An Improved Back Propagation Neural Network Forecasting Model Using Variation Fireworks Algorithm for Short-time Traffic Flow

摘要: Short-term traffic flow has complicated non-linear characteristic, and it a similar seasonality. This paper presents an improved Back Propagation neural network for Short-time prediction using Variation Fireworks Algorithm. In the new model, we use self-adaption explosion amplitude to resolve contradiction between local search global while maintaining good properties of better fireworks. Furthermore, this self-variation mutation operator, can improve features For increasing diversity, introduce random mapping out-of-boundary sparks instead modulo operator. It is applied real-world data collected from Second Ring Road Beijing, China compared with three models. The consequences show that model higher accuracy be served as forecasting short-term flow. And shows result than (BP) Particle Swarm optimization (PSO-BP) Algorithm (FWA-BP) model.