论文部分内容阅读
为每个服务节点选择一个服务实例,形成在满足全局QoS限制的条件下,具有综合QoS最优的执行计划是服务选择的关键问题.本文分析了优化目标,提出了基于自适应变异遗传算法(self-Adaptive Mutation Genetic Algorithm,AMGA)求解服务选择的组合优化问题.在算法中,适应度函数设计采用对执行计划聚合QoS属性值与约束条件的偏差量作为惩罚约束,变异算子设计采用对服务实例的优良程度自适应的变异概率,提高了基因改良的效率,并引入指数衰减函数保证了算法的收敛性.实验结果表明,基于AMGA算法的服务选择策略比现有基于其它遗传算法的策略能够获得更优的解.
For each service node, a service instance is selected to form an execution plan with comprehensive QoS optimality, which is the key issue of service selection under the condition that the global QoS constraints are satisfied.This paper analyzes the optimization objectives and proposes an adaptive mutation genetic algorithm self-Adaptive Mutation Genetic Algorithm (AMGA) is used to solve the combinatorial optimization problem of service selection.In the algorithm, the fitness function is designed to use the deviation of the QoS attributes and the constraints of the execution plan as the penalty constraint, The goodness of the example adapts to the mutation probability, improves the efficiency of gene modification, and introduces the exponential decay function to ensure the convergence of the algorithm.The experimental results show that the service selection strategy based on the AMGA algorithm can outperform the existing strategies based on other genetic algorithms Get better solution.