BEA.自动算法交易

合集下载
  1. 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. 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. 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)。

Brokered Events
Event Source
Event Event
Event Sink
Message Broker
Message
Event Source

Event
Event
Event Sink
Event routed from source to one or more sink Optionally stored persistently until delivered Various ack modes guarantee delivery, e.g. auto, client, duplicates O.K.
<<Source>> Rule Data Source Adapter Stream
Event-Driven SOA
Service Bus
SOA Services
WLS
createOrder()
Interactive Response
通过复杂的数学模型在金融市场上寻找最佳价格交易的系统, 通常在找到合适买点后电脑会按照提前设定的指令自动下单交 易 扩展:目前算法交易已经在电子交易系统中全面应用,特别是 外汇市场、期货市场,包含下单分割、批量交易、统计套利等 等
算法交易的应用领域
2007年4月,纽约证券交易所和欧洲证券交易所正式合并后, 电子交易已经占到交易量的60%~70%,其中算法交易的比例 逼近半数。 据估计,到2008年,北美所有市场上的全部证券交易中超过 40%使用算法交易进行交易 美国90%以上的证券经理在建立投资组合时至少使用一次算法 交易 东京证券交易所、香港联交所、新加坡交易所已经成为亚洲地 区采用算法交易的主要市场
Send to One Interested Party – Multiple Receivers
Sender
Message Broker
Receiver
Event Sink Event Sink
Event Source
Queue

Event delivered to one receiver
Event Sink
Filter
rules of events
Aggregation Event
detection patterns
Application Model
<<OSGI Service>> Service1 <<OSGI Service>> Service2 OSGI Services can be imported and exported
Synchronous Bi-directional Communication
Please do this for me. I’ll wait until I get the reply.
Request
Consumer Provider
Reply

Physical connection couples request/reply Failure of connection terminates execution
算法交易的广泛应用对金融机构的要求
必须能够在极短的时间(毫秒单位)完成下单交易
能够执行复杂的交易策略和逻辑
能够随时更换不同的交易策略
议程
金融交易的发展动向 - 算法自动交易
算法自动交易的实现 – 用事件方式取代请求方式
BEA的复杂事件服务器
Invocation Models
Request/Reply
Multiple receivers promotes scalability All clients are created equal
Brokered Events – Publish/Subscribe
Send to All Interested Parties
Publisher
Message Broker Event Source
Event Flow
Fork
Inventory Mgmt Service
Join
Online Ordering Application
Order Mgmt Service
Shipping and Fulfilling
Credit Services
Request/Reply vs. Events
Request/Reply
Combined
with brokered event mechanisms
routing
Content-based Transformation
Standards-based;
e.g. XML and Web Services
Enterprise Service Bus - Orchestration Service
Request/Reply Flow
Online Ordering Application
Supplier Service Inventory Mgmt Service
Order Mgmt Service
Credit Services
Event
Asynchronous uni-directional communication

Event Sink Provider
Req-id couples request/reply Reply sent independently of request Consumer assumes reply may be received along time after request
Sophisticated Process Flow
Event Source
Event
ESB
Event
Event Sink
Event Source

Event
Event
Event Sink
Combines basic ESB services with long running stateful processes Supports looping constructs, state management, conditional constructs, fork thread, join, etc. Publish event, wait for event, etc.
Subscribers
Event Sink
Topic
Event Sink subscribe
Event
broadcast to all interested parties, e.g. “Hot Deals” can be added dynamically
Subscribers Durable
subscriptions support guaranteed delivery
Enterprise Service Bus – Routed Events
Intelligent Routing
ESB
Yes
Event Source order count < 100 No Event Sink Event Sink
Path of Execution
Reactivity
Events-Driven Application Models
Asynchronous One-way Communication
Simple Events
Brokered Events
Enterprise Service Busses Event Stream Processing Engines
从某银行的理财产品看算法交易在中国的 应用前景
该业务的操作算法
交易算法


通过计算预期收益来 触发开仓操作
通过对套利空间的判 断来触发平仓操作
交易时机

使用交易软件定时计 算并监控股指期货与 股指现货的价格比
当市场上出现数千种类似的理财产品时
必须实时对市场行情做出反应,比对手更快你才能以合理的价 格获得利润 随时调整交易算法,捕获更多的套利空间
当你作出决策后,能够在多短时间内完成交易?
您能够比对手更快一步么? 利用算法自动完成交易?
金融交易的发展(二) 算法的复杂多变
完成大规模交易而不惊动市场

一秒钟内产生数千个交易指令,其中许多指令瞬间就可以被取消或被新的指令取 代,从而把大额委托化整为零,减小对市场的冲击,并且可以寻求最佳的成交执 行路径,减少交易成本


Request - a request to do something
Reply - an indication that the request was processed
Event

an indication that something occurred in the past
Request/Reply
实时发现交易机会,分散市场风险
金融交易的发展(二) 算法的复杂多变
您的IT系统是如何帮助您从瞬息万变的市场中把握机会?
当客户的投资组合、盈利目标,或者市场的风险指标发生变化 时,您的IT系统能否立即应用新的规则?
当您的对手这样做时,您能否及时发现?
什么是算法交易 (Algorithmic Trading)?
Did it. Sorry it took me so long!!!
Consumer assumes reply received shortly after request sent
Single provider receives request implies limited scalability
Export
Import
OSGI Bundle/Spring Application Context
<<Source>> Data Source Adapter Event Processing Network (EPN)
Stream
Processor Rule
<<Sink>> Stream Listener

Event Stream Processing (ESP)
Aggregating and Correlating Events
IF PRGS Price < PRGS VWAP THEN Buy PRGS
Event Manager Event Rules
Events
Event Source
Event Stream
Component to Component Coupling Flow of Control
Events Many to many Asynchronous Supports dynamic and parallel Dynamically reactive
1 to 1 Synchronous Linear and hierarchical Closed loop
Simple Event
Event
Event Source Evente Events – Simulating Request/Reply
Request Event (req-id) Event Source Consumer Reply Event (req-id)
Brokered Events – Point to Point
Send to One Interested Party
Sender
Message Broker
Receiver
Event Source
Queue
Event Sink
connect
Brokered Events – Point to Point
It happened
Event
Event Source Event Sink

Each event sent independently of another Coupling between events handling by source and sink
Multiple sources may send the same event and multiple sinks may receive same event so highly scaleable
Algorithmic Trading Overview
Island Chen Senior Engineer BEA
BEA Confidential. | 1
议程
金融交易的发展动向 - 算法自动交易
算法自动交易的实现
BEA的复杂事件服务器
金融交易的发展(一) 交易实时性
当市场信息发生变化后,您能在多短时间内作出决策?
相关文档
最新文档