The Improvement Of Teaching-learning Based Optimization Algorithm And Its Application

Teaching-Learning-Based Optimization Algorithm(TLBO) is a kind of new intelligent Optimization Algorithm, which is proposed by Indian scholar Rao etc in 2011. Because of its simple structure, easy to understand, less parameters, strong ability of convergence and good global search capability, which has been successfully applied in many engineering problems, such as continuous large-scale nonlinear optimization, single objective optimization problem such as plane steel frame optimization design; Economic emission load dispatch and multi-objective optimization problem.In this paper, the exploration ability of the TLBO algorithm is poor, which easily plunged into local optimum, TLBO algorithm principle and theory for the further research, puts forward several improvement strategies, and applied to the design of IIR filter and power economic dispatch problem. In this paper, the main research work can be summarized as follows:(1) A new teaching-learning-based optimization algorithm with a multi-learning strategy(MTLBO) is proposed to improve the poor local search capacity of the basic TLBO which the solutions likely to be trapped around the local optima. The opposition-based learning is introduced to broaden the search space and increase the diversity of solutions to improve the global search ability of the algorithm. The employment of multi-learning strategy makes local search more effective to speed up the convergence. To increase the likelihood of solutions jumping out of the areas around the local optima, a stochastic mutation strategy is conducted with a small probability. Experiments are carried out on the test functions and comparisons are made with basic TLBO、I-TLBO and other methods, The results demonstrate that the proposed MTLBO algorithm achieved better performance on both low and high dimensional functions.(2) Owing to the basic TLBO algorithm couldn’t find the global optimal value in solving nonlinear multimodal problems and optimal low accuracy, an improved optimization of teaching and learning(MTLBO) algorithm is proposed, which used for the design of IIR digital filters. In the proposed algorithm, the opposition-based learning technology is introduced, which can broaden the search space and increase the diversity of solutions, to improve the global search ability of the algorithm and to avoid the possibility of algorithm falls into local optimum effectively. The teaching factor is modified, which can effectively balance the global search and local search ability of the algorithm, avoid the invalid iteration of the algorithm at the beginning of the search. The simulation results obtained for two well known benchmark examples both same order and reduced order models of the IIR filters justify the efficacy of the proposed MTLBO algorithm.(3) In order to better balance the exploration and exploitation, a new improved TLBO optimization algorithm is proposed, which used adaptive teaching factor. not only to the poorer individuals to better learning, but also has the choice of the optimal individual has a better to its adjacent individual to study, to improve the level of their own. Also uses the elite strategy to archive of populations history optimal, to keep excellent individuals, to accelerate the convergence speed of the algorithm. By solving the power economic and emission dispatch problem, the example simulation results demonstrate the effectiveness of the algorithm.