间歇生产调度优化模型的分周期逼近算法

There are many uncertainties in the process of intermittent production scheduling, the most important of which is the uncertainty of demand. Considering the uncertain demand of multi-week intermittent production scheduling optimization model, the scheduling time domain is divided into a large number of time periods related to scheduling decisions by using discrete or continuous time expressions, which leads to the existence of a large number of integer variables in the model, which makes it difficult for the model to solve. In this study, the existing solution methods are analyzed, and a periodic approximation algorithm is proposed. The multi-week intermittent production scheduling decision problem is decomposed into the first cycle scheduling decision problem and the rest of the cycle scheduling decision problem, which simplifies the structure and accelerates the solution speed. Through the scheme tree aggregation, the scheme tree expressing the uncertain information of the demand is transformed into several scheme files, and the deterministic method is applied to each scheme file to obtain the scheduling decision, but only the first-cycle scheduling decision is retained, which can reduce the impact of the minimum benefit scheme on the expected benefit and improve the level of the first-cycle scheduling decision. After obtaining a number of candidate scheduling decisions in the first cycle, the optimal scheduling decisions for the rest of the cycles were determined by the three-stage method of time contraction, and the time contraction strategy and compensation strategy were applied to improve the scheduling decision level of the remaining cycles. Finally, the candidate scheduling decisions of the first cycle are evaluated with the expected benefits, and the scheduling decisions of all cycles are determined. The case study proves that the proposed algorithm can improve the decision-making level of intermittent production scheduling, speed up the solution speed, and effectively solve the optimization model of multi-week intermittent production scheduling.