关键词: 杀伤网, 博弈对抗, 设计空间探索, 博弈论, 矩阵博弈, 双重预言算法

Abstract: Aiming at the lack of effectiveness of the existing kill-web design method based on unilateral optimization in game confrontation, and the bottleneck problem caused by the huge strategy space after the introduction of game mechanism, a large space strategy exploration and scheme optimization method for killing net design driven by game confrontation is proposed. In order to effectively model the game confrontation of kill-web design, combined with OODA cycle theory, considering three types of equipment, reconnaissance, accusation and strike, the strategy space and strategy constraints of kill-web game are designed. The mechanism of enemy strike behavior leading to equipment accuracy weakening and then reducing combat effectiveness is introduced to quantify the impact of game on the revenue of both sides, so as to establish a matrix game model of kill-web design. Aiming at the problem that the strategy space of both sides in the model is huge, which makes it difficult for the traditional game solution method to realize the efficient exploration and optimization of the scheme, an Improved Double Oracle with Simulated Annealing (DO-SA) algorithm is designed. The algorithm combines the strategy pool iteration mechanism of the double oracle algorithm and the global search ability of the simulated annealing algorithm, which can effectively explore the mixed strategy Nash equilibrium in the large space game and carry out the efficient optimization of the kill-web design scheme. The case verification results show that the proposed method can realize the efficient solution of the optimal scheme of the kill-web design in the game confrontation scenario. Compared with the traditional unilateral optimization algorithm, it significantly improves the expected revenue of the strategy, and provides theoretical support and decision-making basis for the design of the kill-web in the actual system confrontation.

Key words: kill-web, game confrontation, design space exploration, game theory, matrix game, double oracle algorithm