摘要
提出了以随机A*算法为基础,运用遗传算法来求解不满足先进先出原则的动态网络中最短路径问题的思路,其中通过运用提出的随机Dijkstra算法解决了将遗传算法应用于最短路径问题中的最大障碍——初始种群的产生.最后,以广州市电子地图为基础,对提出的算法进行了验证,验证结果表明遗传算法适合求解非常态且不满足先进先出原则的动态网络中的路径诱导问题.
By applying random A^＊ algorithm, this paper clears out the biggest obstruction between the genetic algorithm and dynamic route guidance, which is how to get the initial generation of genetic algorithm. The developed models and algorithms are implemented with Guangzhou＇s electronic map and their computational performance is analyzed experimentally. The results indicate that dynamic route guidance A^＊ algorithm is suited for route guidance problem in FIFO dynamic network, that dynamic route guidance Q-learning algorithm is suited for route guidance problem in steady non-FIFO dynamic network and dynamic route guidance genetic algorithm is suited for route guidance problem in vertiginous non-FIFO dynamic network.
出处
《交通运输系统工程与信息》
EI
CSCD
2007年第3期45-48,共4页
Journal of Transportation Systems Engineering and Information Technology
基金
国家自然科学基金项目(50578064)
华南农业大学校长基金(2006K017)
