The Design Structure SystemA Method for Managing the DesignofComplexSystems

IEEE TRANSACTIONS ON ENGINEERING MANAGEMENT, VOL. EM-28, NO. 3, AUGUST 1 9 8 1
71
The Design Structure System: A Method for Managing the Design of Complex Systems
ห้องสมุดไป่ตู้
once the techniques discussed here have been used to resolve how the circuits will be iterated. This paper shows how the techniques of the Design Struc­ ture System can be applied to organize the design of a system, to develop an effective engineering plan, to show where esti­ mates are required, and to analyze the flow of information that occurs during the design work. This information flow can be used to determine the consequences of a change in any variable on the rest of the variables in the system, and thus to determine which engineers must be informed and which documents must be changed. This can be used as the basis for scheduling the work arising from the change. HOW WE USE THE DESIGN STRUCTURE SYSTEM The variables which define the design of the product are first listed. For each variable we list the other variables which must be known or assumed in order that this variable be determined. These other variables are called predecessors. For example, in Table I we show a precedence table which might represent the design of an electric car. Table I says that in order to determine variable 4 , valables 1, 2, 3 , 7, 11, and 12 are required, and thus must first be known or estimated. These precedences do not have to be based on mathematical equa­ tions. They could represent qualitative judgments by manage­ ment or engineers given the necessary predecessor informa­ tion, e.g., the choice of a lead- or sulfur-type battery based upon an evaluation of the technical risk. Wherever we refer to "variables*' we could also refer to the engineering tasks to determine and document these variables.
Manuscript received January 28, 1980; revised March 9, 1981. The author is with the Computer Science Department, California State University, Sacramento, CA 95819.
INTRODUCTION "[ENGINEERING design involves the specification of many Mid variables which together define a product, how it is made, and how it behaves. Before some variables can be determined, other variables must first be known or assumed. This implies a precedence order of the variables, and consequently of the tasks of determining these variables. In engineering design this ordering will contain circuits, i.e., A cannot be determined unless Β is first known or assumed, but Β cannot be deter­ mined unless A is first known or assumed. For example, the size of the heat exchanger depends upon the coolant tempera­ ture, but the coolant temperature depends upon the size of the heat exchanger. Circuits are usually handled by making estimates for some of the variables to make a preliminary design, then using the results of the preliminary design to confirm or refine these estimates. This is the process of design iteration. The engineer must resolve the simultaneous relations inexpensively with pencil and paper or computer models so as to avoid the more expensive trial and error process of build­ ing and rebuilding the actual system. Critical path planning and control techniques such as the Program Evaluation and Review Technique (PERT) do not handle the circuits inherent in engineering design. However, critical path can be used to schedule the building of the system
TABLE I PRECEDENCE TABLE Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Description η Passenger capacity specification Size and aerodynamics Motor specifications and weight Total weight Stored energy requirement Battery type and energy density Battery size and weight Cruising speed specification Speed and acceleration performance versus power Acceleration specification Speed and acceleration conformance Structural and suspension design Range specification Cost Consumer demand versus cost Profit Predecessors None 1, 3, 7, i Î, 12 2 , 4 , 6 , 7, 11 1 , 2 , 3 , 7 , 11,12 3 , 6 , 8 , 9, 1 0 , 1 3 None 5,6, None 1 , 2 , 4 , 12 None 8, 9, 10 1 , 2 , 3 , 4 , 7, 11 None 2 , 3 , 4 , 6 , 7 , 12 1, 2, 8, 10, 12, 13 14, 15
D O N A L D V. S T E W A R D ,
MEMBER, IEEE
Abstract—Systems design involves the determination of interde­ pendent variables. Thus the precedence ordering for the tasks of determining these variables involves circuits. Circuits require planning decisions about how tc iterate and where to use estimates. Conventional planning techniques, such as critical path, do not deal with these problems. Techniques are shown in this paper which acknowledge these circuits in the design of systems. These techniques can be used to develop an effective engineering plan, showing where estimates are to be used, how design iterations and reviews are handled, and how information flows during the design work. This information flow can be used to determine the consequences of a change in any variable on the rest of the variables in the system, and thus which engineers must be informed and which documents must be changed. From this, a critical path schedule can be developed for implementing the change. This method is ideally suited to an Automated Design Office where data, computer input and output, and communications are all handled through the use of computer ter­ minals and data bases. However, these same techniques can also be effectively used in classical engineering environments.