Optimization algorithm based on kinetic-molecular theory
来源期刊：中南大学学报(英文版)2013年第12期
论文作者：FAN Chao-dong(范朝冬) OUYANG Hong-lin(欧阳红林) ZHANG Ying-jie(张英杰) AI Zhao-yang(艾朝阳)
文章页码：3504 - 3512
Key words：optimization algorithm; heuristic search algorithm; kinetic-molecular theory; diversity; convergence
Abstract: Traditionally, the optimization algorithm based on physics principles has some shortcomings such as low population diversity and susceptibility to local extrema. A new optimization algorithm based on kinetic-molecular theory (KMTOA) is proposed. In the KMTOA three operators are designed: attraction, repulsion and wave. The attraction operator simulates the molecular attraction, with the molecules moving towards the optimal ones, which makes possible the optimization. The repulsion operator simulates the molecular repulsion, with the molecules diverging from the optimal ones. The wave operator simulates the thermal molecules moving irregularly, which enlarges the searching spaces and increases the population diversity and global searching ability. Experimental results indicate that KMTOA prevails over other algorithms in the robustness, solution quality, population diversity and convergence speed.
FAN Chao-dong(范朝冬)1, OUYANG Hong-lin(欧阳红林)1, ZHANG Ying-jie(张英杰)2, AI Zhao-yang(艾朝阳)3
(1. College of Electrical and Information Engineering, Hunan University, Changsha 410082, China;
2. College of Information Science and Engineering, Hunan University, Changsha 410082, China;
3. Institute of Cognitive Science, Hunan University, Changsha 410082, China)
Abstract:Traditionally, the optimization algorithm based on physics principles has some shortcomings such as low population diversity and susceptibility to local extrema. A new optimization algorithm based on kinetic-molecular theory (KMTOA) is proposed. In the KMTOA three operators are designed: attraction, repulsion and wave. The attraction operator simulates the molecular attraction, with the molecules moving towards the optimal ones, which makes possible the optimization. The repulsion operator simulates the molecular repulsion, with the molecules diverging from the optimal ones. The wave operator simulates the thermal molecules moving irregularly, which enlarges the searching spaces and increases the population diversity and global searching ability. Experimental results indicate that KMTOA prevails over other algorithms in the robustness, solution quality, population diversity and convergence speed.
Key words:optimization algorithm; heuristic search algorithm; kinetic-molecular theory; diversity; convergence
