Abstract: Considering the nonlinear characteristics of competitive power evaluation and the existence of many nonquantitative factors in the evaluation system, a Hopfield neural network model of evaluation on competitive power was built in this paper according to the evaluating data given by the experts. By utilizing the properties of association and memory of neural network that can approximate any nonlinear continuous function, the model can calculate quantified cvaluation and has higher efficiency than that of existing evaluation system. The method of applying Hopfield neural network to competitive power evaluation can overcome subjective factors in the process of expert assessment, avoid complex process of traditional competitive power evaluation and also has broad applicability.

Key words: Evaluation index system, Evaluation of competitive power, Hopfield neural network