},有a≥3,b≥2,c≥1.设P∈Sylp(G),1≠x∈Z(P),则|xG|=|G:CG(x)|||G:P|,因此,p|/|xG|.因|xG|∈cs(G),所以|xG|=q2·r,|CG(x)|=pa·qb-2·rc-1.如果c>1,设R∈Sylr(CG(x)),则存在R1∈Sylr(G),使得RR1,因此,Z(R1)∩R≠1.取1≠y∈Z(R1)∩R,则xy=yx,且(o(x),o(y))=1.由引理2,CG(xy)=CG(x)∩CG(y),因此,|(xy)G|=|G:CG(xy)|=|G:CG(x)∩CG(y)|.进而有|xG|| |(xy)G|,且|yG|||(xy)G|.由|xG|∈cs(G),|yG|∈cs(G),得p3·q2·r||(xy)G|,这与|(xy)G|∈cs(G)矛盾.因此,c=1.如果b>2,设1≠x1∈Z(R1),R1∈Sylr(G).所以,r|/||,又||∈cs(G),则||=p3·q2,|CG(x1)|=pa-3·qb-2·r.设Q∈Sylq(CG(x1)),则存在Q1≤G,使得QQ1,且|Q1|=qb-1.因此,Q∩Z(Q1)≠1.取1≠z∈Q∩Z(Q1),则|zG|=p3·r.同上可得,p3·q2·r||(x1z)G|,这与|(x1z)G|∈cs(G)矛盾.因此,b=2.同理可以证明,a=3.因此,|G|=p3·q2·r.设P∈Sylp(G),Q∈Sylq(G),R∈Sylr(G),1≠x∈Q,1≠y∈R,则|CG(x)|=q2,|CG(y)|=r.又设1≠z∈Z(P),则|CG(z)|=p3;若z∈PZ(P),则|CG(z)|=p2.这说明G为一质幂元群,由引理3知,G非可解且G必有一个主因子H∶=H/N为单K3-群.由引理4知≅A5,A6,L2(7)或L2(8).若≅A5,则||= 24·3·5,进而得到||=23·3·5或||=22·3· 5.如果|G|=23·3·5,则 |N|=3或5.但此时对于任意的1≠n∈cs(G),n>5.由于N为G的一些共轭类的并,矛盾.如果||=22·3·5,则|π(N)|=2.所以,存在一个素数r∈π(G)-π(N)和Gr∈Sylr(G).于是Gr以共轭方式作用在N上.由于G为一质幂元群,所以这个作用是无不动点的,因此N幂零,矛盾.若≅A6,则|G|=23·32·5,此时,G= ≅A6.若≅L2(7),则因CG/N(H/N)=1,得到G/N<~Aut(L2(7)),所以,|G/N|=23·3·7,且|N|=3或7.若|N|=3,则P7∈Syl7(G)无不动点地作用在N上,可得|P7|||N|-1,即7|3-1,矛盾.若|N|=7,则P2∈Syl2(G)无不动点地作用在N上,所以8|7-1,矛盾.若≅L2(8),则|G|=23·32·7,此时,G≅ L2(8).由文献[9]知,cs(L2(8))≠cs(G),矛盾.因此,G≅A6.推论1 设G为一有限群,且Z(G)=1,如果cs(G)={1,40,45,72,90},则G≅A6.定理2 设G为一有限群,且Z(G)=1,如果cs(G)={1,q·r,p3·r,q2·r,p·q2·r,p3·q·r,p4·q2},则G≅S6.证明类似定理1的证明,可以得到|G|=p4· q2·r.易知G为一Crr-群,且G的素图分支为2个.又由G的某个共轭类长为p·q2·r知,G的Sylow p-子群P非交换,且只有Z(P)中的p-元素与Q∈Sylq(G)的q-元素交换.由集合cs(G)可知,∀1≠a∈Z(P),|aG|=q·r或q2·r,且|CG(a)|=p4·q或p4.从而,G无p·q2阶元.若G中含有q2阶元y,则|yG|=p3·r,|CG(y)|=p·q2,这与G中不含p·q2阶元矛盾.故G不含q2阶元.同理可知,G中不含p2·q阶元.下证G既不是Frobenius群,也不是2-Frobenius群.设G=K×|H为Frobenius群.若K∈Sylr(G),则由P非交换可知,P 为一2-群,又是H的Sylow q-子群,从而也是G的Sylow q-子群循环,矛盾.若H∈Sylr(G),对于∀1≠x∈Z(Q),则 p,q|/|xG|,这与|xG|∈cs(G)矛盾.设G为2-Frobenius群,则G有如下的正规列:1HKG,其中K是以H为核的Frobenius群,G/ H是以K/H为核的Frobenius群.由Frobenius群的性质知,|K/H|=r,π(H)∪π(G/H)={p,q}.显然,q2|/|G/K|.若q|||G/K|,则q|r-1.由q||H|,则r|q-1,矛盾.若q|/|G/K|,则q2||H|.假设p||H|且|H|p= pt.由于K是以H为核的Frobenius群,故r|pt-1.又G/H是以K/H为核的Frobenius群,所以,p4-t|r.于是,t=3.这时可以得到|H|=p3·q2.任取H的一个非零q-元x,则|xG||p·r.因|xG|∈cs(G),所以,|xG|=1,且x∈Z(G),矛盾.因此,|H|=q2且|G/K|=p4.由引理5可知,G/K循环,因此,G中存在p4阶元,矛盾.因此,G既不是Frobenius群,也不是2-Frobenius群.所以,由引理6,G有如下的正规列:1H KG,使得G/K,H为π1-群,K/H为一单群.由引理6及引理4得,K/H≅A5,A6,L2(7),L2(8)或L2(17).若K/H≅A5,则π2={3}或π2={5}.如果π2= {3},则π1={2,5}.下面分p=2或5两种情形来讨论.假设p=2,则q=5且r=3.于是,可以得到|G|=24·52·3.因CG/H(K/H)=1,所以,G/H<~Aut(A5)=S5,即|G/H|=22·3·5或23·3·5.于是,|H|=22·5或2·5.设P3∈Syl3(G),则P3互素地作用在H上.由于H幂零,且G中不含15阶元,P3无不动点地作用在H5∈Syl5(H)上.因此,|P3|| |H5|-1,即3|5-1,矛盾.假设p=5,则q=2,且r=3.于是,可以得到|G|=54·22·3.同上可得,矛盾.若π2={5},π1={2,3},下面再分p=3,q=2或p=2,q=3两种情况进行讨论. 假设p=3,q=2,则同上可得,矛盾.故只有p= 2,q=3,r=5,即π1={2,3},π2={5}.因|G|=24· 32·5,故|H||12.又因H为G的一些共轭类的并,但∀1≠n∈cs(G),n≥15,所以,H=1,K≅A5.因K◁—G,故K为G的一些共轭类的并.于是,|K|=1+ 15k1+40k2+45k3+90k4+120k5+144k6,此方程无非负整数解,故K/H≇A5.同理,K/H≇L2(7),L2(8)或L2(17).因此,K/H≅A6.同上可得,H=1,K≅A6.又因CG(K)=1,所以,A6≤G≤Aut(A6).由文献[10]中的引理2可知,只有G≅S6满足6∈πe(G),且10∉πe(G).因此,G≅S6.推论2 设G为一有限群,且Z(G)=1,如果cs(G)={1,15,40,45,90,120,144},则G≅S6.参考文献:[1] HERZOGM.On finite simple groups of order divisible by three primes only[J].J Algebra,1968,10:383-388.[2]ITON.On finite groups with given conjugate typesⅠ[J].Nagoya Math,1953,6:17-28.[3]ITON.On finite groups with given conjugate typesⅡ[J].Osaka J Math,1970,7:231-251.[4]陈贵云.关于Thompson猜想[M].北京:中国科学技术出版社,1992. [5]CHENG Y.On Thompson’s conjectur e[J].J Algebra,1996,185:184-193.[6] HIGMANG.Finite groups in which every element has primepowerorder[J].Journalofthe London Mathematical Society,1957,32(3):335-342.[7]陈贵云.Frobenius群与2-Frobenius群的结构[J].西南师范大学学报:自然科学版,1995,20(5):185-187.[8] WILLAMSJ S.Prime graph components of finite groups[J].J Algebra,1981,69(2):487-513.[9] CONWAYJ H,CURTISR T,NORTONS P,et al.Atlas of finite groups [M].Oxford:Clarendon Press,1985.[10] LUCIDOM S.Prime graph componets of finite almost simple groups [J].Rend Sem Mat Univ Padova,1999,102:1-22.。
Connected linear groups as differential Galois groups
Theorem 1.2 Let G be a connected linear algebraic group de ned over an
A A Y 0 = ( x ?1 + : : : + x ?d(G) + A1)Y
1
d(G)
where Ai; i = 1; : : : ; d(G) are constant matrices and A1 is a matrix with polynomial entries of degree at most e(G). In particular, the only possible singularities of this system are d(G) regular singular points in the nite plane and a (possibly irregular) singular point at in nity.
Previous work of Kovacic 17, 18] reduced this problem to the case of powers of a simple connected linear algebraic group. Our contribution is to show that one is able to realize any connected semisimple group as a Galois group and, when k = C (x); x0 = 1 to control the number and types of singularities when one constructs a system Y 0 = AY; A 2 Mn (C (x)) realizing an arbitrary connected linear algebraic group as its Galois group. More precisely we
INCREASING CLASS-COMPONENT TESTABILITY
INCREASING CLASS-COMPONENT TESTABILITYSupaporn Kansomkeat Department of Computer Engineering Faculty of EngineeringChulalongkorn Univesity Bangkok, 10330, Thailand. (+66) 2218-6988 Supaporn.K@Student.chula.ac.thJeff OffuttInformation and Software EngngGeorge Mason UniversityFairfax, VA 22030, USA(+1) 703-993-1654 / 1651ofut@Wanchai RivepiboonDepartment of Computer EngineeringFaculty of EngineeringChulalongkorn UnivesityBangkok, 10330, Thailand (+66) 2218-6988wanchai.r@chula.ac.thABSTRACTTestability has many effects on software. In general, increasing testability makes detecting faults easier. However, increasing testability of third party software components is difficult because the source is usually not available. This paper introduces a method to increase component testability. This method helps a user test when the component is reused during integration. First, we analyze a component to gather definition and use information about method and class variables. Then, this information is used to increase component testability to support component testing. Increased testability helps to detect errors, and helps testers observe state variables and generate inputs for testing. This paper uses an example to report the effort (in terms of test cases) and effectiveness (in terms of killed mutants).KEY WORDSSoftware Testing, Software Testability, Component Software1.IntroductionSoftware testing is one of the most common ways to assure software quality and reliability, and is made easier by high software testability. Several different definitions of testability have been published. According to the 1990 IEEE standard glossary [8], testability is the “degree to which a component facilitates the establishment of test criteria and the performance of tests to determine whether those criteria have been met.” Voas and Miller [14] explained that testability enhances testing and claimed that increasing component testability is a primary key to improving the testability of component-based software. Their definition of software testability focuses on the “probability that a piece of software will fail on its next execution during testing if the software includes a fault.”Binder [2] defined testability in term of controllability and observability. Controllability is the probability that users are able to control component’s inputs (and internal state). Observability is the probability that users are able to observe component’s outputs. If users cannot control the input, they cannot be sure what caused a given output. If users cannot observe the output of a component under test, they cannot be sure if the execution was correct.Likewise, Freedman [4] considered testability based on the notions of observability and controllability. Observability captures the degree to which a component can be observed to generate the correct output for a given input. Controllability refers to the ease of producing all values of its specified output domain.With object-oriented software, testing needs to place more emphasis on testing the connections among components [9]. Researchers [5, 7, 13] have proposed using information from developers to test object-oriented components. Gallagher and Offutt [5] use information about object states for integration testing, including a finite state machine and a list of which state variables each method defines and uses. Harrold and Rothermel [7] use the data flow relations from the program source to guide the selection of tests. Tsai, Stobart and Parrington [13] seek the definitions and uses of data members from code statements for testing classes. Although they are interested in definitions and uses of data, which are similar to this work, their paper used all variables and source code for testing.An important property of a software component is implementation transparency, which means the implementation is not available. This means that testers have very little information about the internal state of the component. This lack of information makes it difficult to apply directly traditional white-box techniques [15] and difficult to fully exercise the software (lack of controllability) and also difficult to know the result of execution (lack of observability).Testers can increase testability in several ways:-gather information from components withoutsource code-increase observability to monitor outputs-increase controllability to support inputs-choose test criteria and generate test cases forcomponent testing based on the criteriaThis research project specifically assumes the component is objected-oriented. Furthermore, no access to the source is assumed. First, our method analyzes a compiled component to extract definition and use information. Then, the collected information is used toincrease component testability. This information provides ways to detect some errors, to observe state variables and to generate tests for when a component is reused in new environment. Test cases are generated to satisfy couplingcriteria [9]. This method does not consider inheritance and polymorphism relationships.2.BackgroundSeveral definitions of testability were given in Section 1. Our work defines component testability as the degree to which a component supports detection, observability and controllability. Detection focuses on the ease of detecting faults. Observability focuses on the ease of observing outputs. This means that the component supports ways to observe or monitor the results of testing. Moreover, an observable component allows not only the output of tests to be observed, but also intermediate values. Controllability focuses on the ease of controlling component’s inputs. This means that a component supports ways to supply inputs that exercise the component as necessary.This research presents techniques to create and supply tests for a black box component, and applies it by using data flow relations between methods to guide the selection of tests. Beizer [1] defined integration testing as focusing on testing interfaces between methods. Coupling between methods measures the dependency relations by the data and control flow interconnections between methods. Thus, brief overviews of data flow analysis and coupling-based testing are given in this section.2.1 Data Flow AnalysisData flow testing [12] requires tests to execute paths from statements that contain assignments to variables in a program, to statements where variables are used. A definition (def) is a statement where a variable’s value is stored into memory. A use is a statement where a variable’s value is accessed. A definition-use pair (or du-pair)of a variable is an ordered pair of a definition and a use, such that there is a path from the def to the use. Data flow testing criteria are used to select particular definition-use associations to test. Two of the most simple data flow testing criteria were first defined by Laski and Korel [10]. They proposed the all-definitions criterion, which requires that a test should cover a path from each definition to at least one use, and the all-uses criterion, which requires a test to cover a path from each def to all reachable uses.2.2 Coupling-based TestingJin and Offutt [9] proposed coupling-based testing (CBT) as an application of data flow testing to the integration level. CBT requires the program to execute from definitions of variable in a caller to uses of the corresponding variables in the callee unit. The variables can be parameters, global and non-local variables, and external references. Unfortunately, directly applying either the all-defs or the all-uses criterion is very expensive, both in terms of number of du-pairs and the difficulty of resolving the paths. Therefore, CBT is only concerned with definitions of variables just before calls (last-defs) and uses of variables just after calls (first-uses). The criteria are based on the following definitions:•A Coupling-def is a statement that contains a last-def that can reach a first use in another method on at least one execution path•A Coupling-use is a statement that contains a first use that can be reached by a last-def in another method on at least one execution path•A coupling path is a path from a coupling-def to a coupling-useFour levels of coupling-based integration test coverage criteria are defined between two units:•Call-coupling requires that the test cases cover all call-sites of the called method in the caller method•All-coupling-defs requires that, for each coupling-def of a variable in the caller, the test cases cover at least one coupling path to at least one reachable coupling-use •All-coupling-uses requires that, for each coupling-def of a variable in the caller, the test cases contains at least one coupling path to each reachable coupling-use•All-coupling-paths requires that the test cases covers all coupling paths from each coupling-def of a variable to all reachable coupling-usesThese testing criteria cannot be directly applied to component-based software because the black-box nature of components prohibits full control flow and data flow analysis. This paper introduces a new analysis method that extracts just the essential information for data flow criteria from Java bytecode.ponent AnalysisAfter a developer implements an object-oriented component, it is assumed to be included into a library without the source code. Moreover, the developer does not provide information about component testing. Because of this lack of information, approaches that rely on detailed analysis of the program, such as the object testing approaches [5, 7, 13] discussed in section 1, cannot be performed. Therefore, we need a process to analyze components to find def and use information without using the source.This analysis only considers class’s state variables that describe of the class. For each method, the process gathers the definition and use information for every state variable. This information is collected from the intermediate form in Java bytecode, and is called DU-M (Definitions and Uses of Methods). Moreover, the firstuses and the last definitions of variables are indicated by the locations from the intermediate form in Java bytecode.This collected information is stored in the DU-M repository.The analysis process is as follows. First, a Java class,.class file, is transformed into the intermediate form in Java bytecode by Decompiler . Then, the Java bytecode is analyzed to collect the definition and use information.The Decompiler , DJ Java Decompiler, is freeware supported by Author NavExpress [11]. The DU-M analyzer was developed for this research and parses the intermediate form in Java bytecode to create DU-M s. The DU-M s are used to increase component testability, as explained in the following subsections.Figure 1 shows the collected information for a variable x, uddu (d indicates a definition and u indicates a use). The location of the first use is at 1, and last definitions are at 10 and 19.3.1An ExampleWe illustrate our technique with the example of a vending machine. The machine accepts quarter coins and dispenses products when two quarters are received.CoinBox has four state variables. total keeps track of how many quarters have been received over a sequence of transactions. curQtr keeps track of how many quarters have been received within one transaction. seleTy keeps track of a user’s current selection and availSeleVals is a constant that stores the selections that are currently available in the vending machine (these are modeled as integers for convenience in the example). Users can add quarters into the machine (addQtr()), ask the machine to return the quarters inserted and not consumed (returnQtr ()), and ask the machine to vend a specific item (vend()).The machine does not dispense the item if it is not available, if the quarters are insufficient, or if the selection is invalid.The Java source code for this example is shown in Figure 2, however, it should be noted that this source code is not available to a component tester. Each callout in this figure is part of the java-bytecode for the associated method. Each line in the callouts contains the location (statement number), getfield or putfield (getfield indicates a use and putfield indicates a definition), the field number,and the field name used to generate the DU-M . After thecallout, we denote the first use by first-use and the last definition by last-def of variables of each method. The DU-M s of each method of CoinBox example is shown in Table 1. The column DU-M shows the sequence of definition and use of a variable. The column First-use and Last-def show the location of first use and last definition of each variable in each method.Note that testers can use the CoinBox component to control only inputs of quarters and selections. State variables total, curQtr, seleTy and availSeleVals are private, thus cannot be directly controlled by testers. For example when two quarters are added into the machine, it dispenses a product (shown at line 36 in Figure 2).However, if line 38 is removed, the machine still dispenses an item. This would represent a fault that causes the state variable curQtr is incorrect. To increase controllability of state variables, the DU-M s keep state variables to help select inputs to states that should be tested.Method Name Variable Name DU-M First-useLast-def CoinBox()total d 26curQtr d 31seleTyd 36availSeleVals d21addQtr()curQtr u d 2 7returnQtr curQtr d 2vend()seleTy d u29 7curQtr u u u d u 11109totalu d 94 99available()availSeleVals u u 4seleTyu184. Component TestabilityAs said in Section 1, component testability generally refers to how easy it is to test. A high degree of testability indicates that any existing faults can be revealed relatively easily during testing, outputs of state variables can be observed during testing and inputs can easily be selected to satisfy some testing criteria. To increase testability, we provide processes to detect errors and increase observability and controllability.Before components are integrated with exiting software, they should be tested. The DU-M repository from the previous section can be used to perform preliminary checks and detect some faults. As shown in Figure 3, Component Anomaly Detector uses DU-M s. For example, a dd data flow anomaly occurs when a definition is followed by another definition without an intervening use. Figure 4 shows a dd anomaly that would occur if the programmer had made a mistake of writing “total =curQtr-VAL” instead of “curQtr = curQtr-VAL” at line 38in Figure 2.Java-bytecodeTable 1. The DU-Ms of CoinBoxTo increase observability and controllability, we collect definition-use pairs between the last definition in a method and the first use in other methods of variable by using DU-M s of component. These definition-use pairs of a variable are called Definition-Use Coupling for Testing (DUCoT). The DUCoT s for CoinBox’s state variables are shown in Figure 5. This information is used to generate test cases. As shown in Figure 3, First-use and Last-def Generator creates a DUCoT for each variable. A DUCoT is defined next.Definition The DUCoT of variable V is a tuple, DUCoT(V ) = (D L , U F )• D L is a finite set of last definitions of variable V • U F is a finite set of first uses of variable VWe define observation points to monitor the state variables of the component. Note that variables are given values in definition statements. Therefore, observation points should be located immediately before and after definitions of each variable. Also, the output of tests should be observation points to show all state variables of the component. As shown in Figure 3, Visualization Generator provides a method to monitor state variables by using the DUCoT s. This supports monitoring of the status of component variables during testing.For example, when the test case addQtr(7),vend(11)is tested to cover DUCoT(curQtr), the state variable curQtr must be monitored before and after calling addQtr (). Furthermore, every state variable of CoinBox must be monitored after testing. Because addQtr() adds a quarter into the machine, the value of curQtr should be increased.The Visualization Generator uses Java reflection to show values of state variables.1 class CoinBox { 23private int total; 4private int curQtr; 5private int seleTy;67 8 9total = 0;10curQtr = 0;11seleTy = 0;12}1314void addQtr() {1516}171819 20}2122232425262728else if ( seleTy > MAXSEL )29System.err.println ("Wrong selection ");30else if ( !available( ) )31System.err.println ("Selection unavailable");32else {33 if ( curQtr < VAL )34 System.err.println ("Not enough coins");35 else {36 System.err.println ("Take selection");37 total = total+ VAL;38 curQtr = curQtr - VAL;39}40}42}434445 for (int i = 0; i<availSeleVals.length; i++)46 if (availSeleVals[i] == seleTy)47 return true;48return false;49}50} // class CoinBoxTo control inputs to exercise state variables in a component, the DUCoT s are used to generate test inputs according to coupling criteria. This paper uses the coupling-based criteria proposed by Jin and Offutt [9], as defined in Section 2.2. Applying coupling-based testing to component testing requires some minor modification to the terminology.Figure 2. Vending Machine example))))))Figure 4. A dd anomaly exampleFigure 3. The Component Testability ProcessTable 2. The All-coupling-defs of curQtr variable Table 4. The Test SequencesTable 5. The Mutation Operators and Number of Mutants Table 3. The All-coupling-uses of curQtr variable All-coupling-defs requires that for each coupling-def,at least one test case executes a path from the def to at least one coupling-use. A version of All-coupling-defs for a component is given in Definition 1 and the All-coupling-defs for curQtr are shown in Table 2.Definition 1 Let (D L , U F ) be a DUCoT of variable V . TheAll-coupling-defs of variable V is defined as AllCoD (V ) = (D L ×U F ) = { (d, u) | ∀d ∈ D L and ∃u ∈ U F ) }#AllCoD (curQtr)1CoinBox (31) , addQtr (2)2addQtr (7) , vend (11)3returnQtr (2) , addQtr (2)4vend (109) , vend (11)All-coupling-uses requires that for each coupling-def,at least one test case executes a path from the def to each reachable coupling-use. A version of All-coupling-uses for a component is given in Definition 2 and the All-coupling-uses for curQtr are shown in Table 3.Definition 2 Let (D L , U F ) be a DUCoT of variable V . TheAll-coupling-uses of variable V is defined as AllCoU (V ) = (D L ×U F ) = { (d, u) | ∀d ∈ D L and ∀u ∈ U F ) }#AllCoU(curQtr)1CoinBox (31) , addQtr (2)2CoinBox (31) , vend (11)3addQtr (7) , addQtr (2)4addQtr (7) , vend (11)5returnQtr (2) , addQtr (2)6returnQtr (2) , vend (11)7vend (109) , addQtr (2)8vend (109) , vend (11)Test cases can be computed from DUCoT s according to the selected level of coupling-based criteria. As shown in Figure 3, the Coupling-based Test Generator uses DUCoT s to generate test cases depending on the selected criteria. How to use these ideas for testing is illustrated through a case study.5. Case StudyThis section illustrates component testing on the vending machine class in Figure 2. Following our criteria in section 4, we generate 9 test cases for All-coupling-defs and 15 test cases for All-coupling-uses . We derive sequences of method calls for each test. One sequence can be used to realize multiple test cases. For example, the sequence [CoinBox(), returnQtr(), addQtr()] is derived for the test case returnQtr(2),addQtr(2). The sequence [CoinBox(), addQtr(), addQtr(), vend(2), vend(3)] is derived for test case vend(109),vend(11). Moreover, thissequence covers the test case addQtr(7),addQtr(2). In this way, 4 sequences are created for All-coupling-defs and 7sequences are created for All-coupling-uses as shown in Table 4.This example is based on the one used by Harrold et al. [6]. They published 25 test sequences that are grouped into 3 sets for the vending machine class in their paper.We use Harrold et al.’s 25 test sequences [6], named AllSequence , and 7 randomly selected test sequences from them to compare with our sequences. More precisely, we use 5 groups of 7 randomly selected test sequences,named Sampling1, Sampling2, …, Sampling5. Sampling1selects 7 test sequences using prime number order. The Sampling2, Sampling3, Sampling4 and Sampling5 select 3, 2 and 2 test sequences at random from 3 sets (1-16, 17-20, 21-25) that are grouped from 25 test sequences [6].Test sequences1. CoinBox() addQtr() vend(1)2. CoinBox() addQtr() addQtr() vend(2) addQtr() addQtr() vend(3)3. CoinBox() returnQtr() addQtr()4. CoinBox() addQtr() addQtr() vend(2) vend(3)A l l C o D5. CoinBox() returnQtr() vend(1)6. CoinBox() vend(2)A l l C o U7. CoinBox() addQtr() vend(21)The tests were evaluated by their ability to detect mutant-like faults. Delamaro et al. [3] proposed a set of interface mutation operators for integration testing. Our intent focuses on interface faults between methods, thus we use their operators to create faults for our case study.Six operators were used for our example; the others did not apply to our program. All possible mutants are generated according to these mutation operators, resulting in 40 mutants, as summarized in Table 5. These mutants were seeded into the vending machine class by hand.Mutation operator DefinitionNumber of mutants DirVarRepPar Replaces interface variable by each element of parameters of callee2DirVarRepGlo Replaces interface variable by each element of global variables accessed by callee18IndVarRepLoc Replaces non interface variable by each element of local variables in callee10DirVarAriNegInserts arithmetic negation at interface variable uses3DirVarLogNeg Inserts logical negation at interface variable uses5RetStaRepReplaces return statement2Total40Table 6 shows the results of executing each test set with mutants. The All-coupling-uses (AllCoU) tests resulted in a higher mutation score than any other sets,except Harrold et al.’s (AllSequence) tests. The All-Table 6. Results for a Vending Machinecoupling-defs (AllCoD) tests resulted in a higher mutation score than three sets of sampling. Although the AllCoD tests had the same mutation score as two sets of sampling,it has a smaller number of test sequences. The AllSequence tests resulted in 100% mutation score, but 3.6 times more test sequences than AllCoU.Criterion Number of test casesNumber of test sequencesKilled mutants Mutation Scores AllCoD 943792.5%AllCoU 1573997.5%Sampling1-7922.5%Sampling2-73792.5%Sampling3-71947.5%Sampling4-71845%Sampling5-73792.5%AllSequence-2540100%6. ConclusionsThe purpose of this research is to increase component testability. We have defined a method to increase testability of an object-oriented component whose source is not available. This method sets up and develops facilities to support testing. First, we analyze a component to gather definition and use information. This information shows the internal state variables of the component. Then,it is used to increase testability. These facilities provide ways to detect errors, to observe state variables and to control inputs for component testing by supporting test cases generation. This generation relies on the selected level of coupling-based criteria. Therefore, test cases are able to control inputs to fully exercise a component and improve the controllability. Our test set effectiveness is illustrated by a case study. Moreover, the output and intermediate values for each test case can be observed. This paper presents a way to increase testability in terms of detection, observability and controllability. In the future, we hope to apply our component testability ideas to another facet of controllability, specifically by allowing testers to control inputs with the goal of causing specific variables to have specific values during execution.7. AcknowledgmentsThis work was supported in part by Thailand’s Commission of Higher Education, MOE, and Center of Excellence in Software Engineering,Dept. of Computer Engineering, Faculty of Engineering,Chulalongkorn University. Thanks to the Department of Information and Software Engineering, School of Information Technology and Engineering, George Mason University, for hosting the first author during this research project.References:[1] B. Beizer, Software Testing Techniques (Van Nostrand Reinhold, Inc, New York NY, 2nd edition,1990).[2] R. V. Binder, Design for Testability with Object-Oriented Systems, Communications of the ACM , 37(9), 1994, 87-101.[3] M. E. Delamaro, J. C. Maldonado, and A. P. Mathur,Interface mutation: An approach for integration testing, IEEE Transactions on Software Engineering ,27(3), 2001, 228–247.[4] R. Freedman, Testability of software components,IEEE Transactions on Software Engineering , 17(6),1991, 553–563.[5] L. Gallagher and J. Offutt, Integration Testing of Object-oriented Components Using Finite State Machines, Under review.[6] M. J. Harrold, A. Orso, D. Rosenblum, G. Rothermel,M. L. Soffa, and H. Do, Using Component Metadata to Support the Regression Testing of Component-Based Software, Proc. IEEE International Conference on Software Maintenance , Florence, Italy, 2001, 154-163.[7] M. J. Harrold and G. Rothermel, Peforming Data Flow Testing on Classes, Proc. ACM SIGSOFT Foundation of Software Engineering , New Orleans, LA, 1994,154–163.[8] IEEE Standard Glossary of Software Engineering Technology, ANSI/IEEE 610.12, IEEE Press (1990).[9] Z. Jin and J. Offutt, Coupling-based criteria for integration testing, The Journal of Software Testing,Verification, and Reliability , 8(3), 1998, 133–154.[10] J. Laski and B. Korel, A Data Flow Oriented Program Testing Strategy, IEEE Transaction on Software Engineering , 9(3), 1983, 347-354.[11] Atanas Neshkov, “NavExpress DJ Java Decompiler”,.[12] S. Rapps and E. J. Weyuker, Selecting Software Test Data Using Data Flow Information, IEEE Transaction on Software Engineering , 11(4), 1985, 367-375.[13] B.-Y. Tsai, S. Stobart and N. Parrington, Employing Data Flow Testing on Object-oriented Classes, The IEE Proc. – Softw., 148(2), 2001, 56-64.[14] J. M. Voas and Miller K. W, Software Testability:The New Verification, IEEE Software , 12(3), 1995,17-28.[15] Y. Wu, M. Chen, and J. Offutt, UML-Based Integration Testing for Component-Based Software,Proc. 2nd International Conference on COTS-Based Software Systems , Ottawa, Canada, 2003, 251–260.。
关于有限群的正规子群的补子群I
关于有限群的正规子群的补子群I王坤仁【期刊名称】《四川师范大学学报(自然科学版)》【年(卷),期】2003(026)004【摘要】研讨了一个有限群的正规子群的补子群之存在性与共轭性的若干结果.主要的结果如下:设G/K是π-可解的并设H为有限群G的一个Hall π-子群,其中π=π(K),则有:(1) 若K的每个Sylow子群P1在G的某个含P1的Sylow子群中有补子群并且这个补子群在G中半正规,则K在G中有补子群,(2) 若进一步设K在H 中的所有补子群(由(1),这些补子群存在.)在H中共轭,则K在G中的所有补子群在G中共轭.%In this paper, some resutls on the existence and conjugacy of complements of a normal subgroup of a finite group are studied. The main result is as follows. Let G/K be π-solvable and let H be a Hall π-subgroup of Afinite group G with π=π(K), (1) if every Sylow subgroup P1 of K has a complement in a Sylow subgroup P of G containing P1 and all complements of P1 in P are seminormal in G, then K has a complement in G and (2) if, moreover, all complements of K in H are conjugate in H, then all complements of K in G are conjugate in G.【总页数】4页(P331-334)【作者】王坤仁【作者单位】四川师范大学,数学与软件科学学院,四川,成都,610066【正文语种】中文【中图分类】O152.1【相关文献】1.有限群正规子群的算法及S7子群的正规子群 [J], 王积社;饶金平2.关于有限群的正规子群的补子群Ⅱ [J], 王坤仁3.有限群的C-补和SS-拟正规子群 [J], 李世荣;彭峰;白彦如4.有限群的正规子群之与极大子群有关的补的几个结果 [J], 王坤仁5.有限群的Fuzzy拟正规子群和Fuzzy次正规子群 [J], 张桂生因版权原因,仅展示原文概要,查看原文内容请购买。
关于Sylow交的共轭类
北京大学学报(自然科学版),第41卷,第1期,2005年1月Acta Scientiarum NaturaIiumUniversitatis Pekinensis ,VoI.41,No.1(Jan.2005)1)国家自然科学基金(10401034)资助项目收稿日期:2003-10-31;修回日期:2004-09-14On Sylow Intersection Conjugacy Classes1)WANG Baoshan(School of Science of Beihang Uniuersity and Key Laboratory of Mathmatics ,Informatics ,and Behauioral Semantics(Beihang Uniuersity and Peking Uniuersity ),Ministry of Education ,Beijing ,100083,E-mail :bwang@ )Abstract Suppose that G is a finite group and p is a prime such that p I I G I .This paper studies the group in-variant n p (G ),the number of G -orbits of SyIow intersections ,which pIays an important roIe in the research of bIock theory and indecomposabIe moduIes of finite groups.The paper shows some properties for finite groupswith n p (G )=2,in particuIar ,it gives some necessary and sufficient conditions which show when a finite group G with n p (G )=2has the unigue maximaI normaI p -subgroup as an intersection of two distinct SyIow p -sub-groups.Key words syIow intersection ;bIock ;maximaI normaI p -subgroup0IntroductionLet G be a finite group and p a prime divisor of I G I .And denote by !the set of aII intersections of any two SyIow p -subgroups :!={P x !P y \P "Syl p (G );x ,y "G }.SyIow Theorems show that G can act on !by conjugation.Then !is partitioned into G -orbits!=#ni =1!i .Define n p (G )be the number of the G -orbits.It is weII known that aII SyIow p -subgroups are conjugatein G ,so suppose !1=Syl p (G ).It is easy to show that n p (G )=1if and onIy if P $G ,where P "Syl p (G ).The goaI is to investigate the case that n p (G )=2.In another word ,what can be said about the structure of G which acts on proper SyIow intersections transitiveIy Assume that n p (G )=2.Let {P ,D }be a set of representatives of G -orbits (or G -conjugacy cIasses of SyIow intersection )and P >D thoughout this paper ,if n p (G )=2.By the way ,P >D means that D is a proper subgroup of P .With above notation and assumption ,there exists x 0"G -N G (P )such that D =P !P x 0.This paper is organized as foIIows.In section 1,giving some eIementary resuIts on D and a neces-sary and sufficient condition for D = p (G ),where p (G )is the maximaI normaI p -subgroup of G .Insection 2,concentrating on the structure of G when D is a defect group or not.1Some Elementary ResultsThe foIIowing resuIts are easy to be proved ,and are gathered in one Iemma.Some conceptions wiII be introduced before stating the Iemma.Given a chain of p -subgroups!:O 0<O l <…<O n (l )of G ,define the Iength I !I =n ,the finaI subgroup V !=O n ,the initiaI subgroup V !=O 0,the k -th initiaI subchain!k :O 0<O l <…<O k ,and the normaIizerG !=N G (!)=N G (O 0)!N G (O l )!…!N G (O n ).(2)A p -subgroup R of G is radicaI if p (N G (R ))=R.A p -chain !given by (l )is said to be radi-caI if O i = p (N G (!i ))for each i ,i.e ,if O 0is a radicaI p -subgroup of G and O i is a radicaI p -sub-group of N G (!i -l )for each i "0.Denote by !=!(G )the set of aII radicaI p -chains of G.By [l ],the p -IocaI rank plr (G )of G is the number plr (G )=max {I !I :!#!(G )}.Then a subgroup H of G is a triviaI intersection (T.I.)if H g !H =l for every g #G -N G (H ).Lemma 1With the notation and assumption in preuious section ,if n p (G )=2,then the following statements hold :(l )D is a p-radical subgroup of G ,in other words ,D = p (N G (D ));(2)N /D has T.I.Sylow p-subgroups ,where N =N G (D );(3)if O is a p-subgroup of G and P $O l $O 2>D ,thenN G (P )$N G (O l )$N G (O 2).Some resuIts have been known on the SyIow intersection of finite groups.The theorems in [2]and [3]state that :Theorem 2[2]Let G be a finite p-soluable group and P #Syl p (G ).Then there is a Sylow p-sub-group ,O of G with P !O = p (G ),unless ,one of the following situations occur :(l )p =2,P / p (G )inuolues Z 2~Z 2and I p'p(G )I is diuisible by the sguare of a Fermat prime ,or the sguare of a Mersenne prime ;(2)p is Mersenne prime ,P / p (G )inuolues Z p ~Z p ,P / p (G )does not centralize 2(G / p (G )),and (p +l )p diuides I G I .Theorem 3[3]If the p-Sylow subgroups P of G are abelian ,then the intersection of all p-Sylow sub-groups of G appears as intersection of P with one of its conjugates.In the case ,n p (G )=2,what can be said about D as a SyIow intersection ?The Iemma beIow wiII answer the guestion when we have D = p (G ).Lemma 4Suppose that n p (G )=2,then D = p (G )if and only if there exists a normal sub-group ,M ,of G such that P !M =D.北京大学学报(自然科学版)第4l 卷第l期王宝山:关于SyIow交的共轭类Proof The“onIy if”condition is obvious,showing the“if”part is needed.By the assumption that D x C M for every x G G since M is normaI and G acts on the set!2of intersections of any two distinct SyIow p-subgrops transitiveIy.Now consider P n P y;y G G-N G(P).CertainIy suppose that P n P y =D z which is a subgroup of M,for some z G G,then D z C D=P n M.Furthermore,P n P y=D since D and D z have the same order.By[l,7.l],for any finite group G,if plr(G)>0,then plr(G)=l if and onIy if G/0p(G)has T.I.SyIow p-subgroups.Hence the foIIowings hoId for the case.Corollary5Suppose that n p(G)=2,then plr(G)=l if and only if there exists a normal sub-group,M,of G such that P n M=D.G.Robinson has toId that the intersection of a p-radicaI subgroup and a normaI subgroup is aIso a p-radicaI subgroup of the normaI subgroup in[l,4.6].So every p-radicaI subgroup of G must contain the maximaI normaI p-subgroup of G.The next Iemma wiII show the necessary condition when a p-sub-group of G is p-radicaI.Proposition6For n p(G)=2,suppose that0is a p-radical subgroup of G and0is not a Sylow p-subgroup of G,then0p(G)$0$G D.Proof Assume P G Syl p(G)and P>0.If there exists another SyIow p-subgroup,L,of G such that L一P and L>0then we are done.Now suppose converseIy that there is a unigue SyIow p-sub-group,P,that contains0.Then N G(0)n P is the unigue SyIow p-subgroup of N G(0),hence that N G(0)n P=0.But P>N P(0)>0,which is a contradiction.So the assertion is proved.2The Group Theoretic Characterization of!In this section,it is focused on that the group theoretic characterization of G via the moduIar repre-sentation properties of D.In the previous section the reIation between D and0p(G)has been de-scribed.This section first states a very usefuI theorem due to V.I.Zenkov and V.D.Mazurov.Theorem7[4]If G is a finite simple non-abelian group,then G contains a pair of Sylow p-sub-groups with a trioial intersection.For the remainder of the paper,assume n p(G)=2unIess a speciaI assertion.A direct coroIIary from above theorem can be gotten.Corollary8If G is a finite simple group with n p(G)=2,then D=l=0p(G).In particular,G has T.I.Sylow p-subgroups.Example[5]Let G be a non-abelian simple group with a non-cyclic T.I.Sylow p-subgroup P.Then G is isomorphic to one of the following groups:(l)PSL(g),where g=p n and n>2;2(g2),where g=p n;(2)PSU3(3)p=2and G22B2(22m+l);(4)p=3and G22B2(32m+l)and m>l;北京大学学报(自然科学版)第4l卷(5)p=3and G_PSL3(4)or M ll;(6)p=5and G_2F4(2)'or McL;(7)p=ll and G_4.For this reason,non-simpie groups wiii be concentrated on.Lemma9Suppose that G is not a simple group,and N is a minimal normal subgroup of G,if n p(G)=2and N is non-abelian,then D n N=l.Proof N is a direct product of isomorphic simpie groups:N=N l X…X N s,where N i’s are isomorphic simpie groups.If one of them is abeiian,so does N.Therefore N i is a simpie non-abeiian group.By above theorem,there are P',O'G Syl p(G)such thatD x'n N=P'n O'n N=(P'n N)n(O'n N)=l,where x'G G.So D n N=l since N is normai in G.Lemma10Suppose that G is not a simple group and M<.G,maximal normal subgroup of G,if n p(G)=2and D一0p(G)then one of the following statements holds:G=〈a〉M or G=N G(P)M,where a is a p-element,P is a Sylow p-subgroup of G.Proof Since G=G/M is a simpie group,if G is a p-group then G=〈a〉M,or G is a p'-group or non-abeiian simpie group.Then there are P l=P l M/M and P2=P2M/M which are Syiow p-groups of G such that P l n P2=l.Therefore,D x M<(P l n P2)M<P l M n P2M<M for some x G G,hence D<M and P n M>D since D一0p(G).By Frattini argument and iemma l,G=N G(P)M.Now it is introduced that two notions of strongiy and weakiy ciosure which are stated in[6].Definition11Let G be a group,P a Sylow p-subgroup of G,p any prime,and O a subgroup of P.O is strongly closed in P with respect to G prouided that wheneuer x G O and x g G P for g G G,so x g G O.Definition12G is also a group,H a subgroup of G,and K a subgroup or subset of H.SetV(ccl G(K);H)=〈K g I K g<H,g G G〉.(K);H)is the weak closure of K in H with respect to G.If V (ccl denotes conjugate class.)V(cclG(K);H)=K,K is weakly closed in H(with respect to G).(cclGFor above definition,very often one is interested in the case that H=P.In that case,it is easy to show that for a p-subgroup O of P,if O is strongiy ciosed in P(with respect to G),then O is weakiy ciosed in P(with respect to G).Extend iemma4and coroiiary5to the foiiowing statement with above notions.Theorem13Let G be a finite group.If n p(G)=2,the following statements are eguiualent.(G);(l)D=0p(D)>N G(P);(2)NG(3)plr(G)=l;(4)D is strongly closed in P with respect to G;第l期王宝山:关于SyIow交的共轭类(5)D is weakly closed in P with respect to G;(G)has a block with defect0;(6)G/0p(7)there exists a normal subgroup,M,of G such that P n M=D.Proof It is needed onIy to show that(2)r(l)and(5)r(l).FirstIy,to show(5)r(l).By the notion of weakIy cIose,then V(ccl G(D);P)=D,in another word,〈D g I D g<P,g G G〉=D. The remainder proof is simiIar to that of Iemma4.Now to show the second part.Suppose converseIy that D;0p(G).By the AIperin’s main resuIt on fusion(see[7,7.2.6]),if D,D x0<P then D and D x0 are conjugate in N G(P).Then D=D x0for aII x0G G such that D x0C P,since N G(D)>N G(P). This is a contradiction and the assertion foIIows.When is the p-subgroup D of the finite group G a defect group for some p-bIock of G?If D is such a subgroup,what are the group theoretic properties of G?Let R be a compIete discrete rank one vaIuation ring with maximaI ideaI(R)=!R such that its residue cIass fieId F=R/!R has characteristic p and its guotient fieId K=guot(R)has characteristic zero.Then the tripIe(K,R,F)is caIIed a p-moduIar system for G.Let A G{K,R,F}.Then AG={】g"g g I"g G A}denotes the group ring of G over A.LetG GG<H and G=G/H.The naturaI K-aIgebra homomorphism is denoted by#$,()x.#H:KG~K G】x"x x~】x"x~Proposition14Suppose that G with n p(G)=2is not a simple group,M is its maximal normal subgroup and D;0p(G).If D is a defect group for some block B,then#M(e B)=0,where e B is the block idempotent of B.Proof By[8,Chapter5,Theorem8.7],if#M(e B);0,then D n M G Syl p(M).If D<M then P n M=D where P G Syl p(G)and D<P,hence that D=0p(G).This is a contradiction with the assumption.Now suppose that D n M<D.By Iemma l0,suppose that G=〈a〉M and a G D.It wiII be shown that P n M>D n M,where P is aIso a SyIow p-subgroup of G containing D.In fact,Iet p G P-D.If p G M,then the proof is done.Now suppose p G M then p=am where m G M.It is obvious that m G D and m=a-l p G P then P n M>D n M.However,this is aIso a contradiction to D n M G Syl p(M).A p-subgroup D of G is caIIed a strong p-subgroup if N G(D)/D has a strongIy p-embedded sub-group(see the definition in[6]).This notion can be shown in the next proposition.Proposition15Let G be a finite group with n p(G)=2,thenD=0p(DC G(D)).Proof If D G Syl p(DC G(D)),there is nothing to prove.Now suppose D G Syl p(DC G(D)),it is obvious that D is a maximaI SyIow intersection,so D is a strong p-subgroup by the remark in[9].In particuIar,N G(D)aIso has the properties above.Denote N G(D)by N for abbreviation.Now N and D satisfy the conditions in[9,Theorem5],then D is a defect group for some p-bIock of N and hence of G by the First Main Theorem of Brauer.[l0,Theorem l]shows D=0p(DC G(D)).References1Robinson G.Locai Structure ,Vertices and Aiperin ’s Conjecture.Proc London Math Soc ,1996,72(3):312-3302Robinson G.On Syiow Intersections in Finite Groups.Proc Amer Math Soc ,1984,90:21-243Brodkey J S.A Note on Finite Groups with an Abeiian Syiow Subgroup.Proc Amer Math Soc ,1983,14:132-1334Zenkov V I ,Mazurov V D.On the Intersection of Syiow Subgroups in Finite Groups.Aigebra and Logic ,1996,35:236-2405Michier G.Moduiar Representation Theory and the Ciassification of Finite Simpie Groups.Proc Sympos Pure Math ,1987,47:223-2326Gorenstein D.Finite Simpie Groups.New York :Pienum ,19827Gorenstein D.Finite Groups.New York /Evanston /London :Harper &Row ,19688Hirosi Nagao ,Yukio Tsushima ,Representation of Finite Group.Boston San Diego New York /Berkeiey London Syd-ney /Tokyo Toronto :ACADEMIC PRESS INC ,19869Zhang Jiping.Studies on Defect Groups.J Aigebra ,1994,166:310-31610Shi S M.Some Resuits on Defect Groups.J Aigebra ,1991,142:233-238关于Sylow 交的共轭类王宝山(北京航空航天大学理学院数学、信息与行为教育部重点实验室(北京航空航天大学、北京大学),北京,100083,E-maii :bwang@ )摘要设G 为有限群,p 是G 阶数的一个素因子,即p I I G I 。
关于共轭极大子群的一个注记
关于共轭极大子群的一个注记史江涛;张翠;孟伟【摘要】研究n-极大子群皆共轭(或同阶)的有限群,给出了2≤n≤4时n-极大子群皆共轭(或同阶)的有限群的完全分类.【期刊名称】《广西师范大学学报(自然科学版)》【年(卷),期】2010(028)001【总页数】3页(P10-12)【关键词】有限群;n-极大子群;共轭【作者】史江涛;张翠;孟伟【作者单位】北京大学数学科学学院数学及其应用教育部重点实验室,北京,100871;University of Primorska,FAMNIT,Glagoljaska 8,6000Koper,Slovenia;University of Primorska,PINT,Muzejski trg 2,6000Koper,Slovenial;云南民族大学数学与计算机科学学院,云南,昆明,650031【正文语种】中文【中图分类】O152.1极大子群在有限群的结构研究中有非常重要的影响,目前已有很多经典结果。
文献[1]例7.3说明有限群G 的所有极大子群均共轭当且仅当G 为素数幂阶循环群。
文献[2-3]证明了恰有2个极大子群共轭类的有限群必可解。
文献[4]给出非可解群G 恰有3个极大子群共轭类当且仅当G≅PSL2(7)或PSL2(2p),其中p 为素数。
文献[5]利用极大子群共轭类型给出全部交错群和部分对称群一个新的刻画。
此外,文献[6-7]对某些特殊有限群的极大子群共轭类进行了研究。
子群同阶类是子群共轭类的推广。
子群同阶类类数是指子群按阶是否相同进行划分所得到的类的个数。
文献[8-9]分别对极大子群同阶类类数≤2的有限群和极大子群同阶类类数=3的非可解群进行了刻画。
文献[10]研究了非正规极大子群同阶类类数=2的非可解群。
设G=G0>G1>G2>…>Gi>…>Gs=1为群G 的一个子群列,对任意i均有Gi+1是Gi 的极大子群,则称Gn 为G 的n-极大子群。
并分别用符号δn(G)和ωn(G)表示群G 的n-极大子群的共轭类类数和同阶类类数。
- 1、下载文档前请自行甄别文档内容的完整性,平台不提供额外的编辑、内容补充、找答案等附加服务。
- 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
- 3、如文档侵犯您的权益,请联系客服反馈,我们会尽快为您处理(人工客服工作时间:9:00-18:30)。
Arch.Math.85(2005)297–3030003–889X/05/040297–07DOI10.1007/s00013-005-1398-7©Birkh¨a user Verlag,Basel,2005Conjugacy classes andfinite p-groupsByEdith Adan-BanteAbstract.Let G be afinite p-group,where p is a prime number,and a∈G.Denote byCl(a)={gag−1|g∈G}the conjugacy class of a in G.Assume that|Cl(a)|=p n.ThenCl(a)Cl(a−1)={xy|x∈Cl(a),y∈Cl(a−1)}is the union of at least n(p−1)+1distinctconjugacy classes of G.1.Introduction.Let G be afinite group.Denote by Cl(a)={gag−1|g∈G} the conjugacy class of a in G,and by|Cl(a)|the size of Cl(a).If the subset X of G is G-invariant,i.e.,X g={x g|x∈X}=X for all g∈G,then X is the union of m distinct conjugacy classes of G,for some integer m.Setη(X)=m.Given any conjugacy classes Cl(a)and Cl(b),we can check that the product Cl(a)Cl(b)= {xy|x∈Cl(a),y∈Cl(b)}is a G-invariant set.In this note,we will explore the rela-tion between|Cl(a)|andη(Cl(a)Cl(a−1)).These results are the equivalent in conjugacy classes as some of the ones in irreducible characters in[1]and[2].In Theorem A of[2],it is proved that if G is a p-group,χis an irreducible character with degree p n,then the productχχofχwith its complex conjugateχis the sum of at least 2n(p−1)+1distinct irreducible characters.The following is the equivalent for conjugacy classes.Theorem A.Let G be afinite p-group and a∈G.Assume that|Cl(a)|=p n.Then the product Cl(a)Cl(a−1)of the conjugacy class of a in G and the conjugacy class of the inverse of a in G,is the union of at least n(p−1)+1distinct conjugacy classes of G,i.e.,η(Cl(a)Cl(a−1)) n(p−1)+1.In Proposition5.3,it is shown that for every prime p and every integer n 0,there exist a p-group G and a conjugacy class Cl(a)of G such that|Cl(a)|=p n andη(Cl(a)Cl(a−1))= n(p−1)+1.Thus the bound in Theorem A is optimal.Mathematics Subject Classification(1991):20D15.298 E.Adan-Bante arch.math. An application of Theorem A is the following:Theorem B.Let n be a positive integer.Then there exists afinite set S n of positive integers such that for any nilpotent group G and any conjugacy class Cl(a)of G with η(Cl(a)Cl(a−1)) n,we have that|Cl(a)|∈S n.In Proposition5.5,we prove that given any prime p,there exist a supersolvable group and a conjugacy class Cl(a)of G with|Cl(a)|=p andη(Cl(a)Cl(a−1))=2.Thus the previous result does not remain true assuming the weaker hypothesis that the groups are supersolvable.Theorem A is the equivalent in conjugacy classes of Theorem B of[1]. Theorem C.Let p be a prime number.Let G be afinite p-group and Cl(a)be a conjugacy class of G.Then one of the following holds:(i)|Cl(a)|=1andη(Cl(a)Cl(a−1))=1.(ii)|Cl(a)|=p andη(Cl(a)Cl(a−1))=p.(iii)|Cl(a)| p2andη(Cl(a)Cl(a−1)) 2p−1.Given afix prime p>2,observe that Theorem C implies that there are“gaps”among the possible values thatη(Cl(a)Cl(a−1))can take for anyfinite p-group and any conjugacy class Cl(a)in G.The previous result is the equivalent in conjugacy classes of Theorem B of[2].2.Proof of Theorem A.N o t a t i o n.Let G be afinite p-group and N be a normal subgroup of G.Denote by a the element in G/N that contains a.Thus Cl(a)is the conjugacy class of a in G/N. Lemma2.1.Let G be afinite p-group and N be a normal subgroup of G.Let a and b be elements of G.Then(i)Cl(a)Cl(b)is a G-invariant set.If Cl(a)∩Cl(b)=∅then Cl(a)∩Cl(b)=∅.Thusη(Cl(a)Cl(b)) η(Cl(a)Cl(b)).(ii)If,in addition,|N|=p,then either|Cl(a)|=|Cl(a)|or|Cl(a)|=|Cl(a)|p.Further-more,if|Cl(a)|=|Cl(a)|p,thenη(Cl(a)Cl(a−1)) η(Cl(a)Cl(a−1))+(p−1). P r o o f.(i)Clearly if a=gbg−1,then a=gb(g)−1.Thus if Cl(a)∩Cl(b)=∅then Cl(a)∩Cl(b)=∅.Thereforeη(Cl(a)Cl(b)) η(Cl(a)Cl(b)).(ii)Since N is normal,|N|=p and G is a p-group,then N is contained in the center Z(G)of G.Thus given any n∈N,Cl(n)={n}.Suppose that|Cl(a)|=|Cl(a)|.Since|N|=p,we have that|Cl(a)| p|Cl(a)|.Therefore|Cl(a)|p |Cl(a)| |Cl(a)|.Thus|Cl(a)|=|Cl(a)|psince G is a p-group and|Cl(a)|divides|G/N|.V ol.85,2005Conjugacy classes and finite p -groups 299If |Cl (a)|=|Cl (a)|p ,then given any x ∈Cl (a)and any n ∈N ,we have that nx ∈Cl (a).Thus n =nx(x −1)∈Cl (a)Cl (a −1)for any n ∈N .Therefore N Cl (a)Cl (a −1)and ii)follows.P r o o f o f T h e o r e m A.We are going to use induction on the order of G .Let N be a normal subgroup of G of order p .Observe such group exists since G is a p -group.Let |Cl (a)|=p m .Since |G/N |<|G |,by induction we have that η(Cl (a)Cl (a −1)) m (p −1)+1.If |Cl (a)|=|Cl (a)|,i.e.,if m =n ,then by Lemma 2.1(i)we have thatη(Cl (a)Cl (a −1)) η(Cl (a)Cl (a −1)) m(p −1)+1=n(p −1)+1.We may assume then that |Cl (a)|=|Cl (a)|.By Lemma 2.1(ii),we have that m =n −1andη(Cl (a)Cl (a −1)) η(Cl (a)Cl (a −1))+(p −1)=(n −1)(p −1)+1+(p −1)=n(p −1)+1.3.Proof of Theorem B.P r o o f o f T h e o r e m B.Let S n = (p i )t i |p i is a prime number for all i,t i 0and t i (p i −1)+1 n.Observe that the set S n is a finite set of positive integers since 0 t i n and if t i >0then p i n .Let {p 1,...,p r }be the set of distinct prime divisors of |G |.For i =1,...,r ,let P i be the Sylow p i -subgroup of G .Observe that a =r i =1a i ,for some a i ∈P i for i =1,...,r .Since G is nilpotent,we have that Cl (a)=r i =1Cl (a i ),where Cl (a i )is the conjugacy class of a i in P i ,for i =1,...,r .Let m i =η(Cl (a i )Cl (a −1i )).Observe that m i n and|Cl (a)|=r i =1|Cl (a i )|.We can check thatη(Cl (a)Cl (a −1))=ri =1η(Cl (a i )Cl (a −1i )).For each i ,let |Cl (a i )|=p t i i .Since Cl (a i )is the conjugacy class of a i in the p i -group P i ,by Theorem A we have that m i =η(Cl (a i )Cl (a −1i )) t i (p i −1)+1.Thusn t i (p i −1)+1.Therefore |Cl (a)|=r i =1|Cl (a i )|=r i =1p t i i ∈S n .300 E.Adan-Bante arch.math.4.Proof of Theorem C.Lemma4.1.Let G be afinite p-group and Cl(a)be a conjugacy class of G with |Cl(a)|=p.Then one of the following holds:(i)Cl(a)={az|z∈Z}for some subgroup Z of the center Z(G)of G.ThereforeCl(a)Cl(a−1)=Z andη(Cl(a)Cl(a−1))=p.(ii)Cl(a)Cl(a−1)is the union of p−1distinct conjugacy classes of size p and the class Cl(e)={e}.Thereforeη(Cl(a)Cl(a−1))=p.P r o o f.Observe that if z∈Cl(a)Cl(a−1)and|Cl(z)|=1,then z is in the center Z(G) of G.Since z=a g a−1for some g∈G and z∈Z(G),z i∈Cl(a)Cl(a−1)and a g i=az i for all integer i.Thus z Cl(a)Cl(a−1).Set Z= z .(i)If z=e,it follows that|Z| p.Since|Cl(a)|=p and a g i=az i for all integer i, we have that Cl(a)={az|z∈Z}and|Z|=p.Since Z is contained in Z(G),then Cl(a)Cl(a−1)=Z andη(Cl(a)Cl(a−1))=p.(ii)We may assume now that if z∈Cl(a)Cl(a−1)and Cl(z)=1,then z=e.Thus all the conjugacy classes different from Cl(e)are of size p.Observe that a g(a−1)g=e for all g∈G.Thus|Cl(a)Cl(a−1)| p2−p+1=(p−1)p+1.Therefore by Theorem A it follows that Cl(a)Cl(a−1)is the union of p−1distinct conjugacy classes of size p and Cl(e).R e m a r k.Let p be a prime number.(a)Let G be an extra special group of order p3and exponent p.We can check that given any a∈G,where a is not in the center of G,then Cl(a)Cl(a−1)=Z(G)and thus Lemma4.1(i)occurs.(b)Let G be the wreath product of a cyclic group C p2of order p2by a cyclic group C p of order p.Thus|G|=p2p+1.Let a=(c,e,...,e)in G,where c∈C p2has order p2. Observe that a−1=a.Observe also thatCl(a)={(c,e,...,e),(e,c,...,e),...,(e,e,...,e,c)}.Thus|Cl(a)|=p.Let b i=(c,e,...,c−1,e,...,e),where c−1is in the i-position for i=0,...,p−1,i.e.,b0=(cc−1,...,e)=(e,e,...,e),b1=(c,c−1,e,...,e)and so for.Observe that Cl(b0)=Cl((e,e,...,e))has class size1.We can check that Cl(b i) has size p for i=1,...,p−1.Since c=c−1,then Cl(b i)∩Cl(b j)=∅if i=jand i,j=0,...,p−1.Observe that Cl(a)Cl(a−1)=p−1∪i=0Cl(b i).Thus Cl(a)Cl(a−1)is the union of a conjugacy class of size1,namely Cl(b0)and p−1distinct conjugacy classes of size p,namely Cl(b i)for i=1,...,p−1.We conclude that given any prime p, there exist some group G and some conjugacy class Cl(a)of G satisfying the condition in case(ii)of Lemma4.1.P r o o f o f T h e o r e m C.If|Cl(a)|={a},then Cl(a)Cl(a−1)={e}and so(i)holds. Lemma4.1implies(ii)and(iii)follows from Theorem A.V ol.85,2005Conjugacy classes andfinite p-groups301 5.Examples.Lemma5.1.Let G0be a p-group and Cl(g0)be the conjugacy class containing g0∈G0. Assume that Cl(g0)=Cl(g−10).Let N=G0×G0×...×G0be the direct product of p-copies of G0.Let C= c be a cyclic group of order p.Observe that C acts on N by c:(n0,n1,...,n p−1)→(n p−1,n0,...,n p−2)(5.2)for any(n0,n1,...,n p−1)∈N.Let G be the semidirect product of N and C,i.e.,G is the wreath product of G0and C. Set a=(g0,e,...,e)in N,where e is the identity of G0.Then|Cl(a)|=p|Cl(g0)|, Cl(a)=Cl(a−1)andη(Cl(a)Cl(a−1))=η(Cl(g0)Cl(g0−1))+(p−1).P r o o f.Observe that Cl(a)=Cl(a−1)since Cl(g0)=Cl(g−10).Let Cl(g0)Cl(g0−1)=C1∪C2∪···∪C m,where C1,...,C m are distinct conjugacy classes of G0.Thus m=η(Cl(g0)Cl(g−10)).We can check that the distinct conjugacy classes of Cl(a)Cl(a−1)are of the following two types:(i){(x,e,...,e)c|x∈C i,c∈C}for i=1,...,m.(ii){(x,y,...,e,e)c|x∈Cl(g0),y∈Cl(g0−1),c∈C},{(x,e,y,...,e)c| x∈Cl(g0),y∈Cl(g0−1),c∈C}and{(x,e,...,e,y)c|x∈Cl(g0),y∈Cl(g0−1), c∈C}.Observe that there areη(Cl(g0)Cl(g0−1))distinct conjugacy classes of type(i)and exactly p−1distinct conjugacy classes of type(ii).Thusη(Cl(a)Cl(a−1))=η(Cl(g0) Cl(g0−1))+(p−1).Proposition5.3.Given any prime p,and any integer n 0,there exist afinite p-group G and a conjugacy class Cl(a)of G with|Cl(a)|=p n,Cl(a)=Cl(a−1) andη(Cl(a)Cl(a−1))=n(p−1)+1.P r o o f.Observe that if G is an abelian group and a∈G has order p2,then|Cl(a)|=1, Cl(a)=Cl(a−1)andη(Cl(a)Cl(a−1))=1=0(p−1)+1.Thus the statement is true for n=0.Assume by induction that the statement is true for n−1,i.e.,there exist afinitep-group G0and a conjugacy class Cl(g0)of G0with|Cl(g0)|=p n−1,Cl(g0)=Cl(g−10)andη(Cl(g0)Cl(g0−1))=(n−1)(p−1)+ing the notation of Lemma5.1,we have thatη(Cl(a)Cl(a−1))=η(Cl(g0)Cl(g0−1))+(p−1)=(n−1)(p−1)+1+(p−1)=n(p−1)+1. Since|Cl(a)|=p|Cl(g0)|=p×p n−1=p n,the proof is complete.H y p o t h e s i s5.4.Fix a prime p and let F={0,1,...,p−1}be thefinitefield with p elements.Observe that F is also a vector space of dimension1over itself.Let A=Aff(F)302 E.Adan-Bante arch.math. be the affine group of F.Observe that the group A is a cyclic by cyclic group and thus it is supersolvable.Let C be a cyclic group of order p.Set X=F and K=C X={f:X→C}.Observe that K is a group via pointwise multiplication,and clearly A acts on this group(via its action on X).Let G be the wreath product of C and A relative to X,i.e.,G=K◦A.We can check that G is a supersolvable group.Proposition5.4.Assume Hypotheses5.4.Set a=(c,e,e,...,e)in K.Then a∈G, the conjugacy class Cl(a)of G has size p and Cl(a)Cl(a−1)=Cl((e,e,...,e))∪Cl((c,c−1,e,...,e)).Thusη(Cl(a)Cl(a−1))=2.Therefore,given any prime p,there exist a supersolvable group G and a conjugacy class Cl(a)of G with|Cl(a)|=p andη(Cl(a)Cl(a−1))=2.P r o o f.Observe that Cl(a)={(c,e,...,e),(e,c,e,...,e),...,(e,e,...,e,c)}.Thus Cl(a)has p-elements.Observe that{(c,c−1,...,e)y|y∈F\{0}}={(c,c−1,e,...,e),(c,e,c−1,...,e),...,(c,e,...,e,c−1)}. Observe also that{(c,c−1,e,...,e)x|x∈F}={(c,c−1,e,...,e),...,(e,e,...,c,c−1),(c−1,e,...,e,c)}.ThusCl((c,c−1,e,...,e))={(c,c−1,e,...,e)x,(c,e,c−1,...,e)x,...,(c,e,...,c−1,e),(c,e,...,e,c−1)x|x∈F}has(p−1)p=p2−p elements.Since a g(a−1)g=(e,...,e),then Cl(a)Cl(a−1)has at most p2−p+1elements.We conclude thatCl(a)Cl(a−1)=Cl((e,e,...,e))∪Cl((c,c−1,e,...,e)).A c k n o w l e d g e m e n t.I would like to thank Manoj Kumar Yadav for bringing to my attention products of conjugacy classes.I also want to thank Professor Everett C.Dade for useful advise and corrections.V ol.85,2005Conjugacy classes andfinite p-groups303References[1] E.Adan-Bante,Products of characters andfinite p-groups.J.Algebra277(1),236–255(2004).[2] E.Adan-Bante,Products of characters andfinite p-groups II.Arch.Math.82,289–297(2004).Received:16December2004Edith Adan-BanteUniversity of Southern Mississippi Gulf Coast730East Beach BoulevardLong Beach MS39560USAedith.bante@。