论文部分内容阅读
遗传算法等启发式算法在求解旅行商问题时,存在收敛速度较慢、容易出现过早收敛及算法计算效率较低的问题。在模式理论基础上,提出一种新的基因重组算法。根据优良基因模式,设计模式重组算子,运用重构及进化规划的思想设计算法的个体重构算子和个体选择算子。建立一个多目标旅行商问题模型,分析每一轮计算旅行路线适应度值的差异性,采用熵值法确定路程和费用权重。系列实验表明,基因重组算法在求解多目标旅行商问题时,计算效率远高于比较的算法,收敛速度和求解精度也较一般启发式算法有明显改善。
Heuristic algorithms, such as genetic algorithms, have some problems in solving the traveling salesman problem, such as slow convergence, premature convergence, and low computational efficiency. Based on the model theory, a new gene recombination algorithm is proposed. Based on good gene patterns, design pattern reorganization operators, individual reconstruction operators and individual selection operators using the ideological design algorithms of reconstruction and evolution planning. A multi-objective traveling salesman problem model is established. The discrepancy of fitness value of travel route is analyzed in each round. The entropy method is used to determine the distance and cost weight. Experiments show that the genetic recombination algorithm is more efficient than the comparison algorithm in solving multi-objective traveling salesman problem, and the convergence speed and the accuracy of the solution are obviously improved compared with the general heuristic algorithm.