中文摘要：

针对传统遗传算法在求解自动化立体仓库货位优化多目标模型中容易陷于局部最优解以及交叉变异过程中产生大量不可行解等问题,提出了并列选择单亲遗传算法.算法采用了0,1矩阵编码、并列选择算子、单亲变异算子等,有效避免了交叉变异操作产生不可行解的问题.通过对控制参数进行较合理地选取,算法能够综合考虑各子目标的相对优秀个体,从中选取出全局近似最优解,有效降低了算法陷于局部最优解的概率.利用该算法对36种货物的自动化立体仓库货位进行优化,通过比较优化前后的货位对应的拣选时间及货架重心可以看出,优化后的货位对应的拣选效率及货架稳定性均有明显提高.

英文摘要：

In view of the traditional genetic algorithm in solving the warehouse slotting optimization multi-objective model easily trapped into local optima and produced a large number of unfeasible solutions in the cross over and mutation operations,this paper presents parallel selection partheno-genetic algorithm.The algorithm uses the 0,1 matrix encoding,parallel selection operator,single parent mutation operator,which effectively avoid producing unfeasible solutions in cross over and mutation operations.Through the selection of the control parameters,the improved algorithm is able to take into account the relative excellent individual of each sub target,then take out the global approximate optimal solution and effectively reduce the probability of trapping into local optimal solution.Using the algorithm to optimize the storage location of 36 kinds of goods in automated warehouse,and compare the picking time and shelf barycenter before and after optimization.The results show that the picking efficiency and shelf stability of optimized location are improved significantly.