New Teaching-Learning-Based Optimization Algorithm with Course Selection

Abstract

The teaching-learning-based optimization (TLBO) is a new metaheuristic, which searches the best result of NP hard problem with the teaching and learning process. TLBO algorithm is efficient and robust. However, the original TLBO is easy to trap in local optimum and lack of global search capability. To get a better global performance, this paperwork proposes an improved TLBO named graduate teaching learning based optimization algorithm (GTLBO). Distinguished from one teacher in the basic TLBO with only one teacher, GTLBO has multiple teachers, each teacher leads a group of student and does research in different fields. Student is allowed to change their research status when they get no advance for a long time. The student giving up his fields may not be able to find a new result for a long time. That requires another bad-field-tolerant mechanism. This algorithm also proposes a new way to ensure all the field have searching value. GTLBO also has a good local search capability by controlling the number of student in each field. The GTLBO is tested on benchmark functions and got a better result compare with other improved TLBO algorithm.