摘要
针对教与学优化算法(TLBO)局部开发能力差,易陷入局部最优的缺点,提出一种基于反向学习的多学习教与学优化算法(MTLBO)。通过反向学习技术拓展搜索空间,增加解的多样性,进一步增强算法的全局搜索能力。引入多学习机制,使其更有效地进行局部搜索,加快收敛速度。同时提出一种小概率变异策略,增加跳出局部最优的可能性。在基准测试函数上进行验证实验,结果表明,与TLBO算法、ITLBO算法以及其他优化算法相比,该算法在低维和高维函数上都取得了较好的优化效果。
We proposed a teaching-learning-based optimisation algorithm （TLBO） with multi-learning mechanism （ MTLBO）, which is based on the opposition-based learning, aimed at the drawbacks of basic TLBO in poor local development ability and solutions being prone to falling into local optima. It broadens the search space through opposition-based learning technology to increase the diversity of solutions and further improve the global search ability of the algorithm. The introduction of multi-learning mechanism makes the local search more effective i and speeds up the convergence rate. Meanwhile we presented a small probability mutation strategy to increase the likelihood of solutions jumping out of local optima. Verification experiments were carried out on benchmark test functions, results indicated that compared with basic TLBO, ITLBO and other optimised algorithms, the proposed MTLBO algorithm achieved better optimisation effect on both low and high dimensional functions.
出处
《计算机应用与软件》
CSCD
2016年第2期246-249,298,共5页
Computer Applications and Software
基金
新疆"十二五"制造业信息化科技示范工程项目(201130110)
关键词
教与学优化算法
反向学习技术
多学习机制
变异策略
Teaching-learning-based optimisation algorithm Opposition-based learning technology Multi-learning mechanism Mutation strategy
