Research And Application Of Harmony Search Algorithm

With the development of science and technology of the computer, traditional numerical methods are replaced by many intelligence optimization algorithms gradually, and intelligence optimization algorithms imitate natural phenomena and become the main methods for solving complex real-word optimization problems. Intelligent optimization algorithm is able to overcome the computational drawback of traditional numerical methods such as computing complex derivative or gradient and sensitive to the initial value, and it quickly finds a better solution. Recently, a global intelligent optimization algorithm was proposed and named harmony search (HS) algorithm. In this paper, the harmony search algorithm is summarized, analyzed, improved and applied research.The present research of harmony search algorithm is firstly and comprehensively summarized in this paper. The optimization performance improvement and application research of harmony search algorithm are introduced and summarized. For the optimization performance improvement, algorithm parameter setting, operator innovation and harmony search algorithm combine with other intelligence algorithms are analyzed, and the significant and representative applications in some domain are introduced, including engineering optimization problem, the electric power system economy dispatch, information network optimization and neural network training, multi-objective optimization and discrete application and other application fields.A learned harmony search algorithm (LHS) is proposed in this paper. LHS adaptively set parameter HMCR and dynamically adjust PAR to make up HS algorithm weakness such as sensitive to the parameter value. In addition, Differential learning operation is introduced to increase the diversity of population and enhance the global search ability. Based on a large number of numerical experiments, the LHS has demonstrated better performance on solving most unconstrained problems when compared to the other three harmony search algorithms. On the whole, LHS algorithm improves the optimization accuracy of HS algorithm. Besides, the effect of parameters on the performance of the LHS is analyzed and investigated to explore the optimization potential of the LHS as much as possible.The LHS algorithm is applied to solve0-1knapsack problem. Ten classic0-1knapsack problems and ten0-1knapsack examples with large dimension to be choose to test the performance of LHS. Experimental results demonstrate that the optimal solutions obtained by the LHS algorithm are all better than those obtained by the other three HS algorithms on solving0-1knapsack problems.In order to further study LHS algorithm can effectively to solve0-1knapsack problem, the effect of parameters on the performance of the LHS is analyzed and the result show that suitable parameter can accelerate LHS algorithm to find the global optimal solution in a certain degree.