| Summary: | 碩士 === 國立臺灣大學 === 工業工程學研究所 === 92 === Ant colony optimization (ACO) is a new kind of heuristic algorithms in recent years. There are successful implementations of applying ACO to a number of different combinational optimization problems such as traveling salesman problem (TSP), bin packing problem (BPP), job shop problem (JSP), and so on. An improved ACO method, called added to the superior segments and subtracted for inferior segments ant system (ASDSAS) is presented.This method is to avoid stagnation and decrease tour constructions. At first, the ants are classified into two groups- superior and inferior using groups the grouping rule. Only the routing segments traversed by the superior and inferior ants are considered for pheromone updating. Moreover these segments traversed by the superior and inferior ants can be further classified into superior and inferior segments. Pheromone is added to the superior segments and subtracted from the inferior segments. Stochastic factor is adapted in pheromone updating.
ASDSAS is applied to solve TSP, BPP, and JSP. Several TSPLIB and ORLIB are tested and results are compared with other ACO Systems. Results show that ASDSAD can achieve the same solution, some even better, but use lesser computer resources.
|
|---|
