基于改进遗传算法的动态交通分配优化研究

Research of Optimization on Dynamic Traffic Assignment Based on Improved Genetic Algorithm

传统的优化算法大多数由于计算量大或者容易使性能指标落入局部最优值而严重制约了模型的应用与发展,而用遗传算法等新的智能算法求解则会很简洁和方便。文章针对遗传算法在应用中存在的局限性,采用了小生境技术的遗传算法,结合精英保留策略、种群多样性保持方案、新的适应度值标定方式等改进遗传算法。通过对动态交通分配的特点分析,建立了动态交通分配模型,利用改进的遗传算法对模型进行求解。仿真结果体现了动态交通分配模型的有效性和改进遗传算法的优越性,大大提高了动态交通分配模型的实用价值。

Traditional optimization algorithms severely restrict the application and development of the model in that large amount of calculation of them or easy to fall into local optimal value of performance index,but it is simple and convenient to settle such problems via new intelligent algorithms such as the genetic algorithm,etc.Aiming at limitations of the simple genetic algorithm on application,the niche technique is adopted and it combines elitists reservation model,scheme of keeping population diversiform to improve basic genetic algorithm.The dynamic traffic assignment model is established via to analysis its characteristics,and using improved genetic algorithm to solve the model.The simulation result shows the validity of dynamic traffic assignment model and superiority of improved genetic algorithm and improve the practicality of the model of dynamic traffic assignment.