A Self-reflection Teaching-Learning-Based Optimization Algorithm and Its Application

This paper proposes a self-reflection teaching-learning-based (STLBO) algorithm to improve the optimization performance of the TLBO algorithm. A new learning mechanism is employed to enhance the search ability of TLBO algorithm. In the learner phase, a no-best student learns from a random selected outstanding student or an outstanding student learns from teacher again according the fitness of the two mutual learners. After the interaction learning with other students, a self-reflection learning strategy is carried out to realize more detailed search for local information and jump out the local optimal. To evaluate the performance of STLBO algorithm, a comparison with other TLBO variants is presented to address the global optimization problem for robust pole assignment in linear discrete time systems with poles restrict in a circular region, and a robust test with LMI method is carried out. Finally, the simulation results demonstrate the effectiveness and merits of the proposed method.