An improved teaching-learning-based optimization for constrained evolutionary optimization

摘要: Abstract When extending a global optimization technique for constrained optimization, we must balance not only diversity and convergence but also constraints objective function. Based on these two criteria, the famous teaching-learning-based (TLBO) is improved optimization. To convergence, an efficient subpopulation based teacher phase designed to enhance diversity, while ranking-differential-vector-based learner proposed promote convergence. In addition, how select in rank solutions have significant impact tradeoff between address this issue, dynamic weighted sum formulated. Furthermore, simple yet effective restart strategy settle complicated constraints. By adopting e constraint-handling as technique, evolutionary algorithm, i.e., TLBO (ITLBO), proposed. Experiments broad range of benchmark test functions reveal that ITLBO shows better or at least competitive performance against other TLBOs some algorithms.