【作者】 冯珊； 李锋； 周凯波；

【Author】 FENG Shan, LI Feng, ZHOU Kai-bo(Institute of System Engineering, Huazhong University of Science and technology. Wuhan 430074, China)

【机构】 华中科技大学系统工程研究所；

【摘要】 面向一类高复杂度问题求解,设计多种群协适应式共演化的仿生算法框架。在传统演化算法模型体系中嵌入种群适应性行为调整模块,通过多模式集成,构造演化优化模型及其仿真执行程序,采用多智能体建模仿真平台,支持在个体行为自律和交互规则基础上产生群体的自组织和自适应智能行为。这种协适应共演化过程的同态仿真将高效获得优质的全局近优解。更多还原

【Abstract】 To solve the ever increasingly difficult complex problems, authors devoted their efforts on designing an architecture for evolving coadapted subcomponents, i.e. a multi-population cooperative coevolution system. The above system model shows that through embedding sub-population specific adaptive module into the traditional EC model system, can enable the creation of the desired evolutionary optimization model, from which to obtain the solutions. For the system implementation, we utilize the multi-agent platform (operational environment) and the associated infrastructure (tool kits) to support more plausible system simulations. These obtained simulations can replicate the production of individual intelligence-based population self-organization and intelligent self-adaptative behavior. This type of coevolution processes can be very useful in searching the global qusi-optimal solutions of complex problems.更多还原