基于GA的网络最短路径多目标优化算法研究

Research on multi-objective optimization for shortest path algorithm based on GA

针对现有基于遗传算法(GA)优化的网络最短路径算法存在优化目标单一、遗传编码质量低、搜索策略间平衡性差、适应度分配效率与灵活性较低等问题,建立一种多目标优化最短路径自适应GA模型.提出了优先级编码和优先级索引交叉算子,引入了遗传算子参数的模糊控制机制和基于自适应加权的适应度分配方法.实验结果表明,该算法的准确性和稳定性高、复杂度合理,实现了对网络设计优化中多目标最短路径问题的高质量求解.

The singleness of the optimization objective, poor performance of genetic representation, unbalance between searching strategies, and low efficiency of fitness assignment are main problems of the conventional shortest path（SP） genetic algorithms （GA）. Therefore, an adaptive SP multi-objective （MO） GA is proposed. Priority-based genetic encoding and priority-indexed crossover are introduced. Fuzzy logic based genetic operator adaptation and adaptive weight fitness assignment methods are designed. Simulations of the model based on various scale of networks effectively show that the high requirement of SP problem is well fulfilled with high accuracy and stability of the proposed MOGA.