A Hybrid Computational Intelligence Method of Newton's Method and Genetic Algorithm for Solving Compatible Nonlinear Equations

Abstract In order to solve the system of compatible nonlinear equations, author proposes a hybrid computational intelligence method Newton's and genetic algorithm. First, Quasi-Newton Methods (QN) is given. Aiming at local convergence algorithm, it easy cause solution fail. By embedding QN operator in Genetic Algorithm (GA) defining appropriate fitness, thus, algorithm CNLE obtained that combines advantages GA method, which has both faster higher probability solving. Experimental results show that: The value selection p n also directly affects efficiency. Generally speaking, for strong composed multimodal functions, can be larger; For weakly functions with fewer extreme points stronger monotonicity, smaller. It demonstrated this significantly outperforms methods.