为解决具有多个需求节点和多个服务中心的服务网络,如何在预算约束条件下设计与优化服务网络的服务中心配置,使得服务网络运营的总效率最高的决策问题,本文对多个服务中心的服务网络优化问题进行了提炼和描述,并构建了服务网络设计的优化模型,同时证明了该服务网络优化问题是一个NP-完全问题;进一步,针对优化模型的特点,设计了求解模型的混合拆分遗传算法,进行了大规模仿真实验并与传统的多目标遗传算法进行了比较,结果表明本文给出的算法具有较好的求解效率和效果;最后,通过例子说明了本文提出方法的潜在应用价值.
This paper addressed the decision problem of how to locate service centers in a service network with multiple demand nodes to achieve the highest operation efficiency with certain constraints. This work firstly described and formulated the problem of the design and optimization for a multiple-service-center network. It then built an optimization model to solve this problem. The mode is proved to be NP-complete. Furthermore, it developed a hybrid partition genetic algorithm (HPGA) to solve this model in light of the characteristics of this model. Extensive computational experiments were conducted to compare the HPGA with the multi-objective genetic algorithm (MOGA). Much better performance of the proposed algorithm was observed in effectiveness and efficiency. Additionally, an example was used to illustrate the potential application value of the proposed method.