Parallel performance analysis of nonlinear equations algorithm based on hybrid genetic algorithm

Abstract

For a long time, people have done a lot of research on nonlinear equations in theory and numerical calculation. In this paper, the basic theories and methods of genetic algorithm and traditional algorithm are expounded, and the advantages of genetic algorithm and quasi-Newton algorithm are analyzed, thus a new hybrid genetic algorithm for solving nonlinear numerical problems is proposed, and the effectiveness of the algorithm is verified by numerical examples. The hybrid algorithm gives full play to the group search and global convergence of genetic algorithm, and effectively overcomes the initial point sensitivity problem of classical algorithm At the same time, the classical algorithm is introduced into genetic algorithm for local search, which overcomes the shortcomings of slow convergence speed and poor accuracy of genetic algorithm. The algorithm in this paper provides an effective way to solve nonlinear equations from another angle.