Neural network forecast under the organic hybrid model of genetic algorithm and particle swarm algorithm

Abstract

This paper proposes a organic hybrid model of the genetic algorithm and the particle swarm algorithm firstly, then establishes the multi-factor time series forecasting model, designs the BP neural networks, adopts the organic hybrid model of genetic algorithm and particle swarm algorithm to optimize the weight from the input layer to the hidden layer, the weight from the hidden layer to the output layer and the number of neuron nodes in the hidden layer. Finally, it carries on the training with the related power consumption data in 1980–2000 in China, then obtains the network model of the nonlinear relationship between the influencing factors and the power consumption, and forecasts the electricity consumption in 2001–2005, the average absolute error rate of the forecast is 12.08%. The results show that the neural network forecasting model optimized by the organic hybrid of the genetic algorithm and the particle swarm algorithm is not precocious, and it has a high search efficiency, which also makes the accuracy of the power consumption forecast much improved. Thus, the hybrid model can be regarded as an effective method in optimizing the neural network.