System architecture construction is key to the effectiveness of equipment system-of-systems(ESOS). It defines the structure of each component and the relationships between them. However, existing research still exhibits limitations in addressing challenges such as node failures, resource constraints, and environmental uncertainty. To address these issues, an adaptive construction model of ESOS is established with the goal of maximizing operation effectiveness, subject to constraints such as operation loop(OL) relationships, resource allocation, and closure time. First, to simulate scenarios of node failures caused by hostile interference or attacks, an environment with dynamic node availability is constructed. Second, an adaptive construction framework for ESOS is proposed, which combines the Proximal Policy Optimization (PPO) algorithm to achieve rapid construction of ESOS. Finally, the effectiveness of the proposed method is validated through comparative experiments. Experimental results demonstrate that under node failure rates ranging from 10% to 70%, the ESOS constructed by the proposed algorithm significantly outperforms other baseline methods in terms of operation effectiveness. Moreover, the success rate remains stably above 60%, highlighting the robustness and adaptability of the proposed approach.
Liu et al. (2025) studied this question.