随机需求下双产品混合生产的Markov决策过程研究

Hybrid MTO/MTS production decision with stochastic demands based on FMDP

基于按订单生产（make to order，MTO）和按库存生产（make to stock，MTS）两种生产模式的产品，考虑可以忽略生产方式转变带来的安装费用及其它固定费用的情况，建立混合MTO／MTS模式下有限期的Markov决策过程（finite Markov decision process，FMDP）模型．比较研究随机需求下同一生产系统混合生产MTO和MTS产品时，MTO优先模式、MTS优先模式与基于FMDP模型的混合MTO／MTS模式（FMDP混合模式）的效率问题．结果表明，MTO订单量、MTS库存量和产品需求强度是影响FMDP混合模式决策的关键因素，而剩余决策期数、单位MTO订单延迟交货罚金和MTS处理成本则只会在短期内对生产决策产生影响；同时，当企业MTO订单量减少或者MTS产品需求强度降低时，FMDP混合模式比MTS优先模式更优；当MTS产品库存量降低或者MTO产品需求强度降低时，FMDP混合模式同样优于MTO优先模式．

Considering two production modes of products which are make-to-order （MTO） and make-to- stock （MTS） and ignoring setup costs which are from production mode change and other fixed costs, this paper develop a finite Markov decision process （FMDP） model to analyse the hybrid production decision of these two products. A comparative research is done under the situation of hybrid production of MTO and MTS products in a production system with stochastic demand about the efficiency of MTO priority mode, MTS priority mode and hybrid MTO/MTS production decision based on FMDP model （FMDP mode）. The results show that the number of MTO order, MTS inventory and product demand intensity are essential to the decision of FMDP mode, while remaining decision periods, penalty of unit MTO order for late delivery and the cost of handling remaining MTS production impacts are only effective in the short term. When the MTO orders or the demand intensity of MTS production is reduced, FMDP mode is better than the MTS priority mode. With respect to MTO priority mode, the same conclusion is drawn when the stock level of MTS production is reduced or the demand intensity of MTO product is lowered.