Data transmission among multicast trees is an efficient routing method in mobile ad hoc networks(MANETs). Genetic algorithms(GAs) have found widespread applications in designing multicast trees. This paper proposes a stable quality-of-service(QoS) multicast model for MANETs. The new model ensures the duration time of a link in a multicast tree is always longer than the delay time from the source node. A novel GA is designed to solve our QoS multicast model by introducing a new crossover mechanism called leaf crossover(LC), which outperforms the existing crossover mechanisms in requiring neither global network link information, additional encoding/decoding nor repair procedures. Experimental results confirm the effectiveness of the proposed model and the efficiency of the involved GA. Specifically, the simulation study indicates that our algorithm can obtain a better QoS route with a considerable reduction of execution time as compared with existing GAs.
实际交通环境规划最优路径的重要问题是无人车智能导航,而无人车全局路径规划研究主要在于模拟环境中算法求解速度的提升,考虑大部分仅路径距离最优或局限于当前道路的自身状况,本研究针对实际环境中的其他因素及其未来的变化和动态路网中无人车全局路径规划的复杂任务,基于预测后再规划的思想提出面向实际环境的无人车驾驶系统框架,并结合深度Q学习和深度预测网络技术提出一种快速全局路径规划方法(deep prediction network and deep Q network, DP-DQN),从而利用时空、天气等道路特征数据来预测未来交通状况、求解全局最优路径。基于公开数据集的试验和评价后发现,本研究提出的方法与Dijkstra、A*等算法相比,行车时间最高降低了17.97%。
