Research On Improved Teaching Learning-Based Optimization Algorithm And Its Application

At present,the combination optimization problem exists widely in various fields of the national economy.The solution to the combination optimization problem provides scientific support for economic construction and macro decision-making,and improves the scientific,correct and forward-looking decision-making.However,as the size of the problem increases,the search space of combinatorial optimization problems increases exponentially,and the search terrain becomes more rugged than small size problems.The traditional method is usually slow and the quality of solution is poor.Therefore,the theory of combinatorial optimization problems and the effective solution of combinatorial optimization problems are still central issue in academia.Since the intelligent optimization algorithm usually obtains a high-quality feasible solution from an arbitrary feasible solution,it has become an important method to solve the complex combinatorial optimization problem.Teaching Learning Based Optimization,which is a new type of swarm intelligence optimization algorithm,has received extensive attention due to its unique operating mechanism and relatively good performance.However,the TLBO algorithm still has weaknesses such as low solution accuracy and slow convergence speed.In this thesis,the advantages and disadvantages of TLBO algorithm are analyzed through theory and experiment,and an improved self-excited TLBO algorithm is proposed.The main research contents of this paper are as follows:(1)In this thesis,the TLBO algorithm is analyzed from a theoretical perspective.The performance of the TLBO algorithm and three comparison algorithms are compared on the CEC2017 standard test set through simulation experiments.The experimental results show that the TLBO algorithm is easy to fall into the local optimum in the process of solving high-dimensional problems and multi-peak functions.Besides,the quality of solution which obtained by TLBO algorithm is poor.(2)Because there is no self-learning method in the standard TLBO algorithm,the TLBO algorithm randomly explores the search space according to the current state.Namely,the TLBO algorithm lacks effective self-directing.This paper designs a new selflearning strategy for balancing the global search ability and local search ability of the algorithm.(3)The self-learning strategy is introduces into the standard TLBO algorithm and proposes an improved self-incentive-based TLBO algorithm.In this algorithm,a selflearning strategy is used in the learning phase to achieve self-direction in the optimization process.Simulation results show that the performance of improved self-excited TLBO algorithm is significantly better than the performance of standard TLBO algorithm and other comparison algorithms.