Improved Genetic Algorithm for Dynamic QoS Routing to Multiple Destinations for Multimedia Applications (IGADQoS)

摘要

This study presents a new genetic method for solving dynamic multicast routing problem, which is found in multimedia applications. Multicast services in multimedia applications require the optimization of QoS parameters namely cost, end-to-end delay and each link must meet delay and bandwidth constraints. This study proposes an improved genetic algorithm for the construction of QoS multicast tree which has the following features: Multicast tree adopts for dynamic changes; all the links in the tree must meet delay constraint and bandwidth constraint; cost and end-to-end delay is better than other heuristic algorithms; the encoding method helps to perform dynamism; improved genetic operators and heuristic local search operation; Multicast routing over unicast. We have also performed a comparative study of selection mechanisms in GA using simulation and listed the best one for our problem. Experimental results show that our improved genetic algorithm has higher search success rate, convergence rate, dynamic request success rate and least cost than existing methods.