Multi-Agent Systems for Distributed Geospatial Modeling, Simulation and Computing

Multi-Agent Systems for Distributed Geospatial Modeling, Simulation and Computing

Genong Yu (George Mason University, USA), Liping Di (George Mason University, USA), Wenli Yang (George Mason University, USA), Peisheng Zhao (George Mason University, USA) and Peng Yue (George Mason University, USA)
Copyright: © 2009 |Pages: 9
DOI: 10.4018/978-1-59140-995-3.ch025
OnDemand PDF Download:
$37.50

Abstract

Multi-agent system is specialized in studying the collective effects of multiple intelligent agents. An intelligent agent is a computer system with autonomous action in an environment. This technology is especially suitable for studying geospatial phenomena since they are complex in nature and call for intertwined actions from different forces. This chapter describes multi-agent systems and their application in geospatial modeling, simulation and computing. Geospatial data integration and mining are discussed.
Chapter Preview
Top

Introduction

This section reviews applications of multi-agent systems in geospatial data integration, modeling, simulation, and computing. Agent is a computer system with autonomous action in some environment (Weiss, 2000). This is by no means a comprehensive or concise definition for an agent. As pointed out by many scientists, no consensus has been reached on the definition of an agent (Franklin & Graesser, 1997; Russell & Norvig, 2002; Weiss, 2000). Readers who are interested in the different interpretations and typologies of agents in different contexts may refer to Weiss (2000) and Nwana (Nwana, 1996) for an in-depth discussion. In this context, the definition focuses on the most commonly agreed-upon property of agents - namely, their autonomous ability. This is the backbone in forming multi-agent systems. Multi-agent systems are formed by many agents that interact with each other. These systems emphasize the interaction between agents and the emergent effects from relatively simple individual behaviors of agents. One of the key promises of multi-agent systems is its capability to decompose complex geospatial problems into manageable pieces. Agents communicate and interact with each other through an understanding of common ontology (or domain-specific vocabulary) and communication languages. Agents can be managed and discovered through a centralized directory, peer-to-peer discovery, or hybrid mechanism. Agent mobility provides a mechanism to extend stabilities and sustainability of geospatial services in a heterogeneous distributed environment. Table 1 summarizes some of the most popular properties of agents that are important for interactions in a distributed multi-agent system. Table 2 lists some popular platforms for developing multi-agent systems.

Table 1.
Properties of an agent(Franklin & Graesser, 1997; Russell & Norvig, 2002; Weiss, 2000)
    Property    Description
    Reactive    reaction based on its sense
    Autonomous    responses based on its own experiences
    Rational    maximize its own interest
    goal-oriented    pursue an goal
    temporally continuous    deals with continuous process
    Mobile    able to transport itself from platform to platform
    Communication    interactions on another level of abstraction - language

Key Terms in this Chapter

Intelligent Agent: An intelligent agent is a computer system with autonomous action in some environment.

FIPA: Foundation for Intelligent Physical Agents – an IEEE Computer Society standards organization that promote agent-based technology and interoperability of its standards with other technologies, including web services.

JADE: Java Agent Development Framework. One of the most popular open source agent frameworks implemented in the Java language.

Multi-Agent: Many agents interact with each other while each agent has incomplete information and is restricted in its capabilities.

Ontology: Simply, it can be treated as vocabulary that specifies not only definition, but also relationship among “words”.

Agent-Based Modeling: In artificial life, it is the set of techniques in which relations and descriptions of global variables are represented with an explicit representation of microscopic features of the system, typically in the form of agents that interact with each other and their environment according to rules in a discrete space-time.

Multi-Agent Simulation: The simulation of a multi-agent system where agents are located in an environment. In such an environment, agents sense local neighborhood, hears messages, and send messages to each other.

Utility: It represents the “happiness” of an agent, or the quality of being useful.

Complete Chapter List

Search this Book:
Reset