基于AGA与MPSO的非传统布局仓储货位分配优化

Storage location assignment optimization in non-traditional warehouse base on AGA and MPSO

非传统布局是现代仓储管理的新热点,根据对非传统布局(Fishbone型)特征分析,针对货位分配优化问题,提出以出入库效率和货架稳定性为优化目标,建立多目标优化模型。设计了自适应遗传算法(AGA)和改进的粒子群优化算法(MPSO)进行求解。AGA采用动态自适应策略改进选择、交叉、变异算子,克服初期"早熟",提高末期局部搜索,增强鲁棒性;考虑到PSO搜索过程的非线性复杂特征,引入非线性变化的惯性权重和时变加速的学习因子,提升早期全局搜索能力,改善末期收敛迟钝,优化算法整体性能。采用Matlab进行仿真实验,结合实例验证了本文方法的有效性与通用性。对比实验结果表明AGA在处理此类货位分配优化问题上优势更明显。

Non-traditional layout warehouse is a new hot spot in the current storage industry. In view of the storage location assignment optimization with Fishbone layout, taking inventory efficiency and shelf stability as the optimization objective based on its characters, an optimization model for the storage location assignment is formulated. Then, the adaptive genetic algorithm (AGA) and modified particle swarm optimization (MPSO) are developed to solve the problem. A dynamic adaptive strategy is exploited in AGA to improve selection, crossover and mutation operators, which can overcome premature, improve the capability of local search ,and enhance the robustness. Considering the complex nonlinear characteristics of PSO searching process, the inertial weights and the time-varying accelerated learning factors are introduced to improve the early global search ability, the late convergence and optimize the overall performance. These algorithms are compiled by Matlab, and the effectiveness and versatility are verified through a case. The experiment results show that AGA is superior to other algorithms in dealing with storage location assignment optimization.