摘要
粒子群算法主要用于优化连续性问题。如果用于求解整数规划问题,算法的粒子位置必须解决取整问题;而量子粒子群算法求解整数规划问题具有更高的效率。利用三种取整方法与量子粒子群算法结合,求解非线性整数规划问题,并且与标准粒子群算法求解整数规划问题进行比较。通过对基准函数仿真实验,比较了六种方法求解整数规划问题。实验结果表明,基于随机取整的量子粒子群算法搜索成功率优于其他五种方法,其综合搜索效率更佳。寻找了一种更优的求解整数规划方法。
The standard Particle Swarm Optimization mainly be used to optimize continuous problem.If using to solve integer programming,the position of particle must be rounded.However,it is more efficient for Quantum Particle Swarm Optimization(QPSO) to solve integer programming.Three kinds of rounding location of particle for QPSO are used to optimize the integer programming,and compared to the standard PSO with the same three kinds of rounding location of particle.By the simulation on benchmark functions,six kind of solving integer programming are compared with.The results of experiment show that the QPSO based on random rounding outperforms the other ways and its' searching efficiency is higher than that of the other ways,so a better method of solving integers programming is found.
出处
《科学技术与工程》
2011年第33期8195-8198,8202,共5页
Science Technology and Engineering
基金
福建省教育厅科技项目(JK2011035)资助
关键词
量子粒子群
整数规划
随机取整
优化算法
QPSO integer programming random rounding optimization algorithm