基于层次化模型的多智能体系统设计方法

在开展系统设计时,系统的内在结构和行为特性使其自身呈现出显著的不确定性,进而造成系统控制难度的提高.在此条件下,当系统面临结构、行为或功能的变动时,会导致其在特定场景下产生设计期望之外的行为表现,引发系统局部甚至是全局失能,进而加剧了系统控制的复杂度.出于上述考虑,基于提出的用于分析系统引入变动产生响应的模型框架,针对产品开发项目过程开展层次化的系统建模,围绕用以表征系统行为特性的直接参数开展对系统引入变动后产生反馈效应的分析,提供对系统设计过程管理的改进.在此基础上,提出基于智能体技术的计算机辅助分析技术方案,为系统设计过程中的变动因素识别及效应追踪提供技术基础.

Hierarchical Model-based Methodology of Multi-Agent System Design

In a product development project,the occurred changes would bring opportunities as well as challenges simul-taneously to companies. The former could encourage the company to modify or improve its product design and the oth-er management aspects in order to keep its competitiveness in market, while the latter would lead the company into some potential dilemma in which the brought in changes could not be handled properly under various constraints and then propagates into some other worse consequences. Regarding the issues, we propose, in this paper, a hierarchical methodology of iteratively modeling product development projects. Relying on the suggested conceptual model of identify-ing change occurrence and propagation,we are then enabled to consider analyzing change occurrence and change propa-gation through investigating the characteristics of direct parameters specifying a product development project so as to in-spire discovering the potential change propagation scenarios. With the established hierarchical model of product develop-ment project and the identified characteristics of change propagations, we suggest an agent-based technical solution to assistant analyzing change occurrence that could elicit identifying and tracking change propagations. In this way,the de-sign of the multi-agent system prototype,as the theoretical and practical contribution to companies,could be employed to build up computer-aided decision making tool for better change management.