全局智能优化集成算法研究

Integrated Intelligent Algorithms for Global Optimization

针对遗传算法、模拟退火算法等智能全局优化算法的集成问题开展研究,分析归纳了智能全局优化算法和局部搜索算法的一般规律和特性,给出了全局智能优化算法进行集成的统一框架——全局智能优化集成算法(IGIOA),及IGIOA的设计要素,还给出了评价算法的优化性能指标、时间性能指标、鲁棒性能指标,以及将三指标综合的综合性能指标,为智能集成算法的选取和性能比较提供了依据.

The systematical integrating of the intelligent algorithms for global optimization such as genetic algorithm, simulated annealing algorithm and so on is discussed. The properties and characteristics of these intelligent algorithms and local search algorithms for optimization are ana- lyzed respectively. A unified structure for a class of integrated intelligent algorithms for global optimization, IGIOA, is given, and some key factors in designing a particular integrating intelligent algorithm are presented. Some indices for evaluation and comparison of the integrated intelligent algorithms are proposed involving the evaluations of algorithms for optimization, time cost, and robustness. And a weighted index to combine these three evaluations used in selecting and comparing integrating intelligent algorithms is presented.