Solving the Dynamic Sensor/Weapon-Target Assignment Problem by Generation Strategy Optimization

Abstract

In the background of warship formation air defense, the Dynamic Sensor/Weapon-Target Assignment(DWSTA) has difficulty in dealing with complex constraints. Aiming at this, we put forward an indirect optimization method. We convert the problem with complicated constraints to an assignment generation strategy optimization problem with fewer constraints. The sensor/weapon target allocation is optimized indirectly through the optimization of the generation strategy. At the start of our method, the feasible assignment schemes for each target are generated according to various constraints. Every scheme is composed of the target, radar, and weapon. Then we choose the current optimal assignment(“current” means that the assignment will be optimized in the next steps) from them for each target in the order determined by the generation strategy and these assignments constitute the current optimal overall assignments. In this process, a variety of features representing the performance of feasible assignments are calculated and substituted into the evaluation function of features which is determined by the generation strategy. We choose the assignment with the maximum function value as the current optimal assignment for the target. In the process of optimization, the evaluation function value of the current optimal overall assignments is used to the corresponding generation strategy. While the differential evolution is used to optimize generation strategies, the current optimal overall assignments are optimized indirectly. So we encounter fewer constraints. It can be seen from the results that this method can obtain higher quality solutions after the same generations compared with other methods.