Aprendizado por Reforço com Valores deInfluência em Sistemas Multi-Agente

We propose a new paradigm for collective learning in multi-agent systems (MAS) as a solution to the problem in which several agents acting over the same environment must learn how to perform tasks, simultaneously, based on feedbacks given by each one of the other agents. We introduce the proposed paradigm in the form of a reinforcement learning algorithm, nominating it as reinforcement learning with influence values. While learning by rewards, each agent evaluates the relation between the current state and/or action executed at this state (actual believe) together with the reward obtained after all agents that are interacting perform their actions. The reward is a result of the interference of others. The agent considers the opinions of all its colleagues in order to attempt to change the values of its states and/or actions. The idea is that the system, as a whole, must reach an equilibrium, where all agents get satisfied with the obtained results. This means that the values of the state/actions pairs match the reward obtained by each agent. This dynamical way of setting the values for states and/or actions makes this new reinforcement learning paradigm the first to include, naturally, the fact that the presence of other agents in the environment turns it a dynamical model. As a direct result, we implicitly include the internal state, the actions and the rewards obtained by all the other agents in the internal state of each agent. This makes our proposal the first complete solution to the conceptual problem that rises when applying reinforcement learning in multi-agent systems, which is caused by the difference existent between the environment and agent models. With basis on the proposed model, we create the IVQ-learning algorithm that is exhaustive tested in repetitive games with two, three and four agents and in stochastic games that need cooperation and in games that need collaboration. This algorithm shows to be a good option for obtaining solutions that guarantee convergence to the Nash optimum equilibrium in cooperative problems. Experiments performed clear shows that the proposed paradigm is theoretical and experimentally superior to the traditional approaches. Yet, with the creation of this new paradigm the set of reinforcement learning applications in MAS grows up. That is, besides the possibility of applying the algorithm in traditional learning problems in MAS, as for example coordination of tasks in multi-robot systems, it is possible to apply reinforcement learning in problems that are essentially collaborative

Application of self-organizing techniques for the distribution of heterogeneous multi-tasks in multi-robot systems

This paper focuses on the general problem of coordinating of multi-robot systems, more specifically, it addresses the self-election of heterogeneous and specialized tasks by autonomous robots. In this regard, it has proposed experimenting with two different techniques based chiefly on selforganization and emergence biologically inspired, by applying response threshold models as well as ant colony optimization. Under this approach it can speak of multi-tasks selection instead of multi-tasks allocation, that means, as the agents or robots select the tasks instead of being assigned a task by a central controller. The key element in these algorithms is the estimation of the stimuli and the adaptive update of the thresholds. This means that each robot performs this estimate locally depending on the load or the number of pending tasks to be performed. It has evaluated the robustness of the algorithms, perturbing the number of pending loads to simulate the robot’s error in estimating the real number of pending tasks and also the dynamic generation of loads through time. The paper ends with a critical discussion of experimental results.

Neural Approach in Multi-Agent Routing for Static Telecommunication Networks

The problem of multi-agent routing in static telecommunication networks with fixed configuration is considered. The problem is formulated in two ways: for centralized routing schema with the coordinator-agent (global routing) and for distributed routing schema with independent agents (local routing). For both schemas appropriate Hopfield neural networks (HNN) are constructed.

Teams leading teams: Examining the role of leadership in multi-team systems

A major challenge of modern teams lies in the coordination of the efforts not just of individuals within a team, but also of teams whose efforts are ultimately entwined with those of other teams. Despite this fact, much of the research on work teams fails to consider the external dependencies that exist in organizational teams and instead focuses on internal or within team processes. Multi-Team Systems Theory is used as a theoretical framework for understanding teams-of-teams organizational forms (Multi-Team Systems; MTS's); and leadership teams are proposed as one remedy that enable MTS members to dedicate needed resources to intra-team activities while ensuring effective synchronization of between-team activities. Two functions of leader teams were identified: strategy development and coordination facilitation; and a model was developed delineating the effects of the two leader roles on multi-team cognitions, processes, and performance.^ Three hundred eighty-four undergraduate psychology and business students participated in a laboratory simulation that modeled an MTS; each MTS was comprised of three, two-member teams each performing distinct but interdependent components of an F-22 battle simulation task. Two roles of leader teams supported in the literature were manipulated through training in a 2 (strategy training vs. control) x 2 (coordination training vs. control) design. Multivariate analysis of variance (MANOVA) and mediated regression analysis were used to test the study's hypotheses. ^ Results indicate that both training manipulations produced differences in the effectiveness of the intended form of leader behavior. The enhanced leader strategy training resulted in more accurate (but not more similar) MTS mental models, better inter-team coordination, and higher levels of multi-team (but not component team) performance. Moreover, mental model accuracy fully mediated the relationship between leader strategy and inter-team coordination; and inter-team coordination fully mediated the effect of leader strategy on multi-team performance. Leader coordination training led to better inter-team coordination, but not to higher levels of either team or multi-team performance. Mediated Input-Process-Output (I-P-O) relationships were not supported with leader coordination; rather, leader coordination facilitation and inter-team coordination uniquely contributed to component team and multi-team level performance. The implications of these findings and future research directions are also discussed. ^

Collective intelligence on dynamic manufacturing scheduling optimization

Swarm Intelligence (SI) is a growing research field of Artificial Intelligence (AI). SI is the general term for several computational techniques which use ideas and get inspiration from the social behaviours of insects and of other animals. This paper presents hybridization and combination of different AI approaches, like Bio-Inspired Techniques (BIT), Multi-Agent systems (MAS) and Machine Learning Techniques (ML T). The resulting system is applied to the problem of jobs scheduling to machines on dynamic manufacturing environments.

Aspectos de aprendizagem em optimização

A optimização e a aprendizagem em Sistemas Multi-Agente são consideradas duas áreas promissoras mas relativamente pouco exploradas. A optimização nestes ambientes deve ser capaz de lidar com o dinamismo. Os agentes podem alterar o seu comportamento baseando-se em aprendizagem recente ou em objectivos de optimização. As estratégias de aprendizagem podem melhorar o desempenho do sistema, dotando os agentes da capacidade de aprender, por exemplo, qual a técnica de optimização é mais adequada para a resolução de uma classe particular de problemas, ou qual a parametrização é mais adequada em determinado cenário. Nesta dissertação são estudadas algumas técnicas de resolução de problemas de Optimização Combinatória, sobretudo as Meta-heurísticas, e é efectuada uma revisão do estado da arte de Aprendizagem em Sistemas Multi-Agente. É também proposto um módulo de aprendizagem para a resolução de novos problemas de escalonamento, com base em experiência anterior. O módulo de Auto-Optimização desenvolvido, inspirado na Computação Autónoma, permite ao sistema a selecção automática da Meta-heurística a usar no processo de optimização, assim como a respectiva parametrização. Para tal, recorreu-se à utilização de Raciocínio baseado em Casos de modo que o sistema resultante seja capaz de aprender com a experiência adquirida na resolução de problemas similares. Dos resultados obtidos é possível concluir da vantagem da sua utilização e respectiva capacidade de adaptação a novos e eventuais cenários.

Multi-Agent Based Smart Grid Management and Simulation: Situation Awareness and Learning in a Test Bed with Simulated and Real Installations and Players

The rising usage of distributed energy resources has been creating several problems in power systems operation. Virtual Power Players arise as a solution for the management of such resources. Additionally, approaching the main network as a series of subsystems gives birth to the concepts of smart grid and micro grid. Simulation, particularly based on multi-agent technology is suitable to model all these new and evolving concepts. MASGriP (Multi-Agent Smart Grid simulation Platform) is a system that was developed to allow deep studies of the mentioned concepts. This paper focuses on a laboratorial test bed which represents a house managed by a MASGriP player. This player is able to control a real installation, responding to requests sent by the system operators and reacting to observed events depending on the context.

A new approach for multi-agent coalition formation and management in the scope of electricity markets

This paper presents a new methodology for the creation and management of coalitions in Electricity Markets. This approach is tested using the multi-agent market simulator MASCEM, taking advantage of its ability to provide the means to model and simulate VPP (Virtual Power Producers). VPPs are represented as coalitions of agents, with the capability of negotiating both in the market, and internally, with their members, in order to combine and manage their individual specific characteristics and goals, with the strategy and objectives of the VPP itself. The new features include the development of particular individual facilitators to manage the communications amongst the members of each coalition independently from the rest of the simulation, and also the mechanisms for the classification of the agents that are candidates to join the coalition. In addition, a global study on the results of the Iberian Electricity Market is performed, to compare and analyze different approaches for defining consistent and adequate strategies to integrate into the agents of MASCEM. This, combined with the application of learning and prediction techniques provide the agents with the ability to learn and adapt themselves, by adjusting their actions to the continued evolving states of the world they are playing in.

Cooperation mechanism for team-work based multi-agent system in dynamic scheduling through meta-heuristics

This paper describes a Multi-agent Scheduling System that assumes the existence of several Machines Agents (which are decision-making entities) distributed inside the Manufacturing System that interact and cooperate with other agents in order to obtain optimal or near-optimal global performances. Agents have to manage their internal behaviors and their relationships with other agents via cooperative negotiation in accordance with business policies defined by the user manager. Some Multi Agent Systems (MAS) organizational aspects are considered. An original Cooperation Mechanism for a Team-work based Architecture is proposed to address dynamic scheduling using Meta-Heuristics.

Multi-agent Simulation of Bilateral Contracting in Competitive Electricity Markets

Traditional vertically integrated power utilities around the world have evolved from monopoly structures to open markets that promote competition among suppliers and provide consumers with a choice of services. Market forces drive the price of electricity and reduce the net cost through increased competition. Electricity can be traded in both organized markets or using forward bilateral contracts. This article focuses on bilateral contracts and describes some important features of an agent-based system for bilateral trading in competitive markets. Special attention is devoted to the negotiation process, demand response in bilateral contracting, and risk management. The article also presents a case study on forward bilateral contracting: a retailer agent and a customer agent negotiate a 24h-rate tariff. © 2014 IEEE.

Multi-agent simulation of bilateral contracting in competitive electricity markets

Traditional vertically integrated power utilities around the world have evolved from monopoly structures to open markets that promote competition among suppliers and provide consumers with a choice of services. Market forces drive the price of electricity and reduce the net cost through increased competition. Electricity can be traded in both organized markets or using forward bilateral contracts. This article focuses on bilateral contracts and describes some important features of an agent-based system for bilateral trading in competitive markets. Special attention is devoted to the negotiation process, demand response in bilateral contracting, and risk management. The article also presents a case study on forward bilateral contracting: a retailer agent and a customer agent negotiate a 24h-rate tariff. © 2014 IEEE.

Upper Ontology for Multi-Agent Energy Systems’ Applications

Energy systems worldwide are complex and challenging environments. Multi-agent based simulation platforms are increasing at a high rate, as they show to be a good option to study many issues related to these systems, as well as the involved players at act in this domain. In this scope the authors research group has developed three multi-agent systems: MASCEM, which simulates the electricity markets; ALBidS that works as a decision support system for market players; and MASGriP, which simulates the internal operations of smart grids. To take better advantage of these systems, their integration is mandatory. For this reason, is proposed the development of an upper-ontology which allows an easier cooperation and adequate communication between them. Additionally, the concepts and rules defined by this ontology can be expanded and complemented by the needs of other simulation and real systems in the same areas as the mentioned systems. Each system’s particular ontology must be extended from this top-level ontology.