论文部分内容阅读
求解有独立到达时间和完工时间的单机调度问题目前主要采用启发式算法。为研究仿生类算法的应用可行性,采用蚁群算法,以提前/拖期的总惩罚量达到最小为目标。将邻位工作交换法(AP I)用于局部搜索以提高解的质量,并对每一个解的相邻工作间隔时间进行优化调整。用90个测试算例,将蚁群算法求解结果与分支定界法和禁忌搜索法的结果进行比较。结果表明,蚁群算法与分支定界法和禁忌搜索法的结果相当。
At present, the heuristic algorithm is mainly used to solve the single-machine scheduling problem with independent arrival time and completion time. In order to study the feasibility of application of biomimetic algorithms, an ant colony algorithm is used to minimize the total amount of early / late penalty. The orthogonality work exchange method (AP I) is used for local search to improve the quality of solutions, and the optimization of the adjacent working interval of each solution is made. With 90 test cases, the result of ant colony algorithm is compared with the result of branch and bound method and tabu search method. The results show that the ant colony algorithm is equivalent to the result of the branch-and-bound method and tabu search method.