An efficient parameter-adaptive genetic algorithm for service selection with end-to-end QoS constraints

With the growing number of Web Services that provide similar functionality but differ in QoS level, the dynamic service composition becomes a decision problem on which component services should be selected under the end-to-end QoS constraints. Local selection is an efficient strategy with polynomial time complexity but can not handle global constraints. Traditional global selection approach based on integer linear programming can handle global constraints, but its poor performance renders it inappropriate for practical applications with dynamic requirements. In this paper, we address this problem and proposes a parameter-adaptive genetic algorithm. First, the problem is modeled as a nonlinear optimization problem under several constraints. Then, we utilize the swarm intelligence search feature of genetic algorithms to search out the global approximate optimal solution quickly. According to the variation of evolution degree and population diversity, we adjust the mutation probability and crossover probability to ensure global convergence and convergence speed. The theoretical analysis and experimental results show the efficiency and feasibility of our approach. Copyright © 2014 Binary Information Press.