基于Agent的计算金融学建模与仿真方法研究

VIP免费
3.0 牛悦 2024-11-19 4 4 1.54MB 134 页 15积分
侵权投诉
复杂系统与复杂性科学是 21 世纪的科学,其研究方法是近代以来科学方法论
的又一场革命,将为人们提供全新的了解自然界奥秘的手段。当前,对于自然
象、社会、政治、军事、管理、生物、经济、特别是金融等领域复杂适应系统与复杂性
的研究,呼唤新的建模与仿真方法的出现,基Agent 的建模与仿真方法应运
生。由于金融市场是一个复杂适应系统,所以基于 Agent 计算金融学建模与仿
真方法学作为当前研究金融市场复杂适应系统的科学方法,成为当前建模与仿
领域的研究热点。
本论文以复杂适应系统理论为立论基础与研究方法,以金融市场的建模与仿
真为立题背景,进行了基于 Agent 计算金融学建模与仿真方法普适理论的相
研究,将金融市场复杂适应系统看作由系统中Agent 制与环境中的资源
制、及其它 Agent 机制不断地进行相互作用,并以构建动态机制的形式相互聚集、
层层涌现出的动态机制。并将这个普适理论应用到股票市场的建模与仿真中,
期使基于 Agent 计算金融学建模与仿真方法学成为一套完善的建模与仿真理论 ,
并能够为指导具体金融市场复杂适应系统的建模与仿真研究提供一定帮助。
本论文结合前人的研究成果,并从复杂适应系统理论的基本内容与观出发
对金应系性进行了分在详了 基 Agent
的计算金融学建模与仿真方法的本思想和主要特点的基础上,提出了金融市场
复杂适应系统建模与仿真的框架和研究思路,并因此界定了基于 Agent 算金
学建仿的研在此框架下过 引机制来刻 Agent
不断进行适应与学习进化的动态属性,将 Agent 在不断的适应与进化过程中涌现
出的功能结构模式看作是动态的 Agent Agent 机制又被看作是金融市场
复杂适应系统复杂现象涌现的构造模式。给出了基Agent 算金融学建模
仿真方法学意义下,包括 Agent 机制、信号、转换函数、资源机制以及基于 Agent
的计算金融学模型的概念及形式化描述。建立了规范基于 Agent 算金融学
模与仿真方法的流程,这对于规金融市场复杂适应系统建模与仿真,减少
模与仿真的复杂度以及提高模型的重用性与可用性均具有重要的意义。
最后,论文还将基于 Agent 的计算金融学建模与仿真方法应用到股票市场的
建模与仿真研究中,不仅详细描述了基于 Agent 股票市场建模与仿真的基本
置、具体过程、及相应的实现算法,还利用面向对象的思想,采用 Java 语言实现了
一个基于 Agent 的股票市场仿真模型的原型系统。
本论文的主要创新点是:
1.在深入研究和分析了基于 Agent 的计算金融学建模与仿真方法的基础上,通过
I
Agent 性 ,
Agent 在不断的适应与进化过程中涌现出的功能结构模式看作是动态的 Agent
制,而 Agent 机制又被看作是金融市场复杂适应系统复杂现象涌现的构造模式;
2.于机Agent ;在引入概念础上文特别对
Agent 计算金融学模型中的 Agent 机制进行了深入的研究,提出了基于机制的
Agent 模型结构框架,将 Agent 看作为由感知基本机制、操作基本机制、决策基本
机制、适应控制基本机制 4大基本机制相互作用而构成的动态机制。并分别对各基
本机制的结构及其相应的算法进行了详细地阐述;
3.引入了资源机制的概念;将资源机制看作是一种特殊的 Agent 机制,是由感知
基本机制、操作基本机制、决策基本机制 3大基本机制相互作用而动态构建的能处
理所有与资源有关操作的一种功能结构模式。并分别对各基本机制的结构及其
应的算法进行了详细地阐述;从而,将金融市场复杂适应系统完全统一到机制
造的动态网络之中;
4.将金融市场复杂适应系统看作是由 Agent 机制与环境中的资源机制进行相互作
用,并以构建动态机制的形式相互聚集、层层涌现出的动态机制。同时,特别强调
了信号在金融市场复杂适应系统的机动态涌现过程的作用;从而使金融市
复杂适应系统模型成为了一个动态的受生成过程[23]
5.Agent 仿Agent
Agent 计算金融学模型的生命期概念。其生命个方面的模型:首先
在系统分段由家参与建立的念模型;是在系统建模由建
专家建立的设计模型;最后是在系统仿真由计算机家参建立的计
模型。
关键词:机制 计算金融学 基于 Agent 的建模与仿真方法
Agent 机制 资源机制 信号 转换函数。
ABSTRACT
II
The theories of complex systems and complexity are the sciences of 21 century,
whose research method is another revolution for scientific methodology since modern
times, and will provide people with new solutions to uncover the miracles of the nature.
The researches needed for complex systems and complexity in natural phenomena,
society, polity, military, management, biology,economy, especially finance impose a
great challenge to the current theories and methods of modeling and simulation.While
the novel Agent-Based Modeling and Simulation methodology, which is the hot topic
today in modeling and simulation field, can fulfill the requirements.
Based on the theories of complex systems and complexity, the paper has conducted
research on Agent-Based Computational Finance Modeling and Simulation general
methodology and problems relevant to the modeling and simulation of finance market,
which are also complex systems. In the approach of Agent-Based Computational
Finance Modeling and Simulation, it considers finance market complex adaptive system
as emergent dynamic mechanism in which agents mechanism of system interact
ceaselessly with resource mechanism of environment and else agents mechanism, one
another assemble and level upon levels construct. And this general methodology is
applied to modeling and simulation of stock market,the purpose is to make Agent-Based
Computational Finance Modeling and Simulation become a perfect theory of modeling
and simulation, which can instruct the research for finance market complex adaptive
systems.
This thesis summarize research harvest of complex adaptive systems, and from the
basic views of complex adaptive systems theory, complexity of finance market complex
adaptive systems was also was analyzed and discussed. This thesis expatiate on the
basic idea and the primary characteristic of Agent-Based Computational Finance
Modeling and Simulation approach. A research approach for finance market complex
adaptive systems modeling and simulation was proposed. And a novel framework for
Agent-Based Computational Finance Modeling and Simulation was proposed
creatively,which bounded the research contents of Agent-Based Computational Finance
Modeling and Simulation impliedly.In this framework, concept of agent was extend
from adaption and evolution of agent. Concept of agent mechanism was import.And it
considers mechanism as emergent pattern of finance market complex adaptive
systems.Thereafter, concept and formal specifications of the agent mechanism, the
III
signal, the transition function, the resource mechanism and the Agent-Based
Computational Finance Model was defined and represented formally according to
Agent-Based Computational Finance Modeling and Simulation framework.And it
establish canonical flow of Agent-Based Computational Finance Modeling and
Simulation approach for directing modeling and simulation of finance market complex
adaptive systems. Consequently, complication was reduced on modeling and
simulation,and usability of model was improved.
Finally, approachv of Agent-Based Computational Finance Modeling and
Simulation is applied to modeling and simulation of stock market. Not only the basic
setup,the process and relevant arithmetic of agent-based stock market modeling and
simulation are detailedly described ,but also a prototype of agent-based stock market is
designed with java language base on idea of Object Oriented.
Innovation of This thesis is that:
1. This thesis depict dynamic attribute of agents which are evolving and adapting
ceaselessly through importing concept of mechanism.And a novel framework for
Agent-Based Computational Finance Modeling and Simulation was proposed
creatively. It considers emergent structure pattern of agent as dynamic agent
mechanisms when agents are evolving and adapting ceaselessly. And it considers
agent mechanism as emergent pattern of finance market complex adaptive
systems.
2. Author study the agent mechanism in Agent-Based Computational Finance
Model, and put forward framework of mechanism based agent model. It considers
the agent as evolutive mechanism which was set through four essence mechanism,
namely perception mechanism, action mechanism, decision-making mechanism,
adaption control mechanism, interacting. And all the structure and the arithmetic
of the essence mechanism was described.
3. Conception of resource mechanism is proposed, it is considered as a special
agent mechanism. So finance market complex adaptive systems is a dynamic
mechanisms that is constructed by agent mechanisms. And resource mechanism is
constructed by three essence mechanism, namely perception mechanism, action
mechanism, decision-making mechanism, interacting. It can disposal all
manipulation correlative with resource. All the structure and the arithmetic of the
essence mechanism was described in resource mechanism at the same time.
IV
4. It considers finance market complex adaptive system as emergent dynamic
mechanism networks in which agents mechanism of system interact ceaselessly
with resource mechanism of environment and else agents mechanism, one another
assemble and level upon levels construct. Especially, it emphasize effect of signal
in dynamic emergence of finance market complex adaptive systems. Therefore,
finance market complex adaptive system like as a dynamic Constrained generating
procedures.
5. It put forward lifecycle conception of Agent and Agent-Based Computational
Finance Model. Lifecycle come down to three process model: notional model,
design model, computing model.
Key words: Mechanism, Computational Finance, Agent Based
Modeling and Simulation, Agent Mechanism, Resource
Mechanism, signal, Transition Function.
V
中文
ABSTRACT
引言......................................................................................................................1
§1.1 研究背景...............................................................................................................1
§1.2 研究意义...............................................................................................................4
§1.3 基于 Agent 计算金融学建模与仿真方法的研究现.......................................5
§1.3.1 基于 Agent 计算金融学建模与仿真方法学的研究....................................5
§1.3.2 基于 Agent 建模方法的金融市场模型研究................................................7
§1.4 本论文的主要研究内容.....................................................................................12
§1.5 本论文的主要创新点.........................................................................................14
第二章 基于 Agent 计算金融学建模与仿真的相关研究领域概述............................16
§2.1 复杂性科学与复杂适应系统理论.....................................................................16
§2.1.1 复杂性科学的......................................................................................16
§2.1.2 复杂适应系统理论的基本内容与观点......................................................20
§2.1.3 复杂适应系统理论的主要特点..................................................................22
§2.1.4 复杂适应系统的基本概念..........................................................................24
§2.1.5 复杂适应系统的基本特..........................................................................26
§2.2 Agent 理论..........................................................................................................28
§2.2.1 Agent 的概念...............................................................................................28
§2.2.2 Agent 的研究范.......................................................................................29
§2.2.3 Agent 技术...................................................................................................30
§2.3 基于 Agent 的建模与仿真方法.........................................................................32
§2.3.1 基于 Agent 建模与仿真方法概述..............................................................32
§2.3.2 基于 Agent 建模与仿真的源与发......................................................33
§2.3.3 基于 Agent 的建模与仿真方法的基本内容..............................................33
§2.3.4 基于 Agent 的建模与仿真的研究现......................................................34
§2.3.5 基于 Agent 建模与仿真概念下 Agent 与对象(Object)比较.................37
第三章 金融市场复杂适应系统的建模与仿真分析....................................................39
§3.1 金融市场是一个复杂适应系统.........................................................................39
§3.2 金融市场复杂性的.....................................................................................40
§3.3 统的金融市场建模方法的不.....................................................................42
§3.4 基于 Agent 的计算金融学.................................................................................43
VI
第四章 基于 Agent 的计算金融学建模与仿真方法....................................................46
§4.1 基于 Agent 的计算金融学建模与仿真概述............................................46
§4.1.1 主要概念与思想..........................................................................................46
§4.1.2 主要特点......................................................................................................47
§4.2 基于 Agent 计算金融学建模与仿真框架.........................................................48
§4.2.1 系统建模与仿真的框架......................................................................48
§4.2.2 基于 Agent 计算金融学建模与仿真框架..................................................48
§4.3 基于 Agent 计算金融学模型的校核确认.....................................................50
§4.3.1 模型的校核确认......................................................................................50
§4.3.2 涌现现象......................................................................................................51
§4.4 基于 Agent 的计算金融学模型的建模要.....................................................53
§4.5 基于 Agent 计算金融学模型的形式化描述.....................................................53
§4.5.1 有关机制的概念..........................................................................................54
§4.5.2 信号的形式化描述......................................................................................55
§4.5.3 机制的形式化描述......................................................................................55
§4.5.4 Agent 机制的形式化描述...........................................................................56
§4.5.5 资源机制的形式化描述..............................................................................56
§4.5.6 环境的形式化描述......................................................................................57
§4.5.7 基于 Agent 的计算金融学模型的形式化描述..........................................57
§4.6 基于机制技术的特点.....................................................................................6058
§4.7 基于 Agent 计算金融学建模与仿真研究步骤.................................................58
§4.7.1 定义建模与仿真目标................................................................................589
§4.7.2 目标系统复杂性特分析及概念模型的建立..........................................60
§4.7.3 合理选择抽象层..................................................................................5960
§4.7.4 信号流分析..............................................................................................5960
§4.7.5 建立设计 Agent 模型..............................................................................5960
§4.7.6 建立基于 Agen 计算金融学的设计模型...............................................5960
§4.7.7 进行基于 Agent 机制的仿真....................................................................601
§4.7.8 模型的动态校核确认............................................................................601
§4.7.9 涌现结果的分析........................................................................................601
§4.7.10 断是否符合要..................................................................................601
§4.7.11 模型修正..................................................................................................601
第五章 基于机制的 Agent 建模....................................................................................62
VII
§5.1 Agent 模型研究现..........................................................................................62
§5.2 基于机制的 Agent 建模.....................................................................................63
§5.2.1 基于机制的 Agent 建模框架......................................................................63
§5.2.2 基本机制的系统结构..................................................................................64
§5.3 感知基本机制.....................................................................................................65
§5.4 操作基本机制.....................................................................................................68
§5.5 决策基本机制.....................................................................................................70
§5.6 适应控制基本机制.............................................................................................72
§5.7 基于机制 Agent 建模的关键技术.....................................................................74
§5.7.1 转换函数的动态构建算法..........................................................................74
§5.7.2 的信用分算法..................................................................................75
§5.7.3 基本机制的适应性进化算法......................................................................76
§5.7.4 感系数的设置..........................................................................................79
§5.7.5 Agent 机制中的信号传递...........................................................................80
§5.7.6 适应控制基本机制的适应控制算法..........................................................81
§5.7.7 Agent 中其它基本机制的适应控制算法...................................................82
第六章 环境模型............................................................................................................83
§6.1 资源机制.............................................................................................................83
§6.1.1 基本机制的系统结构..................................................................................84
§6.1.2 交易产品交易处理机制..............................................................................84
§6.1.3 感知算法......................................................................................................86
§6.1.4 编码算法......................................................................................................86
§6.1.5 交易匹配与计算算法..................................................................................87
§6.2 金融市场时模型.............................................................................................87
§6.3 息共享机制建模的探讨.................................................................................89
第七章 基于 Agent 计算金融学建模与仿真方法的应用............................................92
——股票市场的建模与仿真
§7.1 研究背景.............................................................................................................92
§7.2 建模与仿真的目标.............................................................................................93
§7.3 系统分析.............................................................................................................93
§7.4 Agent 机制模型..................................................................................................94
§7.4.1 感知基本机制..............................................................................................94
§7.4.2 操作基本机制..............................................................................................97
VIII
§7.4.3 决策基本机制..............................................................................................99
§7.4.4 适应控制基本机制..................................................................................1001
§7.5 资源机制...........................................................................................................101
§7.5.1 感知基本机制..........................................................................................1012
§7.5.2 决策基本机制..........................................................................................1023
§7.5.3 操作基本机制..........................................................................................1034
§7.6 仿真语言的选择.............................................................................................1034
§7.6.1 Java 对基于 Agent 股票市场仿真模型的支持......................................1045
§7.6.2 Java 性能实际测试....................................................................................105
§7.7 发环境...........................................................................................................106
§7.8 系统设计原.................................................................................................1067
§7.9 主要机制的设计.........................................................................................1078
§7.9.1 机制基的设计......................................................................................1078
§7.9.2 Agent 机制的设计.................................................................................108
§7.9.3 资源机制的设计....................................................................................110
§7.9.4 基于 Agent 的股票市场仿真模型设计框架..........................................1101
§7.10 仿真结果分析...............................................................................................1112
第八章 结与展望......................................................................................................117
§8.1 ...........................................................................................................117
§8.2 研究展望...........................................................................................................118
参考.....................................................................................................................1210
IX
引言
第一章 引言
§1.1 研究背景
来,人们对金融市场认识和研究的不断深入,发现标准金融学理论
以解现实金融市场中在的大量异常现象,如小公司效应、应、年末效
等,因而对金融市场的研究入了境之中。标准金融学:
1.是理性的。们对获取的信计是无偏的,出的资决策是最
,来经济的预测公正的。
2.是同的。间没别,具有同的理性。
3.风险厌恶型的。它是在定的收益选择风险资决策。
4.对不同资风险态度是一的。
5.市场有。市场价格充应了所有信,市场价格即为理性价格
标准金融学的这观点使它对现实中在的理性行为从下
作 为 现 代 标 准 金 融 理 论 支 柱 之 一 的 市 场 有 假 说 (Efficient Markets
HypothesisEMH )为,不受扭曲地全证券价格到了反映
么就市场是有。此假说的一个最主要市场
的信对于将来预测毫无意义,市乎没有原。然而,场中
模式率波Volatility ClusteringFat
Tails记忆(long memory等,表明市场着某种内在的动系统。
金融学的基本观点,受到怀疑EMH 好像
金融市场是一个复杂的经济系统,它资主体和
学、资策市场期等内观环境因政策及市场因
息获杂的关系决定了金融市场运行系统的运行
直接系统外部素则通过用于金融市场
系统,同会引系统管是还是外部不是
影响金融市场的运行,而是相互渗透相互影响的。金融市场中的资主体是具有
身目的与主动性的、积极“活 个体。它并不象当金融研究领域中占据着
统治统金融理为的具有完全的理性,用原进行
的完全理性个体,而是能够与环境以及其体进互作用,从而进行
性行为的个体,这的个体能够不断的互过程中,不断地
“积验”,并根据学到的经改变的结构和行为方式。当环境发生
化的可以从而使能够尽量谋得
观的化和个体分化可以从这些投资主体的行为规源。
1
摘要:

摘要复杂系统与复杂性科学是21世纪的科学,其研究方法是近代以来科学方法论的又一场革命,将为人们提供全新的了解自然界奥秘的手段。当前,对于自然现象、社会、政治、军事、管理、生物、经济、特别是金融等领域复杂适应系统与复杂性的研究,呼唤新的建模与仿真方法的出现,基于Agent的建模与仿真方法应运而生。由于金融市场是一个复杂适应系统,所以基于Agent的计算金融学建模与仿真方法学作为当前研究金融市场复杂适应系统的科学方法,成为当前建模与仿真领域的研究热点。本论文以复杂适应系统理论为立论基础与研究方法,以金融市场的建模与仿真为立题背景,进行了基于Agent的计算金融学建模与仿真方法普适理论的相关研究,将...

展开>> 收起<<
基于Agent的计算金融学建模与仿真方法研究.doc

共134页,预览10页

还剩页未读, 继续阅读

作者:牛悦 分类:高等教育资料 价格:15积分 属性:134 页 大小:1.54MB 格式:DOC 时间:2024-11-19

开通VIP享超值会员特权

  • 多端同步记录
  • 高速下载文档
  • 免费文档工具
  • 分享文档赚钱
  • 每日登录抽奖
  • 优质衍生服务
/ 134
客服
关注