Optimization design of terminal sensitive projectile based on neural network and ant colony algorithm

The overall optimization design of the terminal sensitive projectile is a complicated and difficult work because the complex structure of projectile and multiple influencing factors. Several main parameters influencing the system effectiveness of terminal sensitive projectile were optimized with Back-Propagation Neural Network (BPNN) and Ant Colony Algorithm (ACA) in this paper. Firstly, the basic theories of Neural Network and Ant Colony Algorithm were described. Then the work process of terminal-sensitive projectile was given, and several influencing parameters of system effectiveness were analyzed. Based on the nonlinear relationship between parameters and effectiveness, the BPNN with three layers was established, and was trained with data samples. Using the trained BPNN as adaptive function, the optimization model was built on the basis of the ant colony algorithm theory. Finally, the globally optimal values of parameters were obtained by numerical simulating computation. The results show that the combined BPNN and ACA is an effectual way for the optimization design of terminal sensitive projectile.