Prediction model of sintering burden based on information entropy and chaos PSO algorithm

Considering the characteristics of the nonlinear, complexity and relativity of the sintering burden system, the prediction model of sintering burden is established by BP neural network. In addition, a new optimization method of the sintering experiment is proposed, based on information entropy and chaotic improved particle swarm algorithm. The initial particle colony is produced by information entropy to increase the variety of the initial colony. The strategy of dynamic nonlinear adjustment is used for the inertia weight in this paper according to the iteration times, so as to improve the algorithm´s searching capability. And the traversal characteristic of chaos optimization is introduced to overcome effectively the local convergence of standard particle swarm algorithm. The simulating results show that the improved particle swarm algorithm has faster converge, fewer iteration times and stronger global optimization ability.