A novel adaptive weight selection algorithm for multi-objective multi-agent reinforcement learning

To solve multi-objective problems, multiple reward signals are often scalarized into a single value and further processed using established single-objective problem solving techniques. While the field of multi-objective optimization has made many advances in applying scalarization techniques to obtain good solution trade-offs, the utility of applying these techniques in the multi-objective multi-agent learning domain has not yet been thoroughly investigated. Agents learn the value of their decisions by linearly scalarizing their reward signals at the local level, while acceptable system wide behaviour results. However, the non-linear relationship between weighting parameters of the scalarization function and the learned policy makes the discovery of system wide trade-offs time consuming. Our first contribution is a thorough analysis of well known scalarization schemes within the multi-objective multi-agent reinforcement learning setup. The analysed approaches intelligently explore the weight-space in order to find a wider range of system trade-offs. In our second contribution, we propose a novel adaptive weight algorithm which interacts with the underlying local multi-objective solvers and allows for a better coverage of the Pareto front. Our third contribution is the experimental validation of our approach by learning bi-objective policies in self-organising smart camera networks. We note that our algorithm (i) explores the objective space faster on many problem instances, (ii) obtained solutions that exhibit a larger hypervolume, while (iii) acquiring a greater spread in the objective space.

Multi-apprentice learning for meta-heuristics parameter tuning in a multi agent scheduling system

The scheduling problem is considered in complexity theory as a NP-hard combinatorial optimization problem. Meta-heuristics proved to be very useful in the resolution of this class of problems. However, these techniques require parameter tuning which is a very hard task to perform. A Case-based Reasoning module is proposed in order to solve the parameter tuning problem in a Multi-Agent Scheduling System. A computational study is performed in order to evaluate the proposed CBR module performance.

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.

Experience generalization for multi-agent reinforcement learning

On-line learning methods have been applied successfully in multi-agent systems to achieve coordination among agents. Learning in multi-agent systems implies in a non-stationary scenario perceived by the agents, since the behavior of other agents may change as they simultaneously learn how to improve their actions. Non-stationary scenarios can be modeled as Markov Games, which can be solved using the Minimax-Q algorithm a combination of Q-learning (a Reinforcement Learning (RL) algorithm which directly learns an optimal control policy) and the Minimax algorithm. However, finding optimal control policies using any RL algorithm (Q-learning and Minimax-Q included) can be very time consuming. Trying to improve the learning time of Q-learning, we considered the QS-algorithm. in which a single experience can update more than a single action value by using a spreading function. In this paper, we contribute a Minimax-QS algorithm which combines the Minimax-Q algorithm and the QS-algorithm. We conduct a series of empirical evaluation of the algorithm in a simplified simulator of the soccer domain. We show that even using a very simple domain-dependent spreading function, the performance of the learning algorithm can be improved.

Cooperative Learning Model based on Multi-Agent Architecture for Embedded Intelligent Systems

Cooperative systems are suitable for many types of applications and nowadays these system are vastly used to improve a previously defined system or to coordinate multiple devices working together. This paper provides an alternative to improve the reliability of a previous intelligent identification system. The proposed approach implements a cooperative model based on multi-agent architecture. This new system is composed of several radar-based systems which identify a detected object and transmit its own partial result by implementing several agents and by using a wireless network to transfer data. The proposed topology is a centralized architecture where the coordinator device is in charge of providing the final identification result depending on the group behavior. In order to find the final outcome, three different mechanisms are introduced. The simplest one is based on majority voting whereas the others use two different weighting voting procedures, both providing the system with learning capabilities. Using an appropriate network configuration, the success rate can be improved from the initial 80% up to more than 90%.

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.

Case-based reasoning for meta-heuristics self-parameterization in a multi-agent scheduling system

A novel agent-based approach to Meta-Heuristics self-configuration is proposed in this work. Meta-heuristics are examples of algorithms where parameters need to be set up as efficient as possible in order to unsure its performance. This paper presents a learning module for self-parameterization of Meta-heuristics (MHs) in a Multi-Agent System (MAS) for resolution of scheduling problems. The learning is based on Case-based Reasoning (CBR) and two different integration approaches are proposed. A computational study is made for comparing the two CBR integration perspectives. In the end, some conclusions are reached and future work outlined.

Meta-heuristics self-parameterization in a multi-agent scheduling system using case-based reasoning

This paper proposes a novel agent-based approach to Meta-Heuristics self-configuration. Meta-heuristics are algorithms with parameters which need to be set up as efficient as possible in order to unsure its performance. A learning module for self-parameterization of Meta-heuristics (MH) in a Multi-Agent System (MAS) for resolution of scheduling problems is proposed in this work. The learning module is based on Case-based Reasoning (CBR) and two different integration approaches are proposed. A computational study is made for comparing the two CBR integration perspectives. Finally, some conclusions are reached and future work outlined.

Scenarios Generation for Multi-Agent simulation of Electricity Markets based on Intelligent Data Analysis

This document presents a tool able to automatically gather data provided by real energy markets and to generate scenarios, capture and improve market players’ profiles and strategies by using knowledge discovery processes in databases supported by artificial intelligence techniques, data mining algorithms and machine learning methods. It provides the means for generating scenarios with different dimensions and characteristics, ensuring the representation of real and adapted markets, and their participating entities. The scenarios generator module enhances the MASCEM (Multi-Agent Simulator of Competitive Electricity Markets) simulator, endowing a more effective tool for decision support. The achievements from the implementation of the proposed module enables researchers and electricity markets’ participating entities to analyze data, create real scenarios and make experiments with them. On the other hand, applying knowledge discovery techniques to real data also allows the improvement of MASCEM agents’ profiles and strategies resulting in a better representation of real market players’ behavior. This work aims to improve the comprehension of electricity markets and the interactions among the involved entities through adequate multi-agent simulation.

Demonstration of the Multi-Agent Simulator of Competitive Electricity Markets

Electricity markets are complex environments with very particular characteristics. A critical issue concerns the constant changes they are subject to. This is a result of the electricity markets’ restructuring, performed so that the competitiveness could be increased, but with exponential implications in the increase of the complexity and unpredictability in those markets’ scope. The constant growth in markets unpredictability resulted in an amplified need for market intervenient entities in foreseeing market behavior. The need for understanding the market mechanisms and how the involved players’ interaction affects the outcomes of the markets, contributed to the growth of usage of simulation tools. Multi-agent based software is particularly well fitted to analyze dynamic and adaptive systems with complex interactions among its constituents, such as electricity markets. This paper presents the Multi-Agent System for Competitive Electricity Markets (MASCEM) – a simulator based on multi-agent technology that provides a realistic platform to simulate electricity markets, the numerous negotiation opportunities and the participating entities.

Data Mining Approach to support the Generation of Realistic Scenarios for Multi-Agent simulation of Electricity Markets

This paper presents the Realistic Scenarios Generator (RealScen), a tool that processes data from real electricity markets to generate realistic scenarios that enable the modeling of electricity market players’ characteristics and strategic behavior. The proposed tool provides significant advantages to the decision making process in an electricity market environment, especially when coupled with a multi-agent electricity markets simulator. The generation of realistic scenarios is performed using mechanisms for intelligent data analysis, which are based on artificial intelligence and data mining algorithms. These techniques allow the study of realistic scenarios, adapted to the existing markets, and improve the representation of market entities as software agents, enabling a detailed modeling of their profiles and strategies. This work contributes significantly to the understanding of the interactions between the entities acting in electricity markets by increasing the capability and realism of market simulations.

Social Presence Approach Within the Question and Answering eLearning Model: An Experiment with a Multi-Agent System

The model of Questions Answering (Q&A) for eLearning is based on collaborative learning through questions that are posed by students and their answers to that questions which are given by peers, in contrast with the classical model in which students ask questions to the teacher only. In this proposal we extend the Q&A model including the social presence concept and a quantitative measure of it is proposed; besides it is considered the evolution of the resulting Q&A social network after the inclusion of the social presence and taking into account the feedback on questions posed by students and answered by peers. The social network behaviorwas simulated using a Multi-Agent System to compare the proposed social presence model with the classical and the Q&A models

A Multi-agent System for Emergency Decision Support

This paper describes the multi-agent organization of a computer system that was designed to assist operators in decision making in the presence of emergencies. The application was developed for the case of emergencies caused by river floods. It operates on real-time receiving data recorded by sensors (rainfall, water levels, flows, etc.) and applies multi-agent techniques to interpret the data, predict the future behavior and recommend control actions. The system includes an advanced knowledge based architecture with multiple symbolic representation with uncertainty models (bayesian networks). This system has been applied and validated at two particular sites in Spain (the Jucar basin and the South basin).

A hybrid approach to multi-agent decision-making

In the aftermath of a large-scale disaster, agents' decisions derive from self-interested (e.g. survival), common-good (e.g. victims' rescue) and teamwork (e.g. fire extinction) motivations. However, current decision-theoretic models are either purely individual or purely collective and find it difficult to deal with motivational attitudes; on the other hand, mental-state based models find it difficult to deal with uncertainty. We propose a hybrid, CvI-JI, approach that combines: i) collective 'versus' individual (CvI) decisions, founded on the Markov decision process (MDP) quantitative evaluation of joint-actions, and ii)joint-intentions (JI) formulation of teamwork, founded on the belief-desire-intention (BDI) architecture of general mental-state based reasoning. The CvI-JI evaluation explores the performance's improvement

RFID together with multi-agent systems to control global value chains

Nowadays, the cooperative intelligent transport systems are part of a largest system. Transportations are modal operations integrated in logistics and, logistics is the main process of the supply chain management. The supply chain strategic management as a simultaneous local and global value chain is a collaborative/cooperative organization of stakeholders, many times in co-opetition, to perform a service to the customers respecting the time, place, price and quality levels. The transportation, like other logistics operations must add value, which is achieved in this case through compression lead times and order fulfillments. The complex supplier's network and the distribution channels must be efficient and the integral visibility (monitoring and tracing) of supply chain is a significant source of competitive advantage. Nowadays, the competition is not discussed between companies but among supply chains. This paper aims to evidence the current and emerging manufacturing and logistics system challenges as a new field of opportunities for the automation and control systems research community. Furthermore, the paper forecasts the use of radio frequency identification (RFID) technologies integrated into an information and communication technologies (ICT) framework based on distributed artificial intelligence (DAI) supported by a multi-agent system (MAS), as the most value advantage of supply chain management (SCM) in a cooperative intelligent logistics systems. Logistical platforms (production or distribution) as nodes of added value of supplying and distribution networks are proposed as critical points of the visibility of the inventory, where these technological needs are more evident.