【摘要】 粒子群算法在求解连续变量问题有了比较成功的应用,但是对离散变量问题方面的应用研究却相对滞后。针对离散优化问题,提出了一种遗传粒子群算法。算法使用了交叉、变异等遗传算子替代传统粒子群算法的速度-位移公式,克服了传统粒子群算法对组合优化问题编码时出现的信息冗余的问题,提高了搜索效率。应用该算法求解了车辆路径问题,实验结果表明,该算法具有较好的全局收敛能力和较快的收敛速度。在同等条件下,求解效果要明显好于遗传算法和基于速度位移公式的粒子群算法。更多还原

【Abstract】 Particle swarm optimization algorithm is successful for solving the problems of continuous variables,but it is not so good for solving the problems of discrete variables.A genetic particle swarm optimization algorithm(GPSO) is proposed for solving the discrete optimization problems.It uses the crossover and mutation operator instead of velocity-displacement operates to update the particles.The problem of information redundancy in solving combinatorial optimization has been overcome.It is used for solving the vehicle routing problem.Experimental results indicate that GPSO has better global convergence and faster convergence rate.In contrast to the GA with the same operators and the PSO based velocity-displacement operates,GPSO has much better performance.更多还原