Study On Multi-objective Single Machine Scheduling Problem Under Just-in-time Production Mode

Pareto-based multi-objective scheduling solution has been the hot issue of domestic and foreign researchers. Current solutions have three main categories: deterministic algorithm, heuristic algorithms and artificial intelligence algorithms. However, research on multi-objective scheduling problem tended to focus only on the algorithm itself, do not give enough attention for the algorithm evaluation and test data sets. And the researchers only concern one indicators of pareto solution set to improve the algorithm performance ,making it difficult to increase the overall quality of solution sets.To solve the problem, this thesis propose an algorithm evaluation methods"obtain the better solution within reasonable time"considering the time consuming of the method and the quality of the solution set , and integrated single solution quality, performance of the distribution, range of distribution three indicators into the shortest distance weighted method to evaluate the quality of pareto solution set. Then this thesis focus on the on multi-objective single machine scheduling problem under Just-In-Time production scheduling problem, propose an improved Recovering Beam Search algorithm based on sub direction concept. Then, we discuss the condition for the test set and propose 80 groups of test set data. Finally this thesis compare the Recovering Beam Search algorithm to the Sub direction Recovering Beam Search algorithm on the basis of the proposed method of algorithm evaluation.