Analysis of Energy Consumption Prediction Model Based on Genetic Algorithm and Wavelet Neural Network

摘要

This paper analyzed the enterprise process energy consumption systematically with a lot of statistic data starting from energy efficiency, and established the energy consumption prediction model based on genetic algorithm of wavelet neural network (GA-WNN). This paper made previous optimization training with genetic algorithm, which have feature of natural evolution regularity, to the weights and dilation-shift scale of wavelet neural network. Partly replaced gradient descent method of wavelet frame neural network where parameters optimization only with a single gradient direction, overcame the shortcoming that easily into the local minimum and cause oscillation effect of the single gradient descent method. Simulation results showed the effectiveness of the forecasting model, and it is feasible for solving the process energy consumption multi-factor quantitative analysis problem which general mathematical model is difficult to describe.