移动自组织网络Q学习和改进蚁群QoS路由算法

QoS Routing Algorithm Based on Q-Learning and Improved Ant Colony in Mobile Ad Hoc Networks

针对移动自组织网络的QoS路由问题,提出一种结合Q学习和改进蚁群算法的QoS路由算法,该算法综合Q学习和蚁群算法的优点,把Q学习算法的Q值作为蚁群算法的初始信息素,提高了算法初期的收敛速度,同时在路径选择时综合考虑节点的能量和负载.仿真实验表明,该算法在保证QoS需求的前提下,增加了路由的有效性和鲁棒性,降低了能耗,包投递率、网络生存时间等指标均较好.

In view of QoS rounting problem in mobile ad hoc networks, the author proposed a QoS routing algorithm integrated with Q-learning and improved ant colony algorithm. The algorithm combines the advantages of Q-learning with those of ant colony algorithm, and it takes Q value of Q-learning algorithm as the initial pheromone of ant colony algorithm, improves the initial convergence speed of the algorithm, at the same time, takes the node energy and load into account in path selection. Simulation results show that on the premise of guaranteeing QoS demand, the algorithm increases the effectiveness and robustness of routing and reduces energy consumption, and besides, packet delivery ratio, network lifetime and other indicators display better performances.