Application Research Of Bayesian Optimization Algorithm On QoS-Based Multicast Routing Problem

Probabilistic modeling evolutionary algorithms (PMEAs) that incorporate building models into evolutionary algorithms have become a new class of evolutionary algorithms. Bayesian optimization algorithm (BOA) is a representative algorithm of PMEAs to solve high order hard problems. This paper mainly studies BOA and its applications on the QoS-based multicast routing problem in computer networks. Also, this paper discusses how to solve the QoS-based multicast routing problem with multiple restrictions by combining BOA with multi-objective evolutionary algorithms.Firstly, this paper surveys the researching progress of PMEAs and BOA. On the basis of discussing how to map the QoS-based multicast routing problem into a problem that BOA can solve, a QoS-based multicast routing algorithm based on BOA is proposed. The experimental results show that the new algorithm can converge quickly to the optimal solution. Secondly, in order to overcome negative impacts caused by genetic operators in traditional multi-objective evolutionary algorithms, this paper proposes an improving strength Pareto evolutionary algorithm (SPEA), which combines BOA and SPEA to improve SPEA's performance. The experimental tests on the 0/1 multi-objective knapsack problem show that the new algorithm has the stronger capability of multi-objective optimization than SPEA. Thirdly, in order to overcome the disadvantages of single objective QoS-based multicast routing algorithms based on evolutionary algorithms, a multi-objective QoS-based multicast routing algorithm based on Bayes with decision graphs is proposed, which can optimize simultaneously multiple QoS parameters. The experimental results show that the new algorithm is capable of converging quickly to a set of nondominated solutions.