考虑缺陷品数量折扣的随机联合补货及混合骨干差分进化算法

Stochastic Joint Replenishment Considering Discounted Items with Quality Defects and Hybrid Simulated Annealing Bare-Bones Differential Evolutionary Algorithm

【目的】为了满足客户需求的多样性并最大限度地减少缺陷品对顾客满意度的影响。【方法】同时考虑随机需求和随机存在的缺陷项目数量,设计算法对缺陷品进行100%筛选后进行全单位数量折扣处理。【结果】根据骨干差分进化算法(Bare-bones differential evolutionary algorithm,BBDE),结合模拟退火算法(Simulated annealing algorithm,SA)算法的Metropolis准则,设计了基于SA的混合骨干差分进化算法(SA-based hybrid BBDE,SAHBBDE),以提高BBDE的全局寻优能力。【结论】数值实验表明,该算法在总成本最佳、平均值最低及标准误差最小的表现上优于遗传进化算法(Genetic algorithm,GA)、粒子群算法(Particle swarm optimization algorithm,PSO),与骨干粒子群算法(Bare bones PSO algorithm,BBPSO)、差分进化算法(Differential evolution algorithm,DE)和BBDE相比也表现出优异的性能。

[Purposes]With the development of customer consumption level,product quality has been considered to be one of the competitiveness of enterprises.[Methods]In the practice of Joint Replenishment Problem,imperfect items can disrupt normal operations and cost supply chain stakeholders.In addition,because the uncertain needs of the customer will affect replenishment decisions,this article considers both random requirements and the number of random defective items,which are screened 100%and then discounted in full unit quantity.[Findings]Based on the Bare-bones Differential Evolution Algorithm(BBDE),a mixed bare bone differential evolution algorithm(SAHBBDE)based on the Metropolis guidelines of the Simulated Annealing algorithm(SA).[Conclusions]Numerical experiments show that the proposed SAHBBBDE algorithm is superior to genetic evolution algorithm(GA),particle group algorithm(PSO),in total cost,lowest average and least standard error,compared with the bare bones PSO algorithm(BBPSO),differential evolution algorithm(DE)and BBDE.SAHBBBDE also shows excellent performance in large-scale tests compared to differential evolution algorithms(DE)and BBDE.