Monitor system and Gaussian perturbation teaching–learning-based optimization algorithm for continuous optimization problems

Abstract

In this paper, an improved teaching optimization algorithm called monitor system and Gaussian perturbation (GP) teaching–learning-based optimization algorithm (MG-TLBO) is proposed based on several modified variants of TLBO. TLBO is simply divided into two phases: “Teacher phase” and “Learner phase.” To further improve the solution accuracy and efficiency, we introduce two mechanisms in the learner phase, namely, monitor system and self-regulated learning (SRL) theory. In the learner phase, we assume that the monitor is the most outstanding individual in the population and possesses self-learning ability to expand his or her own strengths. In addition, GP is deployed to model the SRL process. Therefore, three different versions of MG-TLBO are proposed and related experiments are carried out. The results show that all three MG-TLBOs are more effective than the original TLBO. Finally, comparison of the experimental results with other representative meta-heuristics confirms the validity of the new MG-TLBO. In particularly, the MG-TLBO exhibits an overwhelming advantage over the TLBO, which indicates that the MG-TLBO well balances the exploration and exploitation behavior. All the aforementioned evidence manifests that the MG-TLBO improves the accuracy and efficiency of the solution of the original TLBO.