Analysis and Design of Metaheuristics Based on Proximate Optimality Principle

包含启发式操作的优化方法称为元启发式。为了改进传统的元启发式算法并构建新的元启发式算法，本文基于近似最优性原理分析了元启发式算法的结构，即最优化问题的优解具有相似的结构。在数值分析的基础上对元启发式算法进行了改进和构建新的元启发式算法，并利用已知的基准问题进行了仿真，验证了所开发方法的性能。

The optimization method including heuristic operations is called metaheuristics. For the purposes of improving conventional metaheuristics as well as constructing new metaheuristics, the structure of metaheuristics is analyzed in this study based on proximate optimality principle which implies that good solutions in optimization problems have similar structure. While the improvement of metaheuristic and the construction of new metaheuristic are done on the basis of the numerical analysis, the performance of the developed method is verified by simulations using well known benchmark problems.