中文摘要：

为了降低备件对流动资佥的占用，针对模具制造设备B类备件需求的随机性、离散性特点，将备件的需求视为一个泊松过程。采用（s，S）库存控制策略，考虑到此策略下备件的库存量和缺货量本身所具有的马尔可夫特性，应用马尔可夫链构建了库存量和缺货量的马尔可夫模型并推导出了状态转移概率矩阵的计算公式。在此基础上，建立了在满足一定服务水平的前提下，以一个盘点周期内期望库存总成本最小为目标的单一备件最优随机库存控制模型。鉴于某些解析式算法求解的局限性，采用了以启发式算法和启发式搜索规则相结合的求解思路，有效地缩小了模型最优解的搜索范围，最终通过MATLAB仿真实现模型求解并得到了启发式规则。以某大型轮胎模具制造企业为例，验证了该算法的有效性，表明该方法可较好地解决随机库存系统的B类备件库存控制问题。

英文摘要：

Study was made to reduce the capital for spare parts inventory in which the de- mand for spare parts was regarded as a Poisson process based on the randomness and discrete characteristics of demands for class B spare parts of mould manufacturing equipments. The （s,S） inventory control policy was used and Markov chain was applied as the inventory level and stockout quantity has Markov property, in which the Markov model was built and the formulas of the state transition probability matrix was deduced. An optimal stochastic inventory control model was established with the goal of minimum long run average total cost of a single spare part in a check period and contentment of certain service level. Because of the limitations of some analytic algorithms, an idea of heuristic algorithm combining heuristic search rule was used to narrow the scope of solution space effectively, the model was solved by MATLAB simulation and heuristic rules were got. Finally, an ex- ample of controlling the spare parts of a large tire mould manufacturing company was presented to illustrate that the algorithms can preferably solve the problem of stochastic inventory control system.