Abstract: For quality assurance, it is very important to make effective quality prediction in the stage of product design and parameter optimization. To do so, by using PSO (particle swarm optimization) and BP (back propagation) neural network, a quality prediction model is established. It is an optimization problem with grey incidence degree between the networks output and input as objective. PSO algorithm is used to optimize the BP neural networks weight coefficient and threshold value. Then, a PSO-GRG (grey relational grade) algorithm is proposed to solve the problem. This algorithm overcomes general BP algorithms shortcomings of slow convergence and local optimum solution. A case problem of injection molding is used to verify the proposed method. Simulation results show that the prediction errors are significantly reduced with the number of iterations being reduced by 87.5%.

Key words: particle swarm optimization(PSO) algorithm, back propagation(BP) neural network, quality prediction; grey relational grade