A methodology for developing software-agent systems in environmental informatics applications
The development of IT systems for the environment is a exigent process, that presupposes interdisciplinary collaboration and blending of opinions, techniques and methods from several disciplines as the natural and environmental sciences, the social and economic sciences, the computer science and engineering. In the specific domain, software agent technology, a novel paradigm for software engineering,has been used fragmentarily, and even more, there is not any practice guide or integrated framework available for assisting the software engineer to develop environmental informatics applications with software agents. The goal of the present thesis is the foundation of an integrated methodology for applying agent technology in the environmental informatics. The framework developed in the present thesis, considers software agents in environmental informatics in three distinct, yet concurrent ways : a) as information carriers, b) as decision-makers, and c) as societal illustrators. The three generic agent types specified can be combined for developing Environmental Information Systems(EIS). A best practice guide for applying agent technology in environmental informatics is proposed, that utilizes the three generic agent types throughout the entire lifecycle of an Environmental Information System. The presented framework is demonstrated in two case studies that undertake the current challenges of environmental informatics. First application is an integrated platform for environmental measurement fusion and management, and for the diffusion of environmental information. This generic-purpose, agent-based, system is demonstrated in the air-quality assessment domain. The second application is an integrated, multi-agent social simulation platform for simulating the diffusion of environmental conservation signals within a society. This platform is demonstrated for the simulation of the urban water supply consumption chain.
I. N. Athanasiadis, A methodology for developing software-agent systems in environmental informatics applications, PhD Thesis, pg. 228, 2005, Aristotle University of Thessaloniki, doi:10.12681/eadd/14946.
You might also enjoy (View all publications)
- Introducing digital twins to agriculture
- Location-specific vs location-agnostic machine learning metamodels for predicting pasture nitrogen response rate
- Machine learning for large-scale crop yield forecasting