摘要
实时、高效的求解大规模路网中的最优路径是动态路径诱导领域的研究难点。针对基本遗传算法在计算大型网络的优化问题时表现出的求解效率低等缺点,在基本遗传算法中引入了子群体和迁移策略,提出了基于并行遗传算法的最优路径选择方法,设计了适用于路径优化的编码方式、适应度函数、遗传操作算子和迁移算子,并采用神经网络预测方法构造了实时动态的路阻矩阵。仿真试验表明:该方法的准确性、实时性和快速性优于基本遗传算法,并且大规模路网中求解效率和求解质量的平衡问题也得以解决。
It is a difficult problem that computing the optimal route of a gigantic traffic network. In order to solve the problem of seeking the answer ineffectively, two new operators including subpopulation and migrate tactics have been adopted. An optimal route choice method based on parallel genetic algorithm is presented and the coding way, the selection, crossover, mutation and migrate operators are designed. The road weight matrix has been composed by the neural network forecasting method. It is indicated by simulation that the presented method has achieved more accurate, real-time and quick guidance than GA and the balance problem of the quality and the efficient of seeking the answer has been solved.
出处
《微计算机信息》
北大核心
2005年第12Z期166-168,32,共4页
Control & Automation
基金
科技部国际重点合作项目(2003DF020009)
关键词
动态路径选择
路径优化
神经网络
路阻矩阵
并行遗传算法
迁移策率
Dynamic Route Choice
route optimization
Neural NetWork
Road Weight Matrix
Parallel Genetic Algorithm
migrate tactics
