基于多种群细菌觅食算法的机械设计优化

摘要

针对传统方法在解决机械优化问题时的不足,提出了一种偏利共生型的多种群细菌觅食算法.利用可行种群和不可行种群两个种群来模拟协同演化,以用于机械设计优化问题的多约束处理.除多种群协同进化外,该算法还采用了自适应的趋化步长,以及均衡的繁殖消亡策略.为验证所提出算法的优化设计能力,在压力容器和焊接悬臂梁优化设计两个实例上进行了测试,并与其他几个算法进行了比较.测试结果表明所提出的多种群细菌觅食算法在最优值以及稳定性上相比于同类算法具有明显优势,能够很好地用于机械设计优化工作.%To overcome the disadvantage of traditional methods in solving the problem of mechanical optimization ,a Multi-Swarm Bacterial Foraging Optimization ( MSBFO) algorithm was proposed ,which uses the idea of commensalism in nature.In the algo-rithm,two swarms were used to represent the feasible population and infeasible population ,to simulate the co-evolution,in order to solve the constrained mechanical design optimization problem.Adaptive step lengths in chemotaxis and equilibrium strategy in re-production are also used in MSBFO.To validate the optimization design ability of proposed algorithm , tested it on instances of pressure vessel design and welded beam optimization design ,and compared its results with several other algorithms.Test results show that the proposed MSBFO algorithm outperforms most comparison algorithms both on the optimal value and the stability of the results,and can be used for mechanical design optimization very well.