Abstract: The key indicator for measuring the combat performance of an aircraft carrier is the sortie rate of carrier-based aircraft, which depends on the support station matching strategy of carrier-based aircraft. Existing works mainly use sequence matching and batch matching methods to match suitable stations for carrier-based aircraft. However, both methods have certain limi-tations, and it is difficult to ensure both real-time and quality of station matching at the same time. Facing the complex and time-varying support environment, it becomes extremely difficult to determine a reasonable support operation matching strategy. Therefore, in this paper, we propose a novel adaptive batch matching decision-making method for carrier-based aircraft support operations based on the batch matching method. First, the optimal time window division strategy is solved by constructing a reinforcement learning method for multi-dimensional environmental state encoding. Then, a highly effi-cient batch matching algorithm is applied within each time window to find the best matching solution for support opera-tions and support stations. The results of multiple sets of simulation experiments based on the publicly available Nimitz aircraft carrier data show that our proposed method can effectively respond to dynamic changes in the support environment, and can quickly solve high-quality support operation assignment plans while meeting real-time requirements.

Key words: Carrier-borne aircraft, Support operations, Real-time scheduling, Reinforcement learning, Case verification