Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.487221
Title: Data distribution strategies for the distributed simulation of multi-agent systems
Author: Oguara, Tonworio.
ISNI:       0000 0001 3455 4683
Awarding Body: University of Birmingham
Current Institution: University of Birmingham
Date of Award: 2007
Availability of Full Text:
Access through EThOS:
Abstract:
Multi-agent systems are increasingly being employed in the development of concurrent systems in a wide range of areas such as telFommunications, mobile robot control, military simulation, computer games and business process modelling. These systems are however quit complex and it can be difficult to formally verify their properties. As a result, the design and implementation of these systems has been largely experimental. To this end, simulation has a fundamental role to play in the development of multi-agent systems, allowing the agent engineer to learn more about the system and its behaviour or investigate the significance of an optional agent architecture or technology. Thus, with simulation strategies, the engineer can test ideas and theories which might either be too expensive, too complex or even too dangerous to attempt in the real physical system and also focus on particular aspects ofthe system. Typically, agent-based systems are large, extremely complex and complicated, therefore the simulation of a realistic large scale model exceed the capabilities of a conventional sequential computing system. Distributed simulation has thus emerged as an appropriate technology to address this problem. This thesis is concerned with a framework for the distributed simulation of multi-agent systems. The framework namely PDES-MAS, exploits the notion of the agents' sphere of influence to achieve the dynamic partitioning and distribution of the agent model. The thesis presents the design and implementation of the PDES-MAS kernel and proposes a series of dynamic data distribution algorithms in this context. Also presented is a detailed quantitative analysis of the algorithms.
Supervisor: Not available Sponsor: Not available
Qualification Name: University of Birmingham, 2007 Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.487221  DOI: Not available
Share: