一种求解生产计划问题的改进量子粒子群算法

摘要

It is known that algorithms for modeling of production planning in process industry is easy to fall into local extreme points. So this paper presents an improved quantum-particle algorithm. According to the constraints of production planning model, this algorithm dynamically calculates particle' s upper and lower limits. It adds boundary amendment and introduces a kind of mutation operation based on probability distribution, thus increases the population diversity and improves the ability to get rid of local extreme points. Finally, comparison is given between improved algorithm and other algorithms. The simulation result shows that the improved algorithm has obvious advantages for giobal search capability and it is an effective way to solve production planning problem.%在生产计划优化过程中,针对流程工业生产计划模型的求解算法易陷入局部最优的问题,为了获得全局最优提出了一种改进的量子粒子群算法.根据生产计划模型的约束条件动态计算粒子的上下限,对粒子进行边界修正,并引入了基于概率分布的变异操作,从而增加了种群的多样性,改善了算法摆脱局部极值点的能力.最后仿真实验对改进算法和基本量子粒子群算法、遗传算法等方法进行了仿真比较,结果表明改进算法在全局搜索能力有了明显的提高,是一种求解生产计划问题的有效方法.