Artificial neural network model based on improved genetic algorithm combined with simulated annealing algorithm for prediction of deformation

This paper presents the effectiveness of the artificial neural network (ANN) based on improved genetic algorithm (GA) combined with simulated annealing algorithm (SA) and its application to prediction of deformation. The paper brings forward the concept of distance that includes absolute distance and relative distance. By applying distance, the new algorithm can control parameters of GA that has self-adaptive quality to avoiding premature. Standard GA algorithm, improved GA, and improved GA combined with SA are also involved for a comparison purpose. The way of determining the topological structure of ANN is also discussed. The experiment on time series data shows that ANN based on the improved GA combined with SA has superiority to ANN based on gradient descent.