Multiagent Systems

The present hypothesis develops the concepts of ChronoShell, postulating that the observed orderliness of the Universe stems from the finite-dimensional nature of the initial superposition of a white hole, which follows from the holographic principle. The self-organization of matter and physical laws is explained through principles of computational efficiency and quasi-fractal dynamics within a finite basis of states. The mathematical apparatus is built on strict variational principles and connections with fundamental physical laws.

This paper introduces an interactive Python debugging framework that combines multi-agent reinforcement learning, Natural Language Processing (NLP), and long-term memory. Two Proximal Policy Optimization (PPO) agents specialize in syntax and logic errors, generating candidate fixes that developers can accept, reject, or refine. A BERT-based module encodes natural language feedback into dense embeddings and quality scores, which shape reward signals for Reinforcement Learning from Human Feedback (RLHF). To support personalization, the system uses dual FAISS indices to retrieve past fixes based on codeerror pairs and developer explanations. Evaluated on a synthetic dataset of 200 Python programs, our approach achieves an 88% syntaxfix rate and 45% logic-fix rate within five suggestions-outperforming one-shot Large Language Model (LLM) baselines. In addition, the system improves the quality of the explanation, as measured by BLEU, ROUGE, and Code-BLEU. By integrating multi-agent specialization, linguistic feedback, and memory-driven retrieval, our framework delivers a more efficient, adaptive, and developer-aligned debugging experience.

Decentralized planning in uncertain environments is a complex task generally dealt with by using a decision-theoretic approach, mainly through the framework of Decentralized Partially Observable Markov Decision Processes (DEC-POMDPs). Although DEC-POMDPS are a general and powerful modeling tool, solving them is a task with an overwhelming complexity that can be doubly exponential. In this paper, we study an alternate formulation of DEC-POMDPs relying on a sequence-form representation of policies. From this formulation, we show how to derive Mixed Integer Linear Programming (MILP) problems that, once solved, give exact optimal solutions to the DEC-POMDPs. We show that these MILPs can be derived either by using some combinatorial characteristics of the optimal solutions of the DEC-POMDPs or by using concepts borrowed from game theory. Through an experimental validation on classical test problems from the DEC-POMDP literature, we compare our approach to existing algorithms. Results show that mathematical programming outperforms dynamic programming but is less efficient than forward search, except for some particular problems. The main contributions of this work are the use of mathematical programming for DEC-POMDPs and a better understanding of DEC-POMDPs and of their solutions. Besides, we argue that our alternate representation of DEC-POMDPs could be helpful for designing novel algorithms looking for approximate solutions to DEC-POMDPs.

Decentralized planning in uncertain environments is a complex task generally dealt with by using a decision-theoretic approach, mainly through the framework of Decentralized Partially Observable Markov Decision Processes (DEC-POMDPs). Although DEC-POMDPS are a general and powerful modeling tool, solving them is a task with an overwhelming complexity that can be doubly exponential. In this paper, we study an alternate formulation of DEC-POMDPs relying on a sequence-form representation of policies. From this formulation, we show how to derive Mixed Integer Linear Programming (MILP) problems that, once solved, give exact optimal solutions to the DEC-POMDPs. We show that these MILPs can be derived either by using some combinatorial characteristics of the optimal solutions of the DEC-POMDPs or by using concepts borrowed from game theory. Through an experimental validation on classical test problems from the DEC-POMDP literature, we compare our approach to existing algorithms. Results sho...

When information or control in a multiagent system is private to the agents, they may misreport this information or refuse to execute an agreed outcome, in order to change the resulting end state of such a system to their benefit. This may result in execution failures. When only information is private, mechanisms such as VCG use payments to create incentives for truthful behavior, and can then guarantee a non-negative utility for all agents. However, when control is also private, such existing mechanisms lose truthfulness and individual rationality: payments should depend on the actual outcome (not on the planned outcome) and some agents should be compensated. We give a more general version of the known negative result in the context of actions with dependencies, and we give a mechanism that can guarantee a nonnegative utility to the agents and is truthful in an ex-post Nash equilibrium.

This article presents an experimental research whose objective is to verify through experimentation if the systems for air conditioning and lighting homes based on the use of Multi-Agent Systems (MAS) are more efficient than traditional systems in terms of perceived environmental comfort. by people and the electricity cost generated. The proposed Consumo eléctrico doméstico, Sistema Multi-Agente (MAS),

Despite the significant progress to extend Markov Decision Processes (MDP) to cooperative multi-agent systems, developing approaches that can deal with realistic problems remains a serious challenge. Existing approaches that solve Decentralized Markov Decision Processes (DEC-MDPs) suffer from the fact that they can only solve relatively small problems without complex constraints on task execution. OC-DEC-MDP has been introduced to deal with large DEC-MDPs under resource and temporal constraints. However, the proposed algorithm to solve this class of DEC-MDPs has some limits: it suffers from overestimation of opportunity cost and restricts policy improvement to one sweep (or iteration). In this paper, we propose to overcome these limits by first introducing the notion of Expected Opportunity Cost to better assess the influence of a local decision of an agent on the others. We then describe an iterative version of the algorithm to incrementally improve the policies of agents leading to higher quality solutions in some settings. Experimental results are shown to support our claims.

Intelligent control has been extensively applied to aerial robots, which is important for autonomous flight applications. However, it is susceptible to noise and uncertainties when integrating different sensors and nonlinear aerial robot dynamics. This paper proposes a robust, neural network-based adaptive hybrid control framework for a unified position and orientation tracking in Hexarotors. The proposed framework is considered for receiving uncertain data from altitude and orientation sensors. It further tackles estimating these timevarying noises, uncertainties, and unmodeled dynamics by synergistically integrating a second-order sliding mode control (SMC) with radial basis function neural networks (RBFNNs). The RBFNN-augmented SMC dynamically adapts control commands, with stability proven via the Lyapunov method. These advancements ensure a simultaneous and accurate position and orientation tracking under varying conditions. The framework is validated through extensive software and hardware in the loop (SIL/HIL) simulations, showcasing significant improvements in tracking accuracy, disturbance rejection, and adaptability compared to traditional methods.

Since several years, multi-agent systems have been used for wide range of application areas and particularly in systems modelling and simulation. Those systems are based on the interaction of several individual components, called agents, which are proactive considering interaction with neighbours and with the surrounding environment. The IRTES-SeT laboratory develops a PEM-Fuel Cell model for detailed layer level simulation based on multi-agent paradigm. Whatever the level of abstraction considered, modelling the electrochemical process requires at least: an electrochemical model, a fluidic model, and a thermal model. These elements are integrated into our simulation model at the microscopic level. The goal of this paper is to present a multi-physical micro-level PEM-Fuel Cell model by using a multi-agent approach.

In this article a multi-agent system is presented which generates melody pitch sequences with a hierarchical structure. The agents have no explicit melodic intelligence and generate the pitches as a result of artificial emotional influence and communication between agents, and the melody's hierarchical structure is a result of the emerging agent social structure. The system is not a mapping from multi-agent interaction onto musical features, but actually utilizes music for the agents to communicate artificial emotions. Each agent in the society learns its own growing tune during the interaction process. Experiments are presented demonstrating that diverse and non-trivial melodies can be generated, as well as a hierarchical musical structure. The generation of novel music is at the heart of many computer-aided composition (CAC) systems. Without some way of generating new material, a CAC will churn out the same material time after time. To avoid this, many systems utilize random numbers. A more recent alternative is the generation of complex structures which are ordered but unpredictable. Popular types of systems that generate structures with such complexity are found in the field of artificial life or A-Life . A-Life investigates systems related to life, their processes, and evolution; it does this most often through computer simulations and models -for example cellular automata. Many A-life systems have two elements in common with have made them attractive to composers for use in CAC: they generate complex data with order and

An information agent is viewed as a deductive database consisting of 3 parts: • an observation database containing the facts the agent has observed or sensed from its surrounding environment. • an input database containing the information the agent has obtained from other agents • an intensional database which is a set of rules for computing derived information from the information stored in the observation and input databases. Stabilization of a system of information agents represents a capability of the agents to eventually get correct information about their surrounding despite unpredictable environment changes and the incapability of many agents to sense such changes causing them to have temporary incorrect information. We argue that the stabilization of a system of cooperative information agents could be understood as the convergence of the behavior of the whole system toward the behavior of a "superagent", who has the sensing and computing capabilities of all agents combined. We show that unfortunately, stabilization is not guaranteed in general, even if the agents are fully cooperative and do not hide any information from each other. We give sufficient conditions for stabilization. We discuss the consequences of our results.

C'est avec tr ès grand plaisir que je souhaite remercier tous ceux qui m'ont aid é ou soutenu durant les dix derni ères ann ées de ma carri ère, que r ésume ce m émoire. Je remercie REN É MANDIAU de m'avoir fait l'honneur de pr ésider mon jury ; ainsi que MARIE-PIERRE GLEIZES, CYRILLE BERTELLE et OLIVIER BOISSIER d'avoir accept é la lourde t âche de rapporteur ; et FRANC ¸OIS CHARPILLET d'avoir participer à mon jury et évaluer mon travail. Nos échanges ont confirm é ou ouvert des perspectives passionnantes. ABDERRAFIAA KOUKAM m'a recrut é dans son équipe. Je le remercie pour la libert é d'action et la confiance qu'il m'a accord ées. Je fais partie des nombreux chercheurs qui lui doivent leur carri ère et qui partagent des valeurs qu'il aura transmises par ses conseils et son exemple. Je lui suis profond ément reconnaissant pour tout ce qu'il a fait et continue à faire pour moi. VINCENT HILAIRE a ét é le premier à avoir propos é une approche organisationnelle dans notre laboratoire. Ses qualit és scientifiques et humaines ont ét é autant d'inspirations pour g érer mes activit és de recherche. C'est avec un grand plaisir que je le remercie d'avoir accept é de participer à mon jury. NICOLAS GAUD a ét é mon premier doctorant. Je garderai un souvenir imp érissable de notre exp érience, étalonnant mes exigences pour le « malheur » de ses successeurs. Je tiens également à le remercier profond ément pour son amiti é, son soutien et sa pr ésence durant ces ann ées de labeur. Il est un pilier de notre équipe de recherche, initiateur et architecte de nombreuses id ées qui ont contribu é aux travaux de recherche pr ésent és dans ce m émoire. SEBASTIAN RODRIGUEZ est une inspiration pour ses qualit és scientifiques et humaines. Architecte des premiers mod èles holoniques dans notre laboratoire, sans lui, les travaux pr ésent és dans ce m émoire n'auraient pas vu le jour. Je le remercie pour m'avoir permis de d écouvrir sa culture et son pays : l'Argentine. Mon amiti é lui est acquise, ainsi qu' à sa famille. DAVID, DAVY, JOLECYN, MICHA ËL, OLIVIER, PHILIPPE et RENAN sans qui la vie au laboratoire et à l'ext érieur n'aurait pas eu ce « je ne sais quoi » éminemment agr éable et divertissant. N'est-ce pas PHILIPPE ! Merci à ARIANE, CINDY, FABRICE, FRANCK, FR ÉD ÉRICK, GILLIAN, JALIL, PABLO et YASSINE ; nos petites causeries au coin caf é et ailleurs resteront des instants privil égi és. Merci à ANSAR, CHRISTOPHE, LUK, MASSIMO et THOMAS pour nos collaborations et échanges scientifiques vivifiants. Une pens ée pour mon fils ARTHUR, sans qui rien n'aurait ét é possible, pour Ève, source de maintes émotions, ainsi que pour AURORE, CHRISTELLE, DELPHINE, JEAN-CLAUDE, JEAN-FRANC ¸OIS, J ÉR ÔME, JOSETTE, MARIE-CLAUDE et tous mes autres proches, pour leur patience, leur amour et leurs encouragements durant cette derni ère d écennie. Merci infiniment. vii SOMMAIRE xi 6.2 Mod élisation et évaluation d'un r éseau de bus .

This paper explores the development and evaluation of physics-specific large-scale AI models, which we refer to as large physics models (LPMs). These models, based on foundation models such as large language models (LLMs) are tailored to address the unique demands of physics research. LPMs can function independently or as part of an integrated framework. This framework can incorporate specialized tools, including symbolic reasoning modules for mathematical manipulations, frameworks to analyse specific experimental and simulated data, and mechanisms for synthesizing insights from physical theories and scientific literature. We begin by examining whether the physics community should actively develop and refine dedicated models,

The decision activity in the field of environmental management is highly complex and involves a great number of contradictory interests (socio-economic, ecological, etc.). Environmental problems often result from the distributed and uncoordinated land use management practices of individual decision-makers that, when taken together, cause significant environmental impacts. To develop feasible and politically acceptable solutions to such problems it is often necessary to foster compromise and consensus among

The problem is complicated due to undetermined constraints of each component service and unpredictable solutions contributed by service providers. It needs the ability to rapidly identify the suitable solutions as well as effectively coordinate them under various constraints. In this paper, an agent-mediated coordination framework for e-service/process composition is proposed. Each agent works as a broker for each service type, posting service constraints, searching suitable solutions and refining the constraints for achieving coherence among the decisions of each service. Based on the framework, a prototype of multi-agent supported e-supply chain composition is implemented. The experimental results indicate the significant effectiveness of the approach.

With e-business emerging as a key enabler to drive supply chains, the focus of supply chain management has been shifted from production efficiency to customer-driven and partnership synchronization approaches. This strategic shift depends on the match between the demands and offerings that deliver the services. To achieve this, we need to coordinate the flow of information among the services, and link their business processes under various constraints. Existing approaches to this problem have relied on complete information of services and resources, and have failed to adequately address the dynamics and uncertainties of the operating environments. The real-world situation is complicated as a result of undetermined requirements of services involved in the chain, unpredictable solutions contributed by service providers, and dynamic selection and aggregation of solutions to services. This paper examines an agent-mediated approach to on-demand e-business supply chain integration. Each agent works as a service broker, exploring individual service decisions as well as interacting with each other for achieving compatibility and coherence among the decisions of all services. Based on the framework, a prototype has been implemented with simulated experiments highlighting the effectiveness of the approach.

Ant algorithms and flocking algorithms are the two main programming paradigms in swarm intelligence. They are built on stochastic models, widely used in optimization problems. However, though this modeling leads to highperformance algorithms, some mechanisms, like the symmetry break in ant decision, are still not well understood at the local ant level. Moreover, there is currently no modeling approach which joins the two paradigms. This paper proposes an entirely novel approach to the mathematical foundations of swarm algorithms: contrary to the current stochastic approaches, we show that an alternative deterministic model exists, which has its origin in deterministic chaos theory. We establish a reactive multi-agent system, based on logistic nonlinear decision maps, and designed according to the influence-reaction scheme. The rewriting of the decision functions leads to a new way of understanding the swarm phenomena in terms of state synchronization, and enables the analysis of their convergence behavior through bifurcation diagrams. We apply our approach on two concrete examples of each algorithm class, in order to demonstrate its general applicability.

This paper presents a decision making system for semiconductor wafer fabrication facilities, or wafer fabs, with hard inter-operation temporal constraints. The decision making system is developed based on a multi-agent architecture that is composed of scheduling agents, workcell agents, machine agents, and product agents. The decision making problem is to allocate lots into each workcell to satisfy both logical and temporal constraints. A dynamic planning-based approach is adopted for the decision making mechanism so that the dynamic behaviors of the wafer fab such as aperiodic lot arrivals and reconfiguration can be taken into consideration. The scheduling agents compute quasi-optimal schedules through a bidding mechanism with the workcell agents. The proposed decision making mechanism uses a concept of temporal constraint sets to obtain a feasible schedule in polynomial steps. The computational complexity of the decision making mechanism is proven to be O(Λ^3⋅L), where Λ is the number of operations of a lot and L is the cardinality of the temporal constraint set. However, there has been little research that addresses hard temporal constraints between operations in wafer fabs. This paper proposes a novel real-time decision making method that deals with hard inter-operation temporal constraints. This paper suggests an agent-based architecture for the decision making, and presents a real-time scheduling algorithm to generate quasi-optimal schedules. The proposed scheduling algorithm consists of two procedures: FEASIBLE_SPACE and OPTIMAL_SCHED. The former finds a feasible solution space for a newly inserted lot, whereas the latter computes the optimal solution among the feasible solution space. Simulation results reveal that the proposed decision making method has sufficiently low computation time for real-time applications, and that it is effective in increasing the throughput rate of the system.

Abstract. The use of expressive marks on musical scores (eg, lamen-toso, affetuoso, etc.) documents well the common assumption that emo-tions play an important role in music performance. This paper describes a model for simulating emotional behaviour in autonomous agents, for ...

Some interpreters use an intermediate machine code, byte-code, for program execution. This paper surveys techniques used to implement instruction dispatching in byte-code interpreters. For comparison purpose, we implemented a sample interpreter which includes dioeerent instruction dispatch techniques. In this article, we compare machine and compiler dependencies and the efficiency of these methods and analyse the results.

In this research, we dene a cooperative multiagent system where the agents use locally designed Bayesian networks to represent their knowledge. Agents communicate via message passing where the messages are beliefs in shared variables that are represented as probability distributions. Messages are treated as soft evidence in the receiver agents, where the belief in the receiving agent is replaced by the publishing agent’s belief. We call this the oracular assumption, where one agent is an expert or more knowledgeable of particular variables. As a result, the agents are organized in a publishersubscriber hierarchy. We compare and contrast our system with the MSBN multiagent model. Finally, we implement a multiagent system for experimentation using our multiagent system and MSBNs. We devise performance measures to compare the two systems. From this comparison, we provide guidance for the design of probabilistic multiagent systems.

We define a cooperative multiagent system where the agents use locally designed Bayesian networks to represent their knowledge. Agents in the newly defined AEBN model communicate via message passing where the messages are beliefs in shared variables that are represented as probability distributions. Messages are treated as soft evidence in the receiver agents, where the belief in the receiving agent is replaced by the publishing agent's belief. We call this the oracular assumption, where one agent is an expert or more knowledgeable of particular variables. As a result, the agents are organized in a publisher-subscriber hierarchy. A central problem of message passing in probabilistic systems is the so called rumor problem, where cycles in message passing cause redundant influence of beliefs. In this paper, algorithms to identify and solve the rumor problem in the context of our multiagent system are presented. The AEBN model is compared and contrasted with the MSBN multiagent model. We implement several multiagent systems for experimentation using our AEBN multiagent system and the MSBN system and we compare them using several performance measures. From this comparison, we provide guidance for the design of probabilistic multiagent systems.

We present a multiagent organization for data interpretation and fusion in which each agent uses an encapsulated Bayesian network for knowledge representation, and agents communicate by exchanging beliefs (marginal posterior probabilities) on shared variables. We call this organization an Agent-Encapsulated Bayesian Network (AEBN) system. Communication of probabilities among agents leads to rumors, i.e. potential double counting of information. We propose a new and correct method to compensate for rumors in AEBN systems by passing extended messages that contain joint probabilities.

Se propone una evaluación basada en Big Data de la interacción entre agentes conversacionales pedagógicos y estudiantes para facilitar la integración de este recurso TIC en las aulas. Así, se espera conseguir que la sociedad pueda aprovecharse de sus beneficios.

Information fusion, in the context of the Generative AI era, must distinguish AI Agents from Agentic AI. This review critically distinguishes between AI Agents and Agentic AI, offering a structured, conceptual taxonomy, application mapping, and analysis of opportunities and challenges to clarify their divergent design philosophies and capabilities. We begin by outlining the search strategy and foundational definitions, characterizing AI Agents as modular systems driven and enabled by LLMs and LIMs for task-specific automation. Generative AI is positioned as a precursor providing the foundation, with AI agents advancing through tool integration, prompt engineering, and reasoning enhancements. We then characterize Agentic AI systems, which, in contrast to AI Agents, represent a paradigm shift marked by multi-agent collaboration, dynamic task decomposition, persistent memory, and coordinated autonomy. Through a chronological evaluation of architectural evolution, operational mechanisms, interaction styles, and autonomy levels, we present a comparative analysis across both AI agents and agentic AI paradigms. Application domains enabled by AI Agents such as customer support, scheduling, and data summarization are then contrasted with Agentic AI deployments in research automation, robotic coordination, and medical decision support. We further examine unique challenges in each paradigm including hallucination, brittleness, emergent behavior, and coordination failure, and propose targeted solutions such as ReAct loops, retrieval-augmented generation (RAG), automation coordination layers, and causal modeling. This work aims to provide a roadmap for developing robust, scalable, and explainable AI-driven systems.

Symmetry breaking and implied constraints can speed up both exhaustive search and the search for a single solution. We experiment with both types of constraint, using three search algorithms (backtracking, local and hybrid) to find single solutions for SAT encodings of three combinatorial problems (clique, set cover and balanced incomplete block design generation). Both show strong positive and negative effects,

Resumen Con el avance de las tecnologías de la informacion se hace cada vez más facil generar y procesar datos con mayor precision acerca de lo que ocurre en un dominio complejo. Las aplicaciones que se desen vuelven en estos entornos se caracterizan por recibir grandes cantidades de datos en poco tiempo, con altas tasas de cambio, como así tambien in certidumbre e incompletitud; a este tipo de torrente de datos -que no se puede almacenar para ser procesado mas adelante -se los conoce con el nombre de "flujos" . En este trabajo presentamos una propuesta para rea lizar procesamiento de flujos de manera inteligente a traves de sistemas argumentativos, proponiendo una combinacion novedosa de (i) tecnicas de procesamiento de flujos de datos y eventos complejos, (ii) herramien tas de argumentacion estructurada, (iii) razonamiento probabilístico y (iv) operadores de revisioín de creencias. Palabras clave: Procesamiento de flujo de datos, Argumentaciín Es tructurada, Razonamiento bajo incertidumbre, Revision de creencias.

Developing computer game agents is often a lengthy and expensive undertaking. Detailed domain knowledge and decision-making procedures must be encoded into the agent to achieve realistic behavior. In this paper, we simplify this process by using the ICARUS cognitive architecture to construct game agents. The system acquires structured, high fidelity methods for agents that utilize a vocabulary of concepts familiar to game experts. We demonstrate our approach by first acquiring behaviors for football agents from video footage of college football games, and then applying the agents in a football simulator.

Computational simulation of research groups at Costa Rican Distance Education University. Social network analysis deals with quantitative studies of human groups and has established a set of measurements to understand their structure and behaviour. Nevertheless, their dynamism is only expressed to a limited extend and multiagent modeling and simulation has become a complementary analytical method. It enables researchers to simulate human interactions on virtual environments and to discover emergent patterns of behavior. This article presents a multiagent system that models and simulates the dynamics of research groups at the Costa Rican Distance Education University. The results of a first scenario point out that the initial amount of available resources in the system does not have a causal relationship with the clustering coefficient, i.e the metric used to measure the reseachers’ tendency to form groups. The concept of resource must be reconsidered to include infrastructure and equipment availability. A second scenario shows that as the number of researchers increases, the clustering coefficient decreases. This effect is atributed to a tendency to equilibrium in the relative costs and benefits of maintaining a given number of researchers. More variables are needed for a reliable simulation.

We explore how the architecture of a product evolves over several generations. We propose a theoretical framework and research approach to study the dynamics of complex product architectures. We illustrate our approach by examining the architecture of software products because they are complex, exhibit fast change rates (like fruit flies in studies of biological evolution), and offer (through their source code) an efficient, reliable, and standardized medium to capture their architecture. The IEEE defines product architecture as "the fundamental organization of a system embodied in its components, their relationships to each other, and to the environment, and the principles guiding its design and evolution" [1]. In the software domain, architecting involves organizing or structuring the code into modules and layers with the appropriate set of dependencies between them . This paper reports results from , in which we provide a theoretical framework, a basic set of metrics, and a research approach for exploring the dynamics of complex software architectures. Then, based on empirical evidence from a case study of an open source project, we uncover several patterns and insights regarding the dynamics of software architectures and their relationships to organizational dynamics. These findings indicate several promising avenues for future research. To explore the dynamics of complex software architectures, we structure our research approach in three steps: 1. Capture the evolution of software architecture properties. 2. Capture the evolution of organizational attributes. 3. Compare the dynamics of product architectures and organizational attributes. This abstract provides only a brief introduction to our approach. Our presentation will include metrics and results. Additional discussion is also available in .

The codification of human intuition and creative problem-solving represents a fundamental challenge in artificial intelligence research. Contemporary AI systems demonstrate exceptional performance in structured, logic-based reasoning tasks, however, their capacity to emulate human-like lateral thinking and intuitive insight remains significantly constrained. The chess grandmaster's ability to instantly evaluate positions through pattern recognition. Being able to process millions of learned configurations to generate intuitive judgments in milliseconds. This provides a concrete computational model for encoding expert intuition. This paper demonstrates how the hierarchical pattern libraries, evaluation heuristics, and position-based reasoning that characterize chess expertise can be systematically translated into the IIA framework's architecture, creating a rigorous bridge between human intuitive expertise and AI-driven innovation. This paper presents a conceptual framework for an Intuitive Inventive Agent (IIA) that bridges this gap by integrating the Theory of Inventive Problem Solving (TRIZ) with lateral thinking methodologies, operationalized through state-of-the-art Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and orchestration architectures employing LangGraph and n8n. This approach synthesizes three key innovations: (1) Automation of TRIZ's systematic contradiction resolution through LLM-powered semantic analysis and deterministic principle retrieval, (2) Incorporation of Edward de Bono's lateral thinking techniques to expand solution spaces beyond conventional boundaries, and (3) Implementation of an Agent orchestration architecture that coordinates multi-agent workflows while maintaining transparency and human interpretability. The framework leverages insights from competitive chess expertise to model rapid, experience-driven judgment patterns characteristic of expert intuition, addressing the distinction between Kahneman's System 1 (fast, intuitive) and System 2 (slow, analytical) cognitive processes. The IIA concept framework advances the state-of-the-art by providing the first unified automation of structured inventive problem-solving and creative lateral thinking within a single computational architecture. This work contributes to the broader goal of developing AI systems that not only process information but demonstrate wisdom. This is achieved by the integration of knowledge, experience, and contextual judgment that characterizes human expertise.

GORAN TRAJKOVSKI, GEORGI STOJANOV, SAMUEL COLLINS , VLADIMIR EIDELMAN, CHRIS. HARMAN AND GIOVANNI VINCENTI 1 Computer and Information Sciences Department, Towson University, 8000 York Road, Towson, MD 21252, USA, E-mail: {gtrajkovski, gvince1}@towson.edu, Phone 1-410704-6310, Fax 1-410-704-3868 2 Department of Computer Science, Mathematics and Science, American University in Paris, 147, Rue de Grenelle, Paris 75007, France, E-mail: gstojanov@aup.fr 3 Department of Sociology, Anthropology and Criminal Justice, Towson University, 8000 York Road, Towson, MD 21252, USA, E-mail: scollins@towson.edu 4 Computer Science Department, Columbia University, 2960 Broadway, New York, NY 10027, USA, E-mail: vae2101@columbia.edu 5 Computer Science Department, Swarthmore College, 500 College Ave, Swarthmore, PA 19081, USA, E-mail: charman1@swarthmore.edu

Autonomous operation of UAS holds promise for greater productivity of atmospheric science missions. However, several challenges need to be overcome before such missions can be made autonomous. This paper presents a framework for safe autonomous operations of multiple vehicles, particularly suited for atmospheric science missions. The framework revolves around the use of piecewise Bezier curves for trajectory representation, which in conjunction with path-following and time-coordination algorithms, allows for safe coordinated operations of multiple vehicles.

With the anticipated explosion of small unmanned aerial vehicles, it is highly likely that operators will be controlling fleets of autonomous vehicles. To fulfill the promise of autonomy, vehicle operators will not be concerned with manual control of the vehicle; instead, they will deal with the overall mission. Furthermore, the one operator to many vehicles is becoming a constant meme with various industries including package delivery, search and rescue, and utility companies. In order for an operator to concurrently control several vehicles, his station must look and behave very differently than the current ground control station instantiations. Furthermore, the vehicle will have to be much more autonomous, especially during non-normal operations, in order to accommodate the knowledge deficit or the information overload of the operator in charge of several vehicles. The expected usage increase of small drones requires presenting the operational information generated by a fleet of heterogeneous autonomous agents to an operator. NASA Langley Research Center's Autonomy Incubator has brought together researchers in various disciplines including controls, trajectory planning, systems engineering, and human factors to develop an integrated system to study autonomy issues. The initial human factors effort is focusing on mission displays that would give an operator the overall status of all autonomous agents involved in the current mission. This paper will discuss the specifics of the mission displays for operators controlling several vehicles.

This presentation will give an overview of a time-critical coordination control law to support multi-vehicle missions that impose both absolute and relative temporal specifications on the trajectories of the vehicles. The proposed cooperative strategy yields robust behavior against external disturbances by allowing the vehicles to negotiate their speeds along the paths in response to information exchanged over an inter-vehicle communication network. Simulation results of a sequential auto-landing scenario will be presented to demonstrate the efficacy of the control law for group coordination. The use of teams of autonomous, cooperating unmanned aerial vehicles (UAVs) is expected to bring unprecedented levels of flexibility, efficiency, and reliability to the execution of a variety of mission. Reconnaissance, surveillance, search and rescue and 3D terrain modeling are just a few examples of the missions that can benefit from such technology. In addition, advances in UAV capabilities are expected to impact other domains; such as air traffic management, where sense and avoid technologies have the potential to reduce the workload at control centers and provide safer management of the airspace. Some of the mentioned missions require the UAVs to maintain a desired inter-vehicle schedule without a fixed arrival time. This type of requirements falls under the category of relative temporal constraints. However, some other missions impose absolute temporal constraints. In this case, the UAVs must not only coordinate with each other, but also arrive at their final destination at a pre-specified time or within a desired window of arrival. The cooperative missions considered in require that the fleet maintain a desired relative timing schedule that ensures that all vehicles execute collisionfree maneuvers. Some missions, however, also impose absolute temporal constraints on the trajectories of the vehicles, in addition to relative temporal assignments. Representative examples are the calibration and validation of satellite instruments for remote sensing, and sequential auto-landing within desired temporal slots. The first refers to the situation in which a formation of aircraft must cooperatively fly a set of predefined trajectories at different levels within the atmosphere, while coordinating their maneuvers with an Earth-observing satellite. In the case of sequential auto-landing, a fleet of aerial vehicles must break up and arrive at the assigned glide path within a desired temporal window and separated by pre-specified safe-guarding time-intervals. In this presentation, we will introduce an extension of the coordination algorithm in [1] (and its multi-leader version adopted in ) to the case of absolute temporal specifications.

Cooperation is a widespread phenomenon in nature that has also been a cornerstone in the development of human intelligence. Understanding cooperation, therefore, on matters such as how it emerges, develops, or fails is an important avenue of research, not only in a human context, but also for the advancement of next generation artificial intelligence paradigms which are presumably human-compatible. With this motivation in mind, we study the emergence of cooperative behaviour between two independent deep reinforcement learning (RL) agents provided with human input in a novel game environment. In particular, we investigate whether evaluative human feedback (through interactive RL) and expert demonstration (through inverse RL) can help RL agents to learn to cooperate better. We report two main findings. Firstly, we find that the amount of feedback given has a positive impact on the accumulated reward obtained through cooperation. That is, agents trained with a limited amount of feedbac...

This article reports on the CBRecSys 2015 workshop, the second edition of the workshop on new trends in content-based recommender systems, co-located with RecSys 2015 in Vienna, Austria. Content-based recommendation has been applied successfully in many different domains, but it has not seen the same level of attention as collaborative filtering techniques have. Nevertheless, there are many recommendation domains and applications where content and metadata play a key role, either in addition to or instead of ratings and implicit usage data. The CBRecSys workshop series provides a dedicated venue for work dedicated to all aspects of content-based recommender systems.

1 Laboratorio de Investigación y Desarrollo en Inteligencia Artificial (LIDIA), Depto. de Ciencias e Ingenieŕıa de la Computación, Universidad Nacional del Sur, Bah́ıa Blanca, Argentina. 2 Consejo Nacional de Investigaciones Cient́ıficas y Técnicas (CONICET), Buenos Aires, Argentina. 3 Depto. de Matemática, Universidad Nacional de Santiago del Estero, Santiago del Estero, Argentina. e-mail: {mlc,mcdb,sag,dcm,grs}@cs.uns.edu.ar

A crescente inserção dos ambientes virtuais no processo de aprendizagem vem trazendo novos desafios a área de Informática na Educação. Entre eles, está o projeto de ambientes que facilitam ao aluno o acesso ao grande volume de informações. Neste contexto, este artigo apresenta um assistente pessoal cujo objetivo é dar apoio aos alunos na utilização dos ambientes virtuais de aprendizagem, fornecendo o suporte apropriado a cada perfil de aluno. O assistente pessoal, modelado em uma abordagem multiagente, faz parte de uma sociedade na qual cada agente controla uma funcionalidade específica, e um agente mediador define qual deles deve entrar em ação a cada momento.

This paper presents some results in the intersection of three technological fields: e-learning, multiagent systems, and standards to improve the development of secure systems. There is a consensus that security is a critical issue for distributed, highly interactive and open learning environments. While a lot of effort in the e-learning domain has been put into delivering infrastructure and providing content, security issues have hardly ever been considered. Agent-oriented methodologies may bring benefits to the conception of multiagent systems, since they gradually are incorporating specifications for general security mechanisms (FIPA standards). Concerning the specification of security requirements for e-learning environments, extended formalisms (such as security use cases) appear as proper modeling tools. The topics covered by this paper converge to the implementation of a multiagent platform (the PMA3), which constitutes an open distributed infrastructure addressing important security issues. PMA3 has been used in the development of elearning environments. The paper concludes with an outlook to ongoing research efforts related to security requirements for e-learning applications.

This paper outlines an approximate algorithm for finding an optimal decentralized control in multi-agent systems. Decentralized Partially Observable Markov Decision Processes and their extension to infinite state, observation and action spaces are utilized as a theoretical framework. In the presented algorithm, policies of each agent are represented by a feedforward neural network. Then, a search is performed in a joint weight space of all networks. Particle Swarm Optimization is applied as a search algorithm. Experimental results are provided showing that the algorithm finds good solutions for the classical Tiger Problem extended to multi-agent systems, as well as for a multi-agent navigation task involving large state and action spaces.

Present space robots such as planetary robots and flexible robots have structural flexibility in their arms and joints that leads to an error in the tip positioning owing to tip deflection. The flexible-link manipulator (FLM) is a non-collocated system that has unstable and inaccurate system performance. Thus, tip-tracking of FLM possesses difficult control challenges. The purpose of this study is to design adaptive intelligent tip-tracking control strategy for FLMs to deal with this control challenges of FLM. A vision sensor is utilized in conjunction with a traditional mechanical sensor to directly measure tip-position in order to address the aforementioned problem. Image-based visual servoing (IBVS), one of several visual servoing control techniques, is more efficient. However, the IBVS scheme faces numerous difficulties that impair the system’s performance in real-time applications, including singularities in the interaction matrix, local minima in trajectory, visibility issues....

Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number.

Phat is an intriguing card game played in several places in Britain whose history has been obscure. Co-authored by Paul Eaton and George Witty, the paper uncovers the invention of the game by English and Irish soldiers in the Portuguese Civil War of the early 19th century and traces its development on both sides of the Atlantic before concluding with examples of how it is still played today in East Anglia.

We present a unified framework where spacetime, gravity, and dynamics emerge from quantum information. We start from a complete quantum graph whose nodes are specific two-qubit clusters in CP 3 and whose edges are weighted by Fubini-Study distances. Locality emerges topologically, dimensionality is detected algorithmically, and geometric tensions are generated by differences between amplitude probabilities. The system's evolution is governed by a minimization principle of an Informational Free Energy, F = U-τ S, giving rise to a dual interaction: an attraction derived from quantum state distinguishability and an entropic repulsion that prevents singularities. We construct the metric tensor directly from the graph's statistical correlations. A significant contribution of this work is the formal demonstration that the holographic principle emerges naturally as a mathematical consequence of the thermodynamic equilibrium, rather than as an external postulate. Numerical simulations confirm the robust emergence of a Lorentzian manifold with a stable, non-integer dimension, suggesting a fractal spacetime structure at the fundamental scale. This demonstrates the framework's viability and offers a testable path toward a quantum gravity theory.