McCormick 기반 혼합정수선형계획법을 이용한 다층 미사일 방어체계의 동적 무기-표적 할당 해법
This study proposes a McCormick-based mixed-integer linear programming (MILP) model for real-time dynamic weapon-target assignment (DWTA) in multi-layer missile defense systems. The proposed model simultaneously considers engagement time windows (ETWs), time-varying interception effectiveness, and inter-layer dependency based on the shoot-look-shoot (SLS) doctrine. To reflect operational engagement conditions, a K-factor-based effectiveness model is introduced by incorporating trajectory progress, engagement geometry, missile kinematics, and environmental conditions. Since the survival probability formulation results in nonlinear and nonconvex multilinear terms, the problem is reformulated into an MILP framework using binary-tree decomposition and McCormick linearization. In addition, a rolling horizon strategy and warm-start mechanism are applied to support real-time decision making under dynamically changing battlefield conditions. Numerical experiments on various threat scenarios demonstrate that the proposed approach significantly reduces computation time compared to nonlinear optimization methods while maintaining high solution quality. In particular, the proposed method achieves near optimal solutions within sub-second computation time even for large-scale instances.The results indicate that the proposed McCormick-based MILP framework provides an effective and practical solution for real-time DWTA problems in multi-layer missile defense systems.