Uncertain Production Environment Under The Conditions Of Production Scheduling Optimization

Production scheduling is the headquarters of production running in enterprises. Therefore, it's necessary to carry on researches on production scheduling in the process enterprise, which will have important theoretical significance and practical values.Scheduling is time distribution to shared resource in order to realize certain purposes. Research to production scheduling mainly fixed on deterministic scheduling problems when scheduling problems were proved to be NPH problems. However, the real world has to be limited, simplified and assumed; deterministic scheduling has finite model description ability and has to give up lots of useful information, which may lead to distortion of models and unfeasible results. So the key to this paper is how to reduce or avoid the effects of disturbs which are brought by uncertain production environment. Here the problem was solved in two ways. 1. Uncertain information was added to the pre-scheduling model as chance constraints before optimal scheduling result that was researched with scheduling stability above some level was attained. 2, Key coefficients whose bounds should be limited strictly were gotten by postoptimal analysis if scheduling was determined. Good performance would be kept if key coefficients had been handled properly.The main contribution and research works are as follow:First, various factors which influence the performance of production scheduling under uncertain production circumstances were analyzed and sorted. And then basic measures of uncertainty were definited. The transforms between chance constraints and their equivalence shapes in the production scheduling model based on uncertain programming were shown and proved, and it settles the theoretical bases for simplifying the complicated models and raising speed of its algorithms.Second, the definition of multi-uncertainty measure was presented following AS theory. Then the production scheduling model based on production scheduling stability under uncertain production circumstances was built by using multi-uncertain chance constraints. It not only guaranteed the stability but also improved the performance of the production scheduling.Third, the interval number coefficient programming model of productionscheduling was established in the third chapter. The function relationship between the object value and decision coefficients was discussed by the analysis of changing of constraint properties in the model. Then the rules which were needed to adjust production to maintain scheduling performance were reached.Finally, some critical issues in the application of production scheduling systems in process industry were listed. The national '863' project, Real-time Production Plan & Scheduling System in Process Industry was used as a typical example to demonstrate the design of equation database and algorithms, which supplied a framework for the applications of production scheduling systems.