Abstract: This paper puts forward the improvement of the traditional simulated annealing algorithm， according to the problem of slow convergence speed.It is improved by using the global harmony search algorithm,which keeps the simulated annealing’s original mechanism,and uses a function to randomly generate the initial solution of the simulated annealing algorithm,thus we to retain the middle optimal solution and prompt update is the advantage of the algorithm.To guarantee the quality of the optimization algorithm and enhance the search efficiency,we use the simulation benchmark test function.The simulation results show that this improved method in convergence speed and optimization quality outperform the traditional simulated annealing algorithm and other algorithms.It is worthy of further study.

Key words: simulated annealing algorithm, global harmony search, improvement, convergence rate