中文摘要：

为使存贮、生产和缺货等费用的总和最小，建立了一种多周期随机需求生产／库存模型，该模型采用（S，Q）策略对生产和库存进行控制，即当成品库存降至8时准备生产，生产量为Q。通过对该模型费用函数特性的分析，设计了一种迭代学习算法，根据该算法可以得出系统的最优生产准备点及最优生产量。将所提出的迭代学习算法与遗传算法进行了比较，结果显示，两者所得到的控制量是吻合的，且迭代学习算法的求解速度更快，从而证实所建立的模型和提出的迭代学习算法是正确有效的。

英文摘要：

In order to minimize the sum of the holding cost, setup cost and shortage cost, a multi -period production/inventory model under random demands was set up, which adopted（s, Q） policy for production/inventory control, that is when the stock level of finished goods was less than s, production preparation started, and the production quantity was Q. And then, through analyzing the model,an iterative learning algorithm was designed, by which the optimal production preparation point and optimal production quantity can be acquired. Lastly, the proposed iterative learning algorithm was compared with Genetic Algorithm （GA）, the results show that the control quantities obrained by the two methods are consistent, and the iterative learning algorithm has higher solution speed. So the correctness and effectiveness of the model and iterative learning algorithm proposed are verified.