Privacy-preserving distributed assignment and guidance for multi-defender-attacker interception

This paper proposes a privacy-preserving, fully distributed autonomous assignment and guidance-interception scheme for large-scale weapon-target assignment (WTA) problems with stringent security/privacy. We first, grounded in optimal control theory and targeting precision guidance, derive an analytical relationship between the optimal interception control law and individual guidance energy consumption. The effect of guidance time on global energy usage is then quantified, thereby obtaining an analytical expression of low energy consumption that describes the energy-matching relationship between defenders and attackers. Building on this foundation, we formulate a multi-defender-attacker (MDA) assignment model. Based on the primal-dual method, we develop a differentially private distributed optimization algorithm for the MDA problems. The proposed algorithm not only guarantees convergence to the optimal solution but also prevents adversaries from inferring the local cost coefficients of the objective function, thereby enhancing system security. Extensive simulations compare the proposed distributed strategy against centralized strategy and heuristic-based distributed strategy, demonstrating that our method consistently attains the global optimum across problem scales and effectively safeguards data privacy throughout the communication process.