调度系统中指派问题的符号算法研究.pdfVIP

Abstract RESEARCH ON THE SYMBOLIC ALGORITHM FOR ASSIGNMENT PROBLEM IN SCHEDULING SYSTEMS Abstract Scheduling problem is one of the NP complete problems which attract the attention of thousands of researchers. Due to the complexity of scheduling problem, diversity of scheduling knowledge and dynamics of manufactural environment, scheduling problems must be solved by considering multiple factors. This dissertation reduces the assignment problem in scheduling systems to the manipulation and calculation of Boolean functions, and solves the problem by symbolic technology. The mainly works of this dissertation are as follows: At first, we study the ordered binary decision diagram and data structures that develop from it and analyze their properties and applications in symbolic technology, which serves as the basis for developing our symbolic algorithm. Secondly, the property and solutions of assignment problem in scheduling systems are analyzed; considering the parallel manufactural phenomenon in practice, we propose a mathematical model for weighted assignment problem and an algorithm for weighted optimal semi-matching in bipartite graphs, which reduces the weighted optimal semi-matching problem to classical weighted matching problem in bipartite graph. Thirdly, the efficiency of symbolic algorithm for optimal semi-matching in bipartite graphs, ASM1 and ASM2 are compared and analyzed, the experiment results show that symbolic algorithm has advantages in processing high-density and large-scaled graphs.