利用自适应混合遗传算法求解平车装载问题  

Solution of Flatcars Loading Problems Based on Adaptive Hybrid Genetic Algorithm

作　　者：吴虎胜[1,2] 张凤鸣[2] 赵法栋[1] 

机构地区：[1]武警工程大学装备工程学院,陕西西安710086 [2]空军工程大学装备管理与安全工程学院,陕西西安710051

出　　处：《铁道学报》2013年第12期1-8,共8页Journal of the China Railway Society

基　　金：国家自然科学基金(71171199)

摘　　要：铁路输送中的多型号平车装载问题旨在满足铁路军事运输和安全的要求下,提高铁路平车利用率,节省铁路运输资源。针对平车优先的平车装载问题,建立数学模型,将改进的自适应遗传算法、模拟退火算法思想、优先适应降序算法相结合,提出一种自适应混合遗传算法。给出算法的编码、遗传算子的设计、适应度函数的定义、交叉概率和变异概率的自适应改变算法。针对平车装载问题实际,对算法编码进行解码,利用优先适应降序算法对装备进行装载;将遗传算法和模拟退火算法思想结合以分别弥补各自局部搜索和全局搜索局限;引入自适应机制使得交叉概率和变异概率能随适应度值改变而改变,抑制了算法"早熟",提高了收敛速度;最后用实例验证了该算法的有效性。In military railway transportation, how to improve the utilization rate of railway flatcars and to satisfy military and safe traffic requirements is the key to loading of multiple-type flatcars. In this paper, to cope with the preferential flatcars loading problem, the mathematical models were established and the adaptive hybrid genetic algorithm was put forward by combining the improved adaptive genetic algorithm, simulative annealing algorithm and best fit decreasing algorithm. The genetic algorithm coding, definition of the fitness function, design of genetic operators and the algorithm which adaptively determined the probabilities of crossover and mutation were theoretically expatiated. Firstly, the idea of the best fit decreasing algorithm was used to load equip ment into flatcars. Secondly, by combining the genetic algorithm with the simulative annealing algorithm their respective local and global searching defects were remedied. Thirdly, by introducing in the adaptive mecha- nism, the probabilities of crossover and mutation were made to change adaptively with fitness function values so as to eliminate the premature convergence. Finally, the effectiveness of the proposed method is verified by case study.

关 键 词：铁路输送 平车装载 优先适应降序算法 自适应混合遗传算法 

分 类 号：U294.893[交通运输工程—交通运输规划与管理]