Providing Automated Semantic Support for Software Agents in Spatial Decision Support Systems
View/ Open
Research Paper 1(Sen LI).pdf (505.7Kb)
Research Paper 2(Sen LI).pdf (725.0Kb)
Appendix.pdf (725.5Kb)
Supporting Document.pdf (2.104Mb)
Date
05/12/2008Item status
Restricted AccessAuthor
Li, Sen
Metadata
Abstract
Spatial Decision Support System(SDSS) is a software system that aimed at assisting decision-makers generate and evaluate alternative solutions to semi and unstructured spatial problems through the integration of spatial data and geo-processing models. There are three contemporary issues that significantly determine the functionality of SDSS, namely, (i) availability and interoperability of the required data and processes that are increasingly searchable through internet, (ii) capability of the interface to hide the technical complexity for inexperienced users, and (iii) software system’s adaptability to a dynamic decision-making environment. In order to ameliorate these limitations, a Multi-Agent based SEmantic driven (MASE) approach has been introduced. Semantic Web Service technology is used for MASE to enrich the descriptions of web based data and processes, and cooperative “software agents” are adopted to construct the software system in a way that modelling the behaviours of a GIS expert. In this paper, we detailed the theoretical background and architecture of MASE. In order to demonstrate the utility of MASE, we have presented the practical work of implementing a prototype MASE system in this paper. A scenario on epidemiology management has been set up firstly with a focus on seeking spatio-temporal characteristics of a user specified epidemic. To construct the application system, we have utilised semantic web technology to publish profiles of a set of imaginary data and processes services into a web accessible registry, and adopted agent-based modelling method to build the software system. The evaluated outcome highlighted the system with a sound autonomy and flexibility to (i)implement accurate service discovering (ii)generate alternative solutions for the use case and (iii)export the result in an easy-understanding way.