Research on Harmony Search BFGS Hybrid Parallel Algorithm Based on Cloud Computing

Abstract

In order to solve the problems of traditional harmony search algorithm such as slow convergence speed and low solution precision, this paper proposes the harmony search BFGS (Broyden-Fletcher-Goldfarb -Shanno) hybrid parallel algorithm based on cloud computing. The generated evolutionary harmony solution vector is used as the initial vector of BFGS algorithm, which then goes through the refining and optimization computing by BFGS algorithm, thus the hybrid parallel algorithm combining the harmony search and BFGS algorithm is constructed, the Map Reduce parallel programming model is used to design the Map and Reduce parallel functions of optimization computing on the cloud platform to construct the hybrid search BFGS hybrid optimization algorithm based on the cloud platform. It performs the programming and comparative experiment on Hadoop platform. The experimental results show that the proposed algorithm outperforms the harmony search algorithm in optimization precision, convergence speed and acceleration ratio, and its parallel computing is close to the linear acceleration ratio.