Harmony Search Algorithm Based On Improved Bat Search Algorithm

The harmony search algorithm is a heuristic search algorithm for simulating the music creation process.This algorithm does not make full use of the current optimal harmony to perform local optimization in the process of mutation to generate new harmony.This paper proposes a bat harmony hybrid algorithms based on collaborative optimization for this problem.The main contents of this paper are as follows:(1)Introduce the bat local search mechanism for collaborative optimization.The bat population and the harmony population are used for collaborative search.In the process of searching,the current optimal solution is shared with the harmony population and the bat population,and the harmony population is globally searched.The bat population performs local search near the current optimal solution.(2)The learning mechanism is introduced in the process of generating harmony,to optimize the quality of harmony by learning from the current optimal solution.When a new harmony is generated,the new harmony property value learns from the current optimal harmony property value with a certain probability to optimize the quality of the newly generated harmony.(3)Improve the out-of-bounds processing of harmony search algorithms.When the newly generated harmony attribute value is out of bounds,the way of replacing the out-of-bound attribute value with the random number in the original harmony algorithm is blind,Because the current optimal harmony value is better than the random value.The method proposed in this paper use the current optimal harmony attribute value other than random value to replacing the attribute value of the out-of-bounds harmony attribute value.(4)Introduce a dynamic search mechanism.Since the current optimal value is getting closer to the global optimal value during the search process,the frequency of the harmony search is larger in the early stage of the search to improve the quality of the global search,and the search frequency of the bat is increased in the later stage of the search to improve the quality of search result.Finally,six standard test functions are selected in the UCI standard data set for simulation experiments and compared with other harmonic search algorithms.The results of simulation comparison experiments show that the proposed algorithm improves on both convergence speed and optimization accuracy.