The concept of unmanned weapon system-of-systems (UWSoS) involves a collection of various unmanned systems to achieve or accomplish a specific goal or mission. The mission reliability of UWSoS is represented by its ability to finish a required mission above the baselines of a given mission. However, issues with heterogeneity, cooperation between systems, and the emergence of UWSoS cannot be effectively solved by traditional system reliability methods. This study proposes an effective operation-loop-based mission reliability evaluation method for UWSoS by analyzing dynamic reconfiguration. First, we present a new connotation of an effective operation loop by considering the allocation of operational entities and physical resource constraints. Then, we propose an effective operation-loop-based mission reliability model for a heterogeneous UWSoS according to the mission baseline. Moreover, a mission reliability evaluation algorithm is proposed under random external shocks and topology reconfiguration, revealing the evolution law of the effective operation loop and mission reliability. Finally, a typical 60-unmanned-aerial-vehicle-swarm is taken as an example to demonstrate the proposed models and methods. The mission reliability is achieved by considering external shocks, which can serve as a reference for evaluating and improving the effectiveness of UWSoS.

Vital nodes play a pivotal part of network structure and dynamics, where finding the minimal size of a set of vital nodes belongs to an NP-hard problem and cannot be solved by a polynomial algorithm. Recent studies of vital nodes identification mostly rely on the structural information, such as collective influence and degree. However, the performance of local-based methods varies for different structure while the complexities of global-based methods are generally high for most situations. Here, we map the problem into an optimization issue based on global information of network structure and propose a belief propagation and node reinsertion (BPR) method with almost linear time complexity, where finding the minimum feeding back vertex set is a key. Compared with several state-of-the-art heuristic methods, the BPR method has advantages of high accuracy and practicability of vital nodes identification and low computational complexity. Under two attack schemes: static and dynamical, extensive experiments of Erdős–Rényi (ER) and Scale-Free (SF) models and real-world networks of the power grid and traffic network convincingly demonstrate that BPR method remarkably outperforms other methods in vital nodes identification. This helps to reassess the operational risk of a network and improve robustness ranging from network design schemes, protection strategies to failure mitigation.

Due to possible disturbances in a conflict environment, the mission performance of flying ad hoc networks (FANET) is highly related to time-varying topology. Currently, the resilience metric is regarded as suitable for evaluating the mission performance of a FANET with time-varying network topology. However, the reported resilience studies contribute little to the theoretical analysis of the impact of the time-varying topology of FANET on a cooperative mission. In this article, an improved resilience analysis method is proposed by defining a new dimension reduction operator, which takes into account the impacts of both isolated and communicated unmanned aerial vehicles. Then, based on random geometry theory, the topological features of FANET are extracted to measure the structure perturbation and recovery. A case study of a joint reconnaissance mission shows that our analysis method is more realistic than the typical method. Our results may be helpful to analyze the resilience of networked dynamic systems and support the more flexible topology control methods design of a FANET.

Heterogeneity in the load capacity of nodes is a common characteristic of many real-world networks that can dramatically affect their robustness to cascading overloads. However, most studies seeking to model cascading failures have ignored variations in nodal load capacity and functionality. The present study addresses this issue by extending the local load redistribution model to include heterogeneity in nodal load capacity and heterogeneity in the types of nodes employed in the network configuration and exploring how these variations affect network robustness. Theoretical and numerical analyses demonstrate that the extent of cascading failure is influenced by heterogeneity in nodal load capacity, while it is relatively insensitive to heterogeneity in nodal configuration. Moreover, the probability of cascading failure initiation at the critical state increases as the range of nodal load capacities increases. However, for large-scale networks with degree heterogeneity, a wide range of nodal load capacities can also suppress the spread of failure after its initiation. In addition, the analysis demonstrates that heterogeneity in nodal load capacity increases and decreases the extent of cascading failures in networks with sublinear and superlinear load distributions, respectively. These findings may provide some practical implications for controlling the spread of cascading failure.

Distributed artificial intelligence is increasingly being applied to multiple unmanned aerial vehicles (multi-UAVs). This poses challenges to the distributed reconfiguration (DR) required for the optimal redeployment of multi-UAVs in the event of vehicle destruction. This paper presents a multi-agent deep reinforcement learning-based DR strategy (DRS) that optimizes the multi-UAV group redeployment in terms of swarm performance. To generate a two-layer DRS between multiple groups and a single group, a multi-agent deep reinforcement learning framework is developed in which a QMIX network determines the swarm redeployment, and each deep Q-network determines the single-group redeployment. The proposed method is simulated using Python and a case study demonstrates its effectiveness as a high-quality DRS for large-scale scenarios.