适于混合整数非线性规划的混合粒子群优化算法  

Hybrid particle swarm optimization algorithm for mixed-integer nonlinear programming

作　　者：贺益君[1] 陈德钊[1] 

机构地区：[1]浙江大学化学工程与生物工程学系,浙江杭州310027

出　　处：《浙江大学学报（工学版）》2008年第5期747-751,共5页Journal of Zhejiang University：Engineering Science

基　　金：国家自然科学基金资助项目(20276063)

摘　　要：通过加强粒子群优化(PSO)算法处理约束和整数变量的能力,使其适于求解混合整数非线性规划(MIN-LP),构建了一种混合粒子群优化(HPSO)算法.建立了种群的约束矩阵来反映其解满足约束的情况,运用Pareto支配概念评价解的优劣,确定种群的局部最优点和全局最优点.通过增设基于距离函数的概率取整操作和随机变异的解修复操作,加快了搜优速率.利用各粒子的局部最优点信息更新速度,采用多粒子群策略增强了种群多样性.实例测试结果显示,与其他算法相比,HPSO算法具有更好的全局寻优能力,收敛速度更快.For improving particle swarm optimization （PSO） to be suitable to solve mixed-integer nonlinear programming （MINLP）, a hybrid particle swarm optimization （HPSO） was proposed, which can enhance the ability of dealing with constraints and integer variables. Constraints satisfaction of each solution was evaluated based on a constraint matrix for particle swarm, and the concept of Pareto domination was used to evaluate the quality of solution and determine the local and global best positions. A random rounding method based on distance function was introduced to dealing with integer variables, and a stochastic mutation based solution repair strategy was embedded for increasing the convergence speed. Both multi-swarms strategy and velocity update strategy which utilizing all particles’ local best information were adopted to enhance the population diversity, which was propitious to improve the global optimization ability. Several case studies were illustrated to verify the effectiveness and superiority of HPSO. The results show that compared with the literatur’s results, HPSO has better global optimization ability and faster convergence speed.

关 键 词：混合粒子群优化(HPS0) 混合整数非线性规划(MINLP) PARETO支配 多粒子群 解修复 

分 类 号：TQ021.8[化学工程]