Research on Mathematical Genetic Algorithms Based on Evolutionary Computer Neural Network of Distribution Function Parameters

Abstract

Genetic algorithm and neural network are both ideas and algorithm skills. The flexibility of this method provides a stage for many researchers and applications. This paper discusses the neural network model based on the improved genetic algorithm under the parameters of the distribution function and its application in multi-variable, multi-step nonlinear economic forecasting. First of all, this article discusses the method of using improved genetic algorithm to learn neural network connection weights, and proposes a secondary optimization strategy. The main innovations or improvement techniques include: initial population generation strategy, crossover operation secondary optimization strategy, mutation operation Secondary optimization strategy, optimal chromosome inventory strategy, etc. This article uses VB for programming. The developed software has a certain versatility and can be used for prediction problems similar to nonlinear economic data.