Evaluation of Underwater Target Scheme Based on Improved Back Propagation Neural Network

Abstract

It is important to evaluate the underwater target scheme scientifically for underwater weapons’ test. Many traditional methods have too many human factors on weight. In this paper, we have built the evaluation index system of underwater target scheme first, and then study the relationship between secondary indexes and evaluation result directly by an improved Back Propagation (BP) neural network. In order to settle the problem that BP neural network is apt to get local optimum, we have optimized the initial weight values and thresholds values by Particle Swarm Optimization (PSO) algorithm. According to compare with the result got by Analytical Hierarchy Process, we have testified that the improved BP neural network can settle the problem better.