An autonomous teaching-learning based optimization algorithm for single objective global optimization

Teaching-learning based optimization is a newly developed intelligent optimization algorithm. It imitates the process of teaching and learning simply and has better global searching capability. However, some studies have shown that TLBO is good at exploration but poor at exploitation and often falls into local optimum for certain complex problems. To address these issues, a novel autonomous teaching-learning based optimization algorithm is proposed to solve the global optimization problems on the continuous space. Our proposed algorithm is remodeled according to the three phases of the teaching and learning process, learning from a teacher, mutual learning and self-learning among students instead of two phases of the original one. Moreover, the motivation and autonomy of students are considered in our proposed algorithm, and the expressions of autonomy are formulated. The performance of our proposed algorithm is compared with that of the related algorithms through our experimental results. The results indicate the proposed algorithm performs better in terms of the convergence and optimization capability.