{"title":"Global numerical optimization using multi-agent genetic algorithm","authors":"Zhong Wei-cai, Liu Jing, Xu Mingzhi, Jiao Licheng","doi":"10.1109/ICCIMA.2003.1238119","DOIUrl":null,"url":null,"abstract":"A new algorithm, Multi-Agent Genetic Algorithm (MAGA), is proposed. It realizes the complex global numerical optimization via agent-agent interactions. All agents are fixed on a lattice, and they will compete or cooperate with their neighbors to increase their own energy. On the other hand, agents can also increase their energy with knowledge. In experiments, 4 multimodal benchmark functions are used to explore the effect of problem of problem dimension on the performance of MAGA. The results on functions with 20/spl sim/10,000 dimensions show that MAGA obtains good performance in solving high dimensional functions. Even when dimension is as high as 10,000, MAGA can still find high quality solutions with very low computational cost.","PeriodicalId":385362,"journal":{"name":"Proceedings Fifth International Conference on Computational Intelligence and Multimedia Applications. ICCIMA 2003","volume":"32 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2003-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"8","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings Fifth International Conference on Computational Intelligence and Multimedia Applications. ICCIMA 2003","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICCIMA.2003.1238119","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 8
Abstract
A new algorithm, Multi-Agent Genetic Algorithm (MAGA), is proposed. It realizes the complex global numerical optimization via agent-agent interactions. All agents are fixed on a lattice, and they will compete or cooperate with their neighbors to increase their own energy. On the other hand, agents can also increase their energy with knowledge. In experiments, 4 multimodal benchmark functions are used to explore the effect of problem of problem dimension on the performance of MAGA. The results on functions with 20/spl sim/10,000 dimensions show that MAGA obtains good performance in solving high dimensional functions. Even when dimension is as high as 10,000, MAGA can still find high quality solutions with very low computational cost.
京公网安备 11010802042870号
京ICP备2023020795号-1
分享
求助内容：
应助结果提醒方式：
扫码关注我们
