复杂体系过程A-GERT网络Bayes学习机制解析与模型设计  

Bayes Learning Mechanism Analysis and Model Design of Complex System of Systems Process A-GERT Network

作　　者：方志耕[1] 陈静邑 张靖如 夏悦馨 熊仪 华晨晨 FANG Zhi-geng;CHEN Jing-yi;ZHANG Jing-ru;XIA Yue-xin;XIONG Yi;HUA Chen-chen(College of Economics and Management,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China)

机构地区：[1]南京航空航天大学经济与管理学院,江苏南京210016

出　　处：《系统工程》2023年第5期151-158,共8页Systems Engineering

基　　金：国家自然科学基金资助项目(72071111)。

摘　　要：复杂体系的自学习是其智能性、复杂性和自适应性的核心驱动因素,然而目前学术界对其自学习机制的深入解析却比较少见。本文设计了一种基于智能体(agent)技术的GERT网络Bayes“互动-模仿”学习机制。首先根据节点逻辑关系构建复杂体系过程A-GERT网络,再结合Bayes理论进行网络节点传递概率自学习,进而搭建起“互动-模仿”学习机制模型。最后以新兴产业技术突破体系为研究对象进行分析,结果表明,在目标效益驱动下,通过Bayes自学习传递概率能产生相应的动态变化,凸’显技术突破关键路径,为企业决策产业技术突破方案提供建议。Self-learning of complex system of systems is the core driving factor of its intelligence,complexity and adaptability.However,the in-depth analysis of its self-learning mechanism in academia is relatively rare.A Bayesian“interaction-imitation”learning mechanism of GERT network based on agent technology is designed.Firstly,the complex system of systems process A-GERT network is constructed according to the logic of nodes,and then the self-learning of node transmission probability is carried out by combining Bayes theory,and then the“interaction-imitation”learning mechanism model is built.Finally,taking the emerging industry technology breakthrough system as the research object,the results show that driven by the target benefit,the Bayes self-learning transfer probability will produce corresponding dynamic changes,highlight the key path of technology breakthrough,and provide suggestions for enterprises to make decisions on industrial technology breakthrough schemes.

关 键 词：复杂体系 GERT网络 智能体技术 BAYES 自学习 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]