一种约束类问题的带权PSO优化方法

An Optimized Method for Solving Constrained Problem Based on Weighted PSO Algorithm

  • 摘要: 为提高带约束类问题的PSO求解质量,将等式约束通过约减策略转化成不等式约束,约减了设计变量个数,降低了粒子的维度;同时将不等式约束事先放在子程序内,在使用PSO寻优计算适应度函数前,排除不在解空间内的解,降低了寻优计算量.利用优化过程中其他粒子的优化信息调整PSO算法的惯性权重,给出了约束类问题的带权PSO优化步骤.采用这种带权的PSO算法以及传统的PSO算法,分别对等式约束处理、不等式约束处理以及两者均处理的3种方案同时优化2个典型实例,对实例求解的最优值、平均值、标准差以及平均优化时间进行对比,结果显示:这种带权PSO算法对约束条件同时处理的方法既能提高解的精度,又能提高优化的求解效率.

     

    Abstract: In order to improve the solution quality of PSO with constrained problems, by converting equality constraint to inequality constraint through reduction stategy, the numbers of designed variables and dimensions of particles are reduced; meanwhile, by placing inequality constraint into subroutine beforehand, solutions that not within the solution space are excluded before using PSO optimization to calculate the fitness function so that the optimization calculation is reduced. By using informations of other particles during the optimization, inertia weights of PSO algorithm are adjusted, also, the optimization steps of the weighted PSO with constraint problems are given. By respectively dealing with the equality constraint, the inequality constraint and both of them through the weighted PSO and the tradional PSO to optimize two typical examples, from the comparisons of the optimal values, mean values, standard deviations and average optimization times which are obtained from the calculations of the examples, it is showed that the method of simultaneous processing to the constraint conditions by using weighted PSO can improve both the accuracy and effiency of the solutions.

     

/

返回文章
返回