求解物流Web服务组合问题的两阶段多目标蚁群算法

针对基于QoS的物流Web服务组合优化问题，提出了两阶段多目标蚁群优化(TMACO)算法。首先，针对原始数据集中存在被支配候选服务而增加算法求解时间的问题，提出了基于Pareto支配的预优化策略；其次，针对属性权重难以确定的问题，提出了不依赖权重的信息素更新策略和启发信息策略；最后，针对基础蚁群算法容易陷入局部最优的问题，提出了懒蚂蚁策略。实验结果表明，TMACO算法具有良好性能，相对于基础蚁群算法、利用解与理想解距离来更新信息素的改进蚁群算法、遗传算法以及用支配程度作为解的个体评价的改进遗传算法，TMACO算法有更高的寻优能力，能够找到更多更优的非劣解。

Two-stage Multi-objective Ant Colony Optimization for Solving Logistics Web Service Composition

A two-stage multi-objective ACO(TMACO) was proposed to solve logistics Web services composition optimization problems. First, to solve the time increases rised by candidate services which was dominated in the raw data, a pre-optimization strategy was proposed based on Pareto dominated ideology; secondly, since the weights of each attribute were difficult to determine, a non-dependent weight pheromone update strategy and inspiration information policy were included in the TMACO; and finally, the basic ACO was easy to fall into local optimum problem, a lazy ant strategy was proposed to solve this problem. The experimental results show that TMACO algorithm has good performance. Its optimization capability is better than that of the basic ACO,the improved ACO which used the distance of the solution and the ideal solution to update the pheromone,the basic genetic algorithm and improved genetic algorithm which applied the dominant factor to evaluate a solution, TMACO optimization algorithm has a higher ability to find more and better Pareto solutions.