Application of Genetic Algorithm to Multiple Modes Scheduling

点击翻译按钮获取中文摘要

Genetic Algorithm (GA) is known well by J.D.Bagley's study in 1967.GA includes generally three genetic operators, selection, crossover and mutation. The lack of dependence on function gradients makes it more suitable to such problems, like as discrete optimization design problems and optimization design problems with non-convexities or disjointness in design space. The method is tried to apply to Multiple Modes Scheduling in this paper. GA is examined through comparing with exact algorithm (enumeration method) and Monte Carlo Method in practical application. The following are the major points of this paper.(1). Design variables which corresponding to the airline and railway routes are coded directlt to a string. (2). Fitness function is used to avoid the premature convergence to a local solution. (3). Although Multiple Modes Scheduling has been a difficult problem, GA could provide some solution to such problem very easily. (4). The method proposed in this paper was shown to be effective for improvement of GA's reliability. The results suggest that GA is more effective for the optimization of large size airline and railway networks.