A Multi-Agent System for Optimal Supply Chain Management

Supply chain management recently has been developing into a dynamic environment that has to accept the changes in the formation of the supply chain. In other words, the supply chain is not static but varies dynamically according to the environmental changes. Therefore, under this dynamic supply chain environment, the priority is given not to the management of the existing supply chain but to the selection of new suppliers and outsourcing companies in order to organize an optimal supply chain. The objective of this research is to develop a multi-agent system that enables the effective formation and management of an optimal supply chain. The multi agent system for optimal supply chain management developed in this research is a multi agent system based on the scheduling algorithm, a cooperative scheduling methodology, which enables the formation of an optimal supply chain and its management. By means of active communications among internal agents, a multi-agent system for optimal supply chain management makes it possible to quickly respond to the production environment changes such as the machine failure or outage of outsourcing companies and the delivery delay of suppliers. This research has tried to suggest a new direction and new approach to the optimal supply chain management by means of a multi-agent system in dynamic supply chain environment

Task Allocation in Case-Based Recommender Systems

Case-based recommender systems can learn about user preferences over time and automatically suggest products that fit these preferences. In this chapter, we present such a system, called CASIS. In CASIS, we combined the use of swarm intelligence in the task allocation among cooperative agents applied to a case-based recommender system to help the user to plan a trip.

Component Agent Systems

One central problem preventing widespread adoption of mobile agents as a code structuring primitive is that current mainstream middleware implementations do not convey it simply as such. In fact, they force all the development to be centered on mobile agents, which has serious consequences in terms of software structuring and, in fact, technology adoption. This chapter discusses the main limitations of the traditional platform-based approach, proposing an alternative: component-based mobile agent systems. Two case studies are discussed: the JAMES platform, a traditional mobile agent platform specially tailored for network management, and M&M, a component-based system for agent-enabling applications. Finally, a bird’s eye perspective on the last 15 years of mobile agent systems research is presented along with an outlook on the future of the technology. The authors hope that this chapter brings some enlightenment on the pearls and pitfalls surrounding this interesting technology and ways for avoiding them in the future.

Three Perspectives on Multi-Agent Reinforcement Learning

This chapter concludes three perspectives on multi-agent reinforcement learning (MARL): (1) cooperative MARL, which performs mutual interaction between cooperative agents; (2) equilibrium-based MARL, which focuses on equilibrium solutions among gaming agents; and (3) best-response MARL, which suggests a no-regret policy against other competitive agents. Then the authors present a general framework of MARL, which combines all the three perspectives in order to assist readers in understanding the intricate relationships between different perspectives. Furthermore, a negotiation-based MARL algorithm based on meta-equilibrium is presented, which can interact with cooperative agents, games with gaming agents, and provides the best response to other competitive agents.

Modeling Knowledge and Reasoning in Conversational Recommendation Agents

This chapter introduces conversational recommender agents that facilitate user navigation through a product space, alternatively making concrete product suggestions and eliciting the user’s feedback. Critiquing is a common form of user feedback, where users provide limited feedback at the feature-level by constraining a feature’s value-space. For example, a user may request a cheaper product, thus critiquing the price feature. One of the most important objectives in a recommender agent is to discover, with minimal user feedback, which are the user’s product preferences. For this purpose, the chapter includes recent research on critiquing-based recommendation and a comparison between standard and recent proposals of recommendation based on critiquing.

RiskMan

The purpose of this chapter is to discuss the use of multi-agent systems to develop virtual reality training systems. We first review these systems and then investigate the architectures used. We demonstrate an example of our own (RiskMan) and then discuss the advantages and drawbacks of using multi-agent agent approaches in the development of virtual reality training systems. The chapter describes the system architecture of a multi-agent system for risk management (RiskMan) to help train police officers to handle high-risk situations. RiskMan has been developed using a high-level scripting language of a game engine, Unreal Tournament 2004. The major modules are a scenario-based expert system, a narrative engine, a game engine, and a graphics engine. The system integrates a simulation agent, trainee agent, communication agent, interface agent, and scripted agents communicating using games technology.

Concurrent Programming with Multi-Agents and the Chemical Abstract Machine

In this chapter, we propose a new concurrent programming approach called MACH (multi-agent extended chemical abstract machine). MACH extends the chemical abstract machine with multiple coexisting agents. This paper focuses on the design, implementation, and verification of MACH. The aim of MACH is to develop a reactive programming language based on an interactive computational model, which we believe is the key to concurrent programming. We present MACH as a simple and efficient programming approach based on a sound theoretical background.

Robust Intelligent Control of Mobile Robots

An appropriate control architecture is a crucial premise for successfully achieving truly autonomous mobile robots. The architecture should allow for a robust control of the robot in complex tasks, while it should be flexible in order to operate in different environments pursuing different tasks. This chapter presents a control framework that is able to control an autonomous robot in complex real-world tasks. The key features of the framework are a hybrid control paradigm that incorporates reactive, planning and reasoning capabilities, a flexible software architecture that enables easy adaptation to new tasks and a robust task execution that makes reaction to unforeseen changes in the task and environment possible. Finally, the framework allows for detection of internal failures in the robot and includes self-healing properties. The framework was successfully deployed in the domain of robotic soccer and service robots. The chapter presents the requirements for such a framework, how the framework tackles the problems arising from the application domains, and results obtained during the deployment of the framework.

A Methodology for Modeling Expert Knowledge for Development of Agent-Based Systems

For intelligent agents to become truly useful in real-world applications, it is necessary to identify, document, and integrate into them the human knowledge used to solve real-world problems. This article describes a methodology for modeling expert problem-solving knowledge that supports ontology import and development, teaching-based agent development, and agent-based problem solving. It provides practical guidance to subject matter experts on expressing how they solve problems using the task reduction paradigm. It identifies the concepts and features to be represented in an ontology; identifies tasks to be represented in a knowledge base; guides rule learning/refinement; supports natural language generation; and is easy to use. The methodology is applicable to a wide variety of domains and has been successfully used in the military domain. This research is part of a larger effort to develop an advanced approach to expert knowledge acquisition based on apprenticeship multi-strategy learning in a mixed-initiative framework.

A Gentle Introduction to Fuzzy Logic and Its Applications to Intelligent Agents Design

The purpose of this chapter is to present the key properties of fuzzy logic and adaptive nets and demonstrate how to use these, separately and in combination, to design intelligent systems. The first section introduces the concept of fuzzy sets and their basic operations. The t and s norms are used to define a variety of possible intersections and unions. The next section shows two ways to estimate membership functions, polling experts, and using data to optimize parameters. Section three shows how any function can be extended to arguments that are fuzzy sets. Section four introduces fuzzy relations, fuzzy reasoning, and shows the first steps to be taken to design an intelligent system. The Mamdami model is defined in this section. Reinforcement-driven agents are discussed in section five. Sections six and seven establish the basic properties of adaptive nets and use these to define the Sugeno model. Finally, the last section discusses neuro-fuzzy systems in general. The solution to the inverted pendulum problem is given by use of fuzzy systems and also by the use of adaptive nets. The ANFIS and CANFIS architectures are defined.