Multi-objective Integrated Optimization Using Optimization, Modeling and Simulation
Multi-objective Integrated Optimization Using Optimization, Modeling and Simulation
复制标题
DOI：
10.1109/smc.2013.603
复制
发表时间：
2013-10
期刊：
影响因子：
--
通讯作者：
H. Katayama;K. Tamura;K. Yasuda
中科院分区：
文献类型：
--
作者：
H. Katayama;K. Tamura;K. Yasuda
本文提出了一种将优化方法、建模和仿真技术有机结合的实用多目标优化框架。将多目标差分进化与径向基函数网络相结合，建立了多目标集成优化框架。该框架可用于减少对具有高计算负荷的模拟器或传感系统的访问次数。通过对典型基准问题的数值实验表明，所提出的多目标综合优化方法可以得到较好的Pareto解，并且大大减少了评价系统性能指标值的函数调用次数。
In this paper, the authors propose a new practical multi-objective optimization framework that combines optimization method, modeling and simulation technologies organically. The new framework is called Multi-Objective Integrated Optimization that combines Multi-Objective Differential Evolution and Radial Basis Function Network. This new framework is used to reduce the number of accesses to a simulator or a sensing system with heavy computational load. According to the numerical experiment on typical benchmark problems, it is shown that the proposed Multi-Objective Integrated Optimization obtains good Pareto solutions with drastic reduction in the number of function calls for evaluating the performance index values of systems.