商品簡介
Development of Multi-Agent Systems Infrastructure: A Practical Approach discusses modular infrastructure to underpin the deployment of real-life software applications, while implementing well-defined theoretical models. The book addresses the challenges associated with deploying multi-agent systems (MAS) and offers a practical approach to how solutions can be developed. The book is based on the author's experience with building the FLASH-MAS framework (A Fast Lightweight Agent Shell) and other related MAS projects. Many applications are built using the paradigm of autonomous agents in very different fields, including manufacturing, robotics, health care, supply chain management, and more. Agents may be deployed individually, may be distributed over the network, or may be used in great numbers to simulate complex interactions, but there are always some invariants in the functionality of a multi-agent system. A multi-agent system contains other entities beside the agents: the environment, support infrastructures, and units continuously processing data; these have their own lifecycle and needs for interaction. Deployment, maintenance, and monitoring of multi-agent systems brings about many practical challenges, such as dealing with the interaction between entities, the construction, deployment, and eventual destruction of individual entities, the management of an open system, and the robustness of the system as a whole. Some of these challenges are common to many or all applications, hence this book comes to the assistance of multi-agent system developers by offering tools and frameworks for the development and deployment of such systems. Development of Multi-Agent Systems Infrastructure is divided into three main parts, dealing with: models and languages for multi-agent systems; the deployment and interaction between entities; and practical aspects of implementing modern MAS frameworks. FLASH-MAS is a framework built on modularity, so the book presents the solutions the author has identified in a modular manner, meaning that they can be replicated and used independently from one another.
