In response to the challenges faced by unmanned swarms in mountain obstacle-breaching missions within complex terrains, such as poor task-resource coupling, lengthy solution generation times, and poor inter-platform collaboration, an unmanned swarm scheduling strategy tailored is proposed for mountain obstacle-breaching missions. Initially, by formalizing the descriptions of obstacle breaching operations, the swarm, and obstacle targets, an optimization model is constructed with the objectives of expected global benefit, timeliness, and task completion degree. A meta-task decomposition and reassembly strategy is then introduced to more precisely match the capabilities of unmanned platforms with task requirements. Additionally, a meta-task decomposition optimization model and a meta-task allocation operator are incorporated to achieve efficient allocation of swarm resources and collaborative scheduling. Simulation results demonstrate that the model can accurately generate reasonable and feasible obstacle breaching execution plans for unmanned swarms based on specific task requirements and environmental conditions. Moreover, compared to conventional strategies, the proposed strategy enhances task completion degree and expected returns while reducing the execution time of the plans.

Addressing critical challenges in unmanned sensing networks, ambiguous topological structures and inadequate networking methodologies, this study establishes a multilayer architecture based on information-value gain mechanisms. We formalize the networking planning problem as a distance-driven convergence task from sensing to combat nodes, proposing a traction-based hierarchical clustering algorithm as the solution. The algorithm incorporates a dual-optimization mechanism: 1) Mean Clustering Phase: Efficiently aggregates local scattered nodes into preliminary clusters; 2) Hierarchical Clustering Phase: Progressively abstracts network topology into a unified sensing network through iterative refinement. Military-specific innovations include: 1) Outlier Traction Strategy: Dynamically reassigns remote nodes to optimize connectivity; 2) Threat-Zone Avoidance: Embeds battlefield geographical constraints during topology formation The simulation experiments verified the rationality of the networking model proposed in this study and the feasibility of the improvement strategies, which can significantly enhance the networking efficiency and practical adaptability in complex battlefield environments.

This paper proposes an integrated dynamic weapon-target assignment (DWTA) framework that unifies physics-based game evaluation with exact combinatorial optimization. Traditional assignment methods relying on nominal kinematics fail against unpredictable adversarial maneuvers. To address this, a physics-informed neural network (PINN) is trained offline to solve the Hamilton-Jacobi-Isaacs equation, mapping the backward reachable tube (BRT). Acting as a high-fidelity approximation of the physical reachability boundary, the learned surrogate evaluates interception feasibility under worst-case evasions. The imposed BRT constraint induces a phase transition that sharply prunes physically unreachable combinations. Operating on this purified search space, a marginal-return column enumeration algorithm solves the resulting nonlinear integer programming problem. Monte Carlo simulations demonstrate that the PINN surrogate maintains high-precision boundaries under extreme dynamics. In large-scale scenarios where commercial solvers time out, the proposed DWTA algorithm consistently delivers high-quality solutions within milliseconds, providing a robust and physically verified decision baseline for closed-loop air defense systems.

In response to the challenges of multiple unmanned aerial vehicles'(UAV) rapid strike missions against multiple enemy targets in a complex urban combat environment, where the optimization of multiple objectives such as energy consumption, flight risk and inter-UAV safety distance must be considered, this paper proposes a two-layer optimization algorithm based on the A* algorithm and the improved particle swarm optimization(PSO) algorithm. The method utilizes the A* algorithm to generate initial paths, which serve as initial solutions for the PSO algorithm, thereby enhancing the early-stage search efficiency. Meanwhile, a particle mechanism with mutation structures is introduced to improve the global search capability of the PSO algorithm in its solution space and to mitigate premature convergence. The simulation results demonstrate that the proposed two-layer optimization algorithm signi-ficantly improves search efficiency and effectively addresses the problem of local optima. The findings verify the feasibility and effectiveness of the proposed algorithm in multi-UAV path planning tasks in a complex urban combat environment.

The high-dynamic nature and intense adversarial engagements inherent in modern warfare impose stringent demands on the evaluation and enhancement of dynamic performance in combat kill-webs. Addressing three critical challenges: (1) combat capability assessment, (2) adversarial resilience quantification, and (3) resilience enhancement strategy formulation, this study proposes an integrated methodological framework for optimizing adversarial resilience in kill-web systems. Building upon operational cycle theory, we establish a comprehensive kill-web model with mission success probability as the core operational capability metric. Through systematic analysis of sabotage process, we develop a resilience quantification framework integrating multi-stage capability value to characterize the kill-web’s functional persistence under progressive node incapacitation. A novel resilience enhancement model is designed through critical nodes protection strategy, and an improved discrete particle swarm optimization algorithm is employed to solve the model. The simulation experiments verify the rationality of the capability calculation and resilience analysis methods of the kill-web and the superiority of the protection strategy. These results provide both theoretical and practical advancements for designing survivable combat systems in modern warfare scenarios.

Multi-objective reinforcement learning (MORL) algorithms predominantly rely on scalarization functions parameterized with the preferences of the decision maker to derive trade-off solutions. However, this is not always feasible or desirable in the deterministic settings where scalarization functions are hard to specify, or where Pareto optimal solutions vary solely due to changes in the multi-objective reward function. Therefore, we consider a goal-augmented dynamic multi-objective Markov decision process (GA-DMOMDP), which enables the learning of Pareto optimal solutions through specifying and pursuing appropriate goals rather than relying on explicit scalarization functions. Restricted to the above GA-DMOMDPs, a multi-objective goal-oriented reinforcement learning (MOGORL) algorithm is further proposed so that the possibly changing Pareto optimal solutions can be tracked. In our algorithm, an on-line learning mode is proposed to continuously detect new goals, and to simultaneously pursue different goals by a hindsight relabeling strategy. Experimental results show that our algorithm can learn the Pareto optimal solutions in the deterministic environments with either static or dynamically changing rewards, regardless of the shape of Pareto optimal fronts, which outperforms generalized MORL algorithms with linear and Chebyshev scalarization functions.

Unmanned combat systems, as a new type of combat force composed of unmanned platforms with different combat capabilities, have significant advantages of high flexibility and low cost. However, the strong confrontational nature of modern warfare makes it vulnerable to attacks and paralysis. The existing recovery methods have problems such as poor effectiveness and slow speed, and it is urgent to enhance the adaptive recovery ability of this system in high-intensity confrontational environments. To address this issue, an optimization strategy is proposed to support the capability recovery of unmanned combat networks, aiming to enhance their rapid recovery ability and continuous combat effectiveness after being attacked. Firstly, an unmanned combat network model is constructed based on the OODA ring theory; Secondly, in combination with the concept of closed combat chains, the evaluation indicators of network combat capabilities are constructed from three dimensions: effectiveness, rapidity and redundancy. Then, a network capability recovery optimization model is established with the maximization of network combat capability as the objective function and the degree of network structure change and node load as the constraint conditions; Furthermore, an improved genetic algorithm is proposed for solution. The performance of the algorithm is enhanced by introducing a penalty function mechanism and adaptive crossover and mutation strategies. Simulation results show that the proposed strategy effectively reduces network structural variation and significantly enhances recovery capability, providing theoretical support for autonomous recovery of unmanned combat systems.

Addressing the challenges inherent in passive unmanned search and rescue missions at sea,including difficulties in target identification,broad search areas,and slow route planning,a strategic process was introduced for maritime search and rescue area planning and a route planning model specifically designed for passive unmanned missions.By thoroughly understanding the emergency response operations at sea and the specific needs for route planning,an optimal routing model have been developed considering factors such as the efficiency of search and rescue area coverage and the cost of rescue routes.The objective function is constructed within these constraints and solved using the whale optimization algorithm.The validity of the model is confirmed through designated scenario experiments,indicating that our proposed model for maritime passive unmanned search and rescue route planning is capable of swiftly identifying the search and rescue area and efficiently discovering a route with reduced costs.

Modern warfare has become confrontations between opposing combat systems. Determining damage to systemic targets (STs) is critical to effective decision-making and operational planning. Entities in an ST have their own functions, and different functions interact with one another to execute tasks. The ST can be abstracted as a combat functional network by abstracting the functions of entities as nodes and the relations between functions as edges. The collaborative behavior of different functions is defined in the efficacy loop model. Multiple network damage indicators are selected, modified, or newly designed from different aspects, and these indicators are aggregated without interference from subjective factors to obtain the composite damage index (CDI). Finally, the effectiveness of CDI is validated and compared with other methods in the case study. In various damage scenarios, this method is able to effectively obtain damage results and bring useful insights to combat operations.

In high-intensity modern urban combat environments, centralized task preallocation for heterogeneous multi-unmanned aerial vehicle (UAV) systems is a key technology for enhancing cooperative operational effectiveness. To address the challenges of diverse target types and the limited flexibility in allocation modes, this paper proposes a collaborative task allocation model for heterogeneous UAVs that simultaneously supports two allocation modes: single UAVs engaging multiple targets and multiple UAVs collaboratively striking a single target. The proposed model comprehensively considers key factors such as overall target damage effectiveness, task execution time, and ammunition consumption, with the objective of maximizing the total strike benefit. To efficiently solve this model, a discrete artificial bee colony algorithm incorporating a genetic structure is developed. During the initialization phase, a probability distribution-based greedy initialization is introduced to improve the quality of initial solutions and accelerate convergence. In the search phase, the algorithm integrates a discrete space-based multimodal neighborhood search strategy and a progressive expansion-based genetic learning mechanism, significantly enhancing its global exploration and local exploitation capabilities in complex solution spaces. Simulation results demonstrate that the proposed method can effectively accomplish cooperative multi-target task allocation across various task scales. It achieves higher solution quality and better computational efficiency, exhibiting superior overall performance compared to conventional algorithms.

A set of unmanned aerial vehicle (UAV) swarm combat effectiveness evaluation methods based on combat process simulation is constructed to address the issue of UAV swarm combat effectiveness evaluation in uncertain situations. Firstly, analyze the operational process of unmanned aerial vehicle (UAV) reconnaissance and strike missions from the perspective of closed-loop combat environment. Secondly, design a combat method calculation module based on the various steps of the closed-loop combat environment, and combine the designed parameter setting module, battlefield situation generation and recording module, global progress controller, and combat data recording module to form a complete simulation system for unmanned aerial vehicle group observation and combat process. Then, based on the available simulation data of the combat process, construct core effectiveness indicators and auxiliary effectiveness indicators, and provide a calculation and coupling method for each indicator. Finally, a scenario is set up to simulate the combat process of unmanned aerial vehicle swarms using different deployment schemes, evaluate and analyze the combat effectiveness, verify the feasibility of the proposed method, and achieve effective evaluation of the combat effectiveness of unmanned aerial vehicle swarms under uncertain situations.

This paper introduces AC-TSP, an Actor-Critic reinforcement learning model designed to solve the Traveling Salesman Problem (TSP). The model employs an encoderdecoder architecture, where the encoder leverages 1D convolutional layers (Conv1D) and positional encoding to extract node features, and the decoder integrates a gated recurrent unit (GRU) with an attention mechanism to generate feasible routes sequentially. Although trained solely on instances with 50 nodes, AC-TSP exhibits strong generalization capabilities across varying problem sizes. Compared to conventional baseline algorithms, AC-TSP consistently delivers high-quality solutions at multiple scales while significantly enhancing computational efficiency.

As a classic optimization problem, the weapon target assignment (WTA) problem is widely studied. However, existing WTA models are mostly established on single-hit destruction and static hit probabilities and take the destroyed target value expectation as the optimization objective, which limits their application to more complex scenarios. In response to the above shortcomings, this paper establishes a two-stage WTA model for a more complex scenario. In this model, a complex network-based assessment method is introduced and applied to better quantify the damage degree of targets, a method for calculating the probability of a target being destroyed when struck by multiple weapons is proposed, and the stage division significantly reduced the dimension of the optimization model. A Monte Carlo tree search enhanced by a metaheuristic algorithm and a memetic algorithm with a coevolution mechanism are used to solve two sub-models of the two-stage WTA model, respectively. The simulation results demonstrate that the two targeted heuristic algorithms can produce better solutions than similar heuristic algorithms, and the WTA model proposed in this paper is applicable to complex scenarios involving dynamic hit probabilities and heterogeneous targets.

A method for evaluating the combat effectiveness of the reconnaissance and strike UAV formation is proposed, which combines the OODA combat cycle theory and combat process simulation. Firstly, establish a combat effectiveness analysis framework based on the combat cycle theory by analyzing the relationship between combat effectiveness and combat cycle capability. Secondly, based on the composition of the combat cycle and combined with typical combat styles of the UAV formation, a simulation model of the combat process is constructed. Thirdly, based on the available simulation results data, a method for analyzing the combat effectiveness of the reconnaissance and strike UAV formation is proposed. Finally, a simulation experiment was conducted to analyze the combat effectiveness of the reconnaissance and strike UAV formation through a scenario, verifying the feasibility of the proposed method in this article, and providing support for the development and improvement of the reconnaissance and strike UAV formation equipment and upgrading of combat styles.

To achieve the autonomy of mobile robots, effective localization is an essential process. Among localization algorithms, the Adaptive Monte Carlo Localization (AMCL) algorithm is most commonly used in many indoor environments. However, when the initial position is unknown, the efficiency and success rate of localization based on the AMCL algorithm decrease with the increasing area of the map. In this paper, an improved MCL algorithm named off-line feature matching and improved particle swarm optimization for Monte Carlo Localization (OFM-IPSO MCL) is proposed. Feature matching is adopted to reduce the online computational burden. Compared with the AMCL algorithm, OFM-IPSO MCL shows better results in the problems of positioning without initial pose and kidnapping robot by using a small number of particles. For positioning without an initial pose, the OFM-IPSO algorithm uses the feature extraction and feature matching methods to find the possible positions of the robot. In the problem of kidnapping robot, a method for determining if the robot has been "kidnapped" is proposed, which determines whether the robot has lost its pose. The validity and efficiency of the OFM-IPSO MCL algorithm are demonstrated by the Robotic Operating System (ROS). Extensive results and comparisons are also provided in this paper.

In response to the significant impact of urban waterlogging on residents, the economy, and urban infrastructure in recent years, this study introduces an innovative wargame-based evaluation approach for emergency rescue plans. The primary goal of this research is to improve emergency rescue capabilities while minimizing costs and identifying gaps in existing emergency rescue plans. To effectively evaluate these capabilities, we extract specific content related to OODA (Observe, Orient, Decide, Act) dynamics in rescue actions. Furthermore, a comprehensive index system is developed to evaluate emergency rescue capabilities in the context of urban waterlogging scenarios. To address the challenges associated with intelligent optimization and evaluation of such systems, we employ a radial basis function neural network and conduct wargame experiments to obtain data and measure capability indices. The evaluation model is trained using data samples to ensure robust performance. In addition to the proposed model evaluation and analysis framework, we also present an evaluation and analysis method for RBF (Radical Basis Function) neural networks and compare the prediction results with those obtained from GRNN (Generalized Regression Neural Network), PNN (Product-based Neural Network), and BP (Back Propagation) neural network algorithms. This model efficiently processes and fits data by simulating expert experience for evaluation purposes. Such an approach takes advantage of machine learning's sensitivity to data characteristics, effectively avoiding the influence of human factors while stably reflecting the mapping relationship between indicators and performance outcomes. This research presents a novel solution with significant implications for the development of urban emergency rescue systems that address the challenges posed by urban waterlogging incidents.

With the widespread application of cruise missile technology and the increasing complexity of mission execution,the mission planning of cruise missile formations is facing new challenges.A multi patrol missile pre war planning task alloca-tion model based on multiple constraints is constructed,with the optimization objectives of mission value return and mission loss cost,while considering various practical constraints such as task execution timing.To solve the model,an improved ge-netic algorithm based on simulated annealing is designed,which adjusts the adaptive selection strategy and adaptive crossover strategy based on the population fitness of different periods,ensuring convergence while increasing population diversity.Fi-nally,the effectiveness of the designed improved algorithm in solving the task allocation problem of multiple patrol missiles was verified through a case study.

The harshness of amphibious landing operations has made unmanned equipment an important development direction for engineering support tasks in amphibious landing operations. This paper proposes a task allocation method based on fitness ranking genetic algorithm for unmanned engineering support tasks in amphibious landing operations. Firstly, an unmanned engineering support task allocation model is constructed considering the optimal objectives of engineering task completion time and equipment consumption, as well as constraints such as task sequence. Secondly, the genetic algorithm is combined with multi-dimensional dynamic list scheduling, and an improved genetic algorithm based on fitness ranking is proposed by introducing a mutation rate updating operation. The task assignment plan is obtained by using the improved genetic algorithm. Finally, simulation experiments based on hypothetical scenarios are conducted to verify the effectiveness and superiority of the algorithm, which can provide technical support for the execution of unmanned engineering support tasks in amphibious landing operations.

Quick and efficient mission planning is essential in maritime search and rescue (SAR). This includes defining the search area and developing an effective strategy. The task is fraught with challenges due to the difficulty of determining location information and the impact of complex meteorological environments. The primary objective of SAR mission planning is the rapid deployment of unmanned surface vehicles (USVs) to the incident area. While many planning algorithms prioritize the shortest route, there’s a lack of mission planning measures that maximize SAR effectiveness. In addition, the joint deployment of USVs increases the success rate compared to individual operations. Therefore, this paper presents a task assignment framework for USVs in SAR missions that considers the probability of success and time constraints. USVs are used to search for lost targets, and the framework consists of the following three modules: (1) a module for predicting the location of the overboard target to be rescued; (2) a module for modeling the probability of mission success; (3) a module for assigning search tasks to USVs. The framework first analyzes the search area. Then, it predicts the target location with a stochastic particle method, which incorporates marine environment forecast data to update the mission target location. To improve the scientific nature of USV search and rescue mission plans, an evaluation model is developed to assess mission capability. Simulation experiments and task scheme analysis validate its effectiveness.