Multi-Stage Sensor Weapon Target Assignment Problem based on Modified MOEA/D

In this paper, we study a novel sensor weapon target assignment (SWTA) problem, called the multi-stage SWTA problem, and the multi-stage refers to the division of the operational process into several interception stages. The objective of the multi-stage SWTA is to maximize the total value of destroyed targets by using as few sensors and weapons as possible. To solve this problem, a modified multi-objective evolutionary algorithm based on decomposition (MOEA/D) has been designed. First, a constructive heuristic mechanism based on the marginal benefit is introduced to ensure the feasibility of the initial population. Second, a novel penalty value is added to the penalty-based boundary intersection to expand the choice of solutions, and combined with the non-dominated solutions selection to improve the convergence of the algorithm. Third, a mechanism for deleting invalid genes is designed to transform infeasible solutions into feasible ones. Experimental results show that compared with existing algorithms, the proposed method provides a high-quality multi-stage SWTA scheme in terms of solution accuracy, convergence, and diversity performance.