【摘要】 本文提出了一种确定多阶段（或单元）串并联拉式系统优化控制策略的有效方法．因整体系统具有较大的状态空间，如果建立整体系统模型，尽管理论性较强，但实用性差．本文通过考虑随机生产可靠性和需求波动两类随机因素，建立了整体系统的分解模型，能方便地确定出串并联拉式系统的优化控制策略，模型为离散时间的马尔可夫过程模型．更多还原

【Abstract】 This paper provides a means to decide the optimal controlled decision of multistage tandem-parallel pull systems. The models of whole systems have good theoretical implication but poor practical performance, because they have too large state spaces. In this paper capacity constrains, stochastic production reliability and demand variability are included, and the decomposed models of whole systems are built. Tile optimal controlled decision for pull systems can be easily decided by the models. Pull systems are modeled as discrete-making Markov processes.更多还原