基于信息熵的Push/Pull策略下供应链不确定性

以3级多产品供应链为例，分别采用Push/Pull/混合Push & Pull /改进Push 4种控制策略运行供应链，采用总熵比量化供应链的全局不确定性.将基于仿真优化(SBO)与遗传算法(GA)相结合，解决了计算量大与不确定因素多的难题.以控制规律中的增益为决策变量，对不确定性进行优化，计算出最优决策变量下的客户满意度、超量库存、延迟交货和总成本等常用性能指标.仿真结果表明:供不应求时，采用混合Push & Pull策略可以降低总成本;供大于求时，采用改进Push策略能最大程度降低供应链的不确定性.

Information Entropy-Based Supply Chain Uncertainty Under Push/Pull Strategies

Taking the 3-level multi-product supply chain as an example，the supply chain was ran by using four strategies， i.e.， Push， Pull， hybrid Push & Pull and improved Push. The global uncertainty of supply chain was quantified by total entropy ratio. Genetic algorithm(GA) was combined with simulation-based optimization(SBO) method to deal with the difficulty of large amounts of computation and uncertainty. The gain in the control law being as the decision variable， the uncertainty was optimized， and other common performance indicators such as customer satisfaction， excess inventory， delayed delivery and total cost were calculated under the optimal decision variables. The simulation results showed that when the demand exceeds supply， the hybrid Push & Pull strategy can reduce the total cost. When the supply exceeds demand， the improved Push strategy can minimize the uncertainty of the supply chain.