3级网状随机性库存控制策略的仿真优化

Simulation-based optimization on three-echelon network stochastic inventory control policies

研究了多制造商,多分销商和多零售商的3级网状随机性库存系统的(r,Q)库存控制策略问题.由于该系统具有顾客到达时间服从泊松分布,随机顾客需求量,随机顾客购买行为,随机订货时间和制造商生产容量有限制等特点,使得解析方法很难描述系统中的多种复杂随机因素并无法求解有效的库存控制策略.为此建立了以总成本最小为目标的数学模型,运用了基于仿真的优化方法,通过将仿真方法与粒子群优化算法相结合对问题进行求解.最后通过仿真实例与比较,验证了模型和基于仿真的粒子群优化方法的可行性和有效性,也表明了基于仿真的优化方法在供应链管理中的适用性.

The problem of how to set the （r, Q） inventory control policies for a three-echelon network stochastic inventory system is studied. Because the customer arrival time is a Possion process, and all of the customer demand, the customer purchasing behavior and the lead time are stochastic and the production capacity is limited in the inventory system, it is difficult for the analytical method to describe various complex and stochastic factors and develop an effective inventory control policy. A mathematical model is built for minimizing the total cost. Next, the simulation-based optimization method is used to solve the problem by combining the simulation method together with the particle-swarm optimization algorithm. The simulation results demonstrate the feasibility and the effectiveness of the mathematical model and the simulation-based particle-swarm optimization method by comparisons, and show the applicability of the simulation-based optimization method in the supply chain management.