多智能体遗传算法用于线性系统逼近

提出了一种新的参数优化方法--多智能体遗传算法,来求解线性系统逼近问题. 该方法中每个智能体代表一个候选解,即搜索空间中的一个实值向量.所有智能体生存在一 个网格状的环境中,且每个智能体占据一个格点不能移动.为了增加能量,它们将与其邻域 进行合作或竞争,也可以利用自身的知识.因此,设计了4个进化算子来模拟智能体间的竞 争、合作、自学习等行为.该方法利用这些智能体与智能体间的相互作用来达到优化逼近模 型中参数的目的;此外,还采用了一种动态扩展搜索空间的方法以解决算法所需的搜索空间 难以确定的问题.实验中,利用一个稳定和一个非稳定的线性系统逼近问题来验证算法的性 能,并与两种新近提出的方法作了比较.结果表明,该文方法优于其它方法,能够用较少的计 算量找到高质量的逼近模型,具有良好的性能和实际应用价值.

Optimal Approximation of Linear Systems by Multi-agent Genetic Algorithm

The problem of optimally approximating linear systems is solved by the multi-agent genetic algorithm(MAGA).In MAGA,each possible solution,a real-valued vectorin the search space,is considered as an agent,and all agents live in a latticelike environ-ment,with each agent being fixed at a lattice-point.In order to increase energies,theycompete or cooperate with their neighbors,and they can also use knowledge.Therefore,four evolutionary operators are designed for simulating the intelligent behaviors of agents,such as competition,cooperation,self-learning and so on.Making use of these agent-agentinteractions,MAGA realizes minimizing the objective function value.At the same time,asearch-space expansion scheme is adopted to find the regions where the optimal parameterslocate.In experiments,two linear systems,a stable one and an unstable one,are used totest the performance of MAGA,and a comparison is made between MAGA and two recentalgorithms.The results show that MAGA can find high quality approximate models withlow computational cost.