Partitioning a pool of capability-heterogeneous human and machine agents into effective teams for multiple concurrent tasks is a fundamental challenge in hybrid human–machine collaboration. We formalize this problem as aMulti-Task Additively Separable Hedonic Game(MT-ASHG), in which every agent is a self-interested player whose utility combines (i) a task-specific proficiency score measuring how well the agent’s skill vector aligns with task requirements, and (ii) an inter-agent compatibility score capturing synergistic or conflicting partnerships. Building on this formulation, we design an Iterative Best Response (IBR) algorithm that lets agents autonomously migrate between task groups to improve their individual payoffs. We prove that the IBR dynamics converge in a finite number of steps to an individually stable partition, in which no agent can unilaterally improve its utility by switching teams, and analyze the computational complexity of the convergence process. To bridge theory and practice, we further introduce an empirical profiling method that extracts capability and compatibility vectors from historical agent interactions, addressing the cold-start problem in newly formed human–machine teams. Extensive experiments on synthetic benchmarks and the Overcooked-AI cooperative environment demonstrate that MT-ASHG consistently outperforms centralized assignment baselines and existing coalition-formation methods in terms of global task completion rate, fairness, and scalability.

In recent years, industrial chain cooperation has evolved into multiplex network structures where product agents are linked through diverse types of interdependencies. While such architectures enhance coordination flexibility, resource-sharing efficiency, and system-level resilience, they also introduce complex hybrid dynamics. These dynamics emerge from fluctuating task demands, continuous changes in network topology as agents join or leave, and variations in production capacities across agents. Their interactions generate cascading cross-layer effects that disrupt load balance and challenge conventional scheduling and resource management strategies. This work addresses the resulting complexity by proposing the hierarchical grouped task migration (HGTM) algorithm, which migrates tasks in groups rather than individually. Leveraging its hierarchical design, HGTM enables effective multilevel load balancing throughout the multiplex structure while keeping computational overhead low. Comprehensive theoretical analysis and experiments show that HGTM enhances task completion rates, improves execution utility, and reduces completion costs. The approach exhibits strong robustness and adaptability, particularly under increasingly dynamic and highly coupled operating conditions.

Industrial software systems underpin complex industrial platforms by integrating cross-disciplinary expertize and sophisticated workflows. Their inherent complexity generates substantial cognitive demands, necessitating contextual, adaptive, and personalized knowledge support aligned with dynamic tasks and evolving expertize. Conventional recommendation approaches struggle to address heterogeneous knowledge sources, dynamic user needs with intent drift, and the domain-specific semantics required in industrial software. To overcome these challenges, we propose a context-aware dialog-based knowledge recommendation (CADKR) model, powered by large language models (LLMs). CADKR fuses dynamic interaction context with domain knowledge through a novel recommendation network (RecNet), enabling robust generalization to unseen scenarios and adaptability to preference drift. For practical deployment, lightweight optimization strategies compress model size by up to 98% without compromising accuracy. A case study in a large chemical industrial park demonstrates the effectiveness of CADKR in enhancing industrial knowledge support, while application deployments in leading new energy vehicle enterprises and the world’s largest circular resource power plant have yielded an independent verification report, providing strong evidence of the proposed method’s practical value.11The report is available athttps://anonymous.4open.science/r/verification-report. Code and data are available athttps://anonymous.4open.science/r/CADKR-TCSS. To meet double-blind requirements, organization and personnel details are anonymized and will be disclosed after acceptance.

In industrial production processes, disruptions within the industrial chain can severely affect the collaborative capabilities of production agents. A notable example occurred during the COVID-19 pandemic, when many agents faced interruption risks and were unable to participate in coordinated production. Ensuring continuity under such conditions requires migrating tasks from disrupted agents to viable alternatives. Designing effective task migration strategies, however, must account for the emergent multiplex nature of modern industrial chains. In these multiplex networked industrial chains, disruption risk in one layer can propagate to others, generating cascading failures across the system. This introduces two key challenges: (1) disruption risk creates mismatches not only between product agents and tasks but also across network layers, enlarging the problem dimensionality; and (2) simultaneous disruptions across multiple agents and layers increase the volume of tasks needing migration, greatly expanding the solution space. To address these challenges, we introduce the notion of a multiplex potential field, which captures cross-layer interdependencies and system-level dynamics in multiplex industrial chains. Building on this concept, we develop a hierarchical contextual task migration algorithm that exploits the multiplex potential field to guide both inter-layer and intra-layer task reallocations. Extensive experiments show that our approach consistently achieves superior utility, markedly improves task completion ratios, and reduces execution costs compared to benchmark algorithms. Furthermore, it attains solution quality comparable to that of the optimal CPLEX solver while requiring substantially less computation time. Finally, a case study on the FAO international food trade network demonstrates that the proposed framework is not only theoretically robust but also practically effective when deployed on large-scale real-world multiplex systems.

In the heterogeneous and dynamically evolving Industrial Internet, industrial agents are required to possess cross-domain adaptability and self-learning capabilities to facilitate task generalization and scalable deployment across diverse operational contexts. However, existing domain adaptation approaches predominantly rely on static feature alignment or domain-invariant assumptions, lacking a systematic consideration of working condition variability and the interplay between self-learning and adaptation. This oversight hampers their effectiveness in real-world industrial scenarios, where agents must operate under complex conditions with limited target domain knowledge. Consequently, these methods often suffer from knowledge shift and insufficient policy generalization. To address these limitations, this article introduces the instance weighting-based domain-adaptive optimization (IW-DAO) framework. IW-DAO combines an instance weighting-based knowledge alignment mechanism with a Bayesian optimization strategy, forming a dynamic self-learning loop tailored for cross-domain adaptation. Specifically, the framework constructs an adaptive knowledge representation in a high-dimensional invariant feature space and formulates a cross-domain performance evaluation estimator to guide the unsupervised learning of knowledge transfer and adaptive optimization via Bayesian iterative search. Extensive experiments on industrial asset management tasks as well as a real-world industrial flow process dataset with various operating conditions demonstrate the effectiveness of IW-DAO. The proposed framework enables industrial agents to evolve autonomously and be deployed efficiently across diverse domains. IW-DAO consistently outperforms baseline and expert-tuned methods, demonstrating strong generalization and adaptability in both industrial asset management and complex flow process scenarios.

From the perspective of service computation, the current service-oriented architecture of industrial chain networks faces issues such as the lack of multiple computing services and insufficient detection capabilities, especially in the context of insider trading detection services for multiplex networked industrial chains. Existing service computation methods are often limited to single-market or simplified network structures, making it difficult to fully capture the dynamic changes and cross-chain propagation characteristics of insider trading within complex multi-layered industrial chain networks. Therefore, developing an efficient and accurate insider trading detection service computation method in the environment of multiplex networked industrial chains remains a critical research challenge. To address this, this paper proposes an Insider Trading Detection Service Computation Method for Multiplex Networked Industrial Chains based on Hybrid Temporal Granularity Scaling (ITDSC-MNICHTGS). This method combines insider trading feature modeling in multiplex networks, an adaptive time granularity adjustment mechanism, and unsupervised learning techniques to accurately identify insider trading behaviors in multiplex networked industrial chains without relying on data labeling. Experimental results show that, compared to traditional detection methods, the proposed method outperforms in metrics such as precision, recall, and$\mathbf{F 1}$-score, effectively improving the reliability and applicability of insider trading detection.

The emergence of diversified computing paradigms and the proliferation of interconnected devices have transformed modern service computing systems into multiplex service networks. In these networks, services are provisioned across multiple interconnected layers, such as infrastructure layer, platform layer, and application layer. A distinctive characteristic of these systems is the presence of hybrid dynamics, characterized by the complex interplay of service request dynamics (random service request arrivals), service network dynamics (topology changes due to agent joins and leaves), and service resource dynamics (fluctuating service provisioning capabilities). These interrelated hybrid dynamics frequently propagate bidirectionally across system layers, creating intricate emergent behaviors and potentially triggering cascading load imbalances that can significantly degrade system performance and reliability. To comprehensively address these multidimensional challenges, we propose a novel and adaptive group-based task migration methodology. By systematically migrating tasks in cohesive groups rather than as isolated individual entities, this innovative approach effectively buffers the destabilizing impact of hybrid dynamics while simultaneously reducing computational and communication overhead associated with frequent decision-making processes. Through rigorous theoretical analysis and extensive experiments, our proposed method shows remarkable advantages in service computing scenarios with hybrid dynamics. The results validate the effectiveness of our group-based migration paradigm in addressing the challenging requirements of modern multiplex service computing systems.

The increasing complexity of industrial information service systems presents significant challenges for anomaly detection, particularly in ensuring service reliability across multilayered service networks. This paper proposes TS-GNN, a novel Temporal-Spatial Graph Neural Network framework that effectively integrates temporal pattern recognition with spatial dependency modeling for anomaly detection in industrial service environments. The framework employs a multi-scale temporal feature extraction mechanism that combines frequency-domain transformation with hierarchical decomposition to capture temporal patterns at different granularities. Subsequently, a graph neural network with attention-based message passing models spatial correlations and anomaly propagation patterns among service nodes. Comprehensive experiments on three benchmark datasets from service computing domains demonstrate that TSGNN achieves superior performance, with F1-scores of 94.74% on SWaT, 85.14% on SMD, and 96.36% on PSM datasets. Compared to the best baseline methods, TS-GNN shows consistent improvements with an average$\mathbf{F 1}$-score enhancement of$\mathbf{0. 7 9 \%}$points, providing an effective solution for enhancing the reliability and robustness of service computing systems.

To better identify user service needs and preferences, proactive conversational interactions are essential, a concept encapsulated by Conversational Recommender Systems (CRS) in service and recommendation domains. A key challenge faced by CRS lies in accurately capturing user preferences from dialogue contexts, particularly in non-stationary environments where traditional methods are hindered by cold-start problems and shifting service demands. Motivated by the strong generalization abilities of In-Context Learning (ICL) in dynamic and unfamiliar scenarios, this paper proposes ICD4CR, a causal decision-making framework grounded in in-context decision-making principles. Leveraging large Language Models (LMs), ICD4CR adopts a data-driven pretraining paradigm, enabling it to infer optimal recommendation strategies from historical dialogue trajectories in analogous service contexts, circumventing the need for explicit user modeling. We introduce a recommendation network that integrates seamlessly with the foundational LM, allowing ICD4CR to function as a fully end-to-end recommendation system. To enhance efficiency and adaptability, adapter-based techniques are employed for knowledge transfer and fine-tuning.

With the rapid development and deep integration of artificial intelligence and automation technologies, autonomous unmanned swarms dynamically organize into multiplex network structures based on diverse task requirements in adversarial environments. Frequent task variations lead to load imbalances among agents and between network layers, significantly increasing the risk of enemy detection and destruction. Existing approaches typically simplify multiplex networks into single-layer structures for task scheduling, failing to address these load imbalance issues. Moreover, the coupling between task dynamics and network multiplexity dramatically increases the complexity of designing task migration strategies, and it is proven NP-hard to achieve such load balancing. To address these challenges, this paper proposes a risk-aware task migration method that achieves dynamic load balancing by matching task requirements with both intra-layer agent capabilities and inter-layer swarm capabilities. Simulation results demonstrate that our approach significantly outperforms benchmark algorithms in task completion cost, task completion proportion, and system robustness. In particular, the algorithm achieves solutions statistically indistinguishable from the optimal solutions computed by the CPLEX solver, while exhibiting significantly reduced computational overhead.

This article addresses distributed optimization problems, in which a group of agents cooperatively minimize the sum of their private objective functions via information exchanging. Building on alternating direction method of multipliers (ADMM), we propose a privacy-preserving and communication-efficient decentralized quadratically approximated ADMM algorithm, termed PC-DQM, for solving such type of problems under the scenario of limited communication. In PC-DQM, an event-triggered mechanism is designed to schedule the communication instants for reducing communication cost. Simultaneously, for privacy preservation, a Hessian matrix with perturbed noise is introduced to quadratically approximate the objective function, which results in a closed form of primal vector update and then avoids solving a subproblem at each iteration with possible high computation cost. In addition, the triggered scheme is also utilized to schedule the update of Hessian, which can also reduce computation cost. We theoretically show that PC-DQM can protect privacy but without losing accuracy. In addition, we rigorously prove that PC-DQM converges linearly to the exact optimal solution for strongly convex and smooth objective functions. Finally, numerical simulation is presented to illustrate the effectiveness and efficiency of our algorithm.

Knowledge-based clustering algorithms can improve traditional clustering models by introducing domain knowledge to identify the underlying data structure. While there have been several approaches to clustering with the guidance of knowledge tidbits, most of them mainly focus on numeric knowledge without considering the uncertain nature of information. To capture the uncertainty of information, pure numeric knowledge tidbits are expanded to knowledge granules in this article. Then, two questions arise: how to obtain granular knowledge and how to use those knowledge granules in clustering. To the end, a novel knowledge extraction and granulation (KEG) method and a granular knowledge-based fuzzy clustering model are proposed in this study. First, inspired by the concept of natural neighbors, an automatic KEG is developed. In KEG, high-density points are filtered from the dataset and then merged with their natural neighbors to form several dense areas, i.e., granular knowledge. Furthermore, the granular knowledge expressed by interval or triangular numbers is leveraged into the clustering algorithm, which is the framework of fuzzy clustering with granular knowledge. To concretize this model into clustering algorithms, the classical fuzzy C-Means clustering algorithm has been selected to incorporate the granular knowledge produced by KEG. Then, the corresponding fuzzy C-Means clustering with interval knowledge granules (IKG-FCM) and triangular knowledge granules (TKG-FCM) are proposed. Experiments on synthetic and real-world datasets demonstrate that IKG-FCM and TKG-FCM always achieve better clustering performance with less time cost, especially on imbalanced data, compared with state-of-the-art algorithms.

Multi-robot task allocation (MRTA) problem has long been a key issue in multi-robot systems. Previous studies usually assumed that the robots must complete all tasks with minimum time cost. However, in many real situations, some tasks can be selectively performed by robots and will not limit the achievement of the goal. Instead, completing these tasks will cause some functional effects, such as decreasing the time cost of completing other tasks. This kind of task can be called “functional task”. This paper studies the multi-robot task allocation in the environment with functional tasks. In the problem, neither allocating all functional tasks nor allocating no functional task is always optimal. Previous algorithms usually allocate all tasks and cannot suitably select the functional tasks. Because of the interaction and sequential influence, the total effects of the functional tasks are too complex to exactly calculate. We fully analyze this problem and then design a heuristic algorithm. The heuristic algorithm scores the functional tasks referring to linear threshold model (used to analyze the sequential influence of a functional task). The simulated experiments demonstrate that the heuristic algorithm can outperform the benchmark algorithms.

As an essential problem in robotics, foraging means that robots collect objects from a given environment and return them to a specified location. On many occasions, robots are required to perform foraging tasks in adversarial environments, such as battlefield rescue, where potential adversaries may damage robots with a certain probability. The longer an individual robot moves through adversarial environments, the higher the probability of being damaged by adversaries. The robot system can gain utility only when the robot brings carried objects back to a predetermined home station. Such a risk of being damaged makes returning home at different locations potentially relevant to the expected utility produced by the robot. Thus, the individual robot faces a dilemma when it responds to the potential risks in adversarial environments: whether to return the carried resources home or continue foraging tasks. In this article, two fundamental environment settings are discussed, homogeneous cases and heterogeneous cases. The former is analyzed as having both the optimal substructure property and the non-aftereffect property. Then, we present a dynamic programming (DP) algorithm that can find an optimal solution with polynomial time complexity. For the latter, it is proven that finding an optimal solution is \( \mathcal {NP} \) -hard. We then propose a heuristic algorithm: A division hierarchical path planning (DHPP) algorithm that is based on the idea of dividing the foraging routes generated initially into a certain number of subroutes to dilute risks. Finally, these algorithms are extensively evaluated in simulations, concluding that in adversarial environments, they can significantly improve the productivity of an individual robot before it is damaged.

Team formation has been extensively studied for complex task crowdsourcing in E-markets, in which a set of workers are hired to form a team to complete a complex task collaboratively. However, existing studies have two typical drawbacks: 1) each team is created for only one task, which may be costly and cannot accommodate crowdsourcing markets with a large number of tasks; and 2) most existing studies form teams in a centralized manner by the requesters, which may place a heavy burden on requesters. In fact, we observe that many complex tasks at real-world crowdsourcing platforms have similar skill requirements and workers are often connected through social networks. Therefore, this paper explores distributed team formation-based batch crowdsourcing for complex tasks to address the drawbacks in existing studies, in which similar tasks can be addressed in a batch to reduce computational costs and workers can self-organize through their social networks to form teams. To solve such an NP-hard problem, this paper presents two approaches: one is to form a fixed team for all tasks in the batch; the other is to form a basic team that can be dynamically adjusted for each task in the batch. In comparison, the former approach has lower computational complexity but the latter approach performs better in reducing the total payments by requesters. With the experiments on a real-world dataset comparing with previous benchmark approaches, it is shown that the presented approaches have better performance in saving the costs of forming teams, payments by requesters, and communication among team members; moreover, the presented approaches have higher success rate of tasks and much better scalability.

Existing studies on the multirobot foraging problem often assume safe settings, in which nothing in an environment hinders the robots’ tasks. In many real-world applications, robots have to collect objects from hazardous environments like earthquake rescue, where possible risks exist, with possibilities of destroying robots. At this stage, there are no targeted algorithms for foraging robots in hazardous environments, which can lead to damage to the robot itself and reduce the final foraging efficiency. A motivating example is a rescue scenario, in which the lack of a suitable solution results in many victims not being rescued after all available robots have been destroyed. Foraging robots face a dilemma after some robots have been destroyed: whether to take over tasks of the destroyed robots or continue executing their remaining foraging tasks. The challenges that arise when attempting such a balance are twofold: (1) the loss of robots adds new constraints to traditional problems, complicating the structure of the solution space, and (2) the task allocation strategy in a multirobot team affects the final expected utility, thereby increasing the dimension of the solution space. In this study, we address these challenges in two fundamental environmental settings: homogeneous and heterogeneous cases. For the former case, a decomposition and grafting mechanism is adopted to split this problem into two weakly coupled problems: the foraging task execution problem and the foraging task allocation problem. We propose an exact foraging task allocation algorithm, and graft it to another exact foraging task execution algorithm to find an optimal solution within the polynomial time. For the latter case, it is proven \( \mathcal {NP} \) -hard to find an optimal solution in polynomial time. The decomposition and grafting mechanism is also adopted here, and our proposed greedy risk-aware foraging algorithm is grafted to our proposed hierarchical agglomerative clustering algorithm to find high-utility solutions with low computational overhead. Finally, these algorithms are extensively evaluated through simulations, demonstrating that compared with various benchmarks, they can significantly increase the utility of objects returned by robots before all the robots have been stopped.

Multiagent pursuit-evasion problems have been widely investigated in a number of related areas. In many situations, the pursuers have energy limitation when pursuing the evader, and the pursuers moving at higher speeds consume energy more quickly. The energy limitation creates two problems for the pursuers. First, constantly moving at the maximum speed is not optimal because the energy will be exhausted quickly. Therefore, it is necessary to control the speed to achieve the optimal energy consumption, but previous studies have not presented a suitable speed-control algorithm. Second, a fraction of pursuers with insufficient energy may fail to complete the sub-tasks, and the evader can vary the escape strategy by considering the energy limitation of the pursuers. Therefore, the pursuers with sufficient energy need to help the pursuers with insufficient energy and address the variety of the evader's escape strategy. The existing pursuing strategies are not self-adaptable to satisfy this requirement. In this paper, we present a pursuing strategy consisting of both a speed-control algorithm and a path planning algorithm as sub-algorithms. Additionally, we integrate the subalgorithms into a self-adaptable cooperation strategy based on simulated annealing, which makes the pursuers adaptably help each other and is adaptable to the variety of escape strategies. The experimental results show that our strategy leads to higher capture success ratios than previous strategies when the energy of pursuers is limited. The experimental data also show that our strategy is more adaptable to the variety of the pursuers' maximum speed and the variety of the evader's escape strategies.

The iterative hard-thresholding algorithm (ISTA) is one of the most popular optimization solvers to achieve sparse codes. However, ISTA suffers from following problems: 1) ISTA employs non-adaptive updating strategy to learn the parameters on each dimension with a fixed learning rate. Such a strategy may lead to inferior performance due to the scarcity of diversity; 2) ISTA does not incorporate the historical information into the updating rules, and the historical information has been proven helpful to speed up the convergence. To address these challenging issues, we propose a novel formulation of ISTA (named as adaptive ISTA) by introducing a novel \textit{adaptive momentum vector}. To efficiently solve the proposed adaptive ISTA, we recast it as a recurrent neural network unit and show its connection with the well-known long short term memory (LSTM) model. With a new proposed unit, we present a neural network (termed SC2Net) to achieve sparse codes in an end-to-end manner. To the best of our knowledge, this is one of the first works to bridge the $\ell_1$-solver and LSTM, and may provide novel insights in understanding model-based optimization and LSTM. Extensive experiments show the effectiveness of our method on both unsupervised and supervised tasks.