Ioannis Athanasiadis | Γιάννης Αθανασιάδης
Information Technologies in Environmental Engineering
Filed Under (Book chapters) by I N Athanasiadis on 07-02-2009
In Advanced Agent-Based Environmental Management Systems, (U. Cortés and M. Poch, ed.), pp. 119-138, 2009.
Abstract This chapter presents a unifying methodology for developing environmental information systems with software agents. Based on the experience reported in recent literature, we abstract common requirements of environmental information systems into agent types, combine state-of-the-art tools from computer science, service-oriented software engineering and artificial intelligence domains, as software agents and machine learning, and illustrate their potential for solving real-world problems. Specifically, two generic agent types are specified that behave as information carriers and decision makers, which provide an appropriate abstraction for deployment with added-value services in environmental management information systems. A concrete pathway for applying these instruments throughout the software lifecycle of an environmental management information system is outlined, along with suggestions for software specification and deployment. The method is demonstrated in two application domains: one for air quality assessment and another for meteorological radar data surveillance.
Read also:
Filed Under (Book chapters) by I N Athanasiadis on 02-02-2007
In Computational Intelligence Based on Lattice Theory, (V. G. Kaburlasos and G. X. Ritter, ed.), pp. 175-193, 2007.
Abstract This chapter introduces a rule-based perspective on the framework of fuzzy lattices, and the Fuzzy Lattice Reasoning (FLR) classifier. The notion of fuzzy lattice rules is introduced, and a training algorithm for inducing a fuzzy lattice rule engine from data is specified. The role of positive valuation functions for specifying fuzzy lattices is underlined and non-linear (sigmoid) positive valuation functions are proposed, that is an additional novelty of the chapter. The capacities for learning of the FLR classifier using both linear and sigmoid functions are demonstrated in a real-world application domain, that of air quality assessment. To tackle common problems related to ambient air quality, a machine learning approach is demonstrated in two applications. The first one is for the prediction of the daily vegetation index, using a dataset from Athens, Greece. The second concerns with the estimation of quartely ozone concentration levels, using a dataset from Valencia, Spain.
Read also:
Filed Under (Book chapters) by I N Athanasiadis on 10-01-2007
In Agent and Web Service Technologies in Virtual Enterprises, (N. Protogeros, ed.), pp. 256-266, Hershey, PA, USA, 2007.
Abstract This chapter introduces a virtual enterprise architecture for environmental information management, integration and dissemination. On a daily basis, our knowledge related to ecological phenomena, the degradation of the natural environment and the sustainability of human activity impact, is growing as a consequence raises the need for effective environmental knowledge exchange and reuse. In this work, a solution among collaborating peers forming a virtual enterprise is investigated. Following an analysis of the main stakeholders, a service-oriented architecture is proposed. Technical implementation options, using Web services or software agents, are considered and issues related to environmental information management, ownership and standardization are discussed.
Read also:
Filed Under (Book chapters) by I N Athanasiadis on 18-10-2006
In Environmental Modelling, Software and Decision Support, (A. Jakeman, A. Voinov, A. E. Rizzoli and S. Chen, ed.), pp. 101-118, 2008.
Abstract In this chapter we investigate the motivation behind the development of modelling frameworks that explicitly target the environmental domain. Despite many commercial and industrial-strength frameworks are available, we claim that there is a definite niche for environmental-specific frameworks. We first introduce a general definition of what is an environmental integrated modelling framework, leading to the outlining of the requirements for a generic software architecture for such frameworks, which identifies the need for a knowledge layer, to support the modelling layer and the experimentation layer, to support the execution of models. The chapter then focuses on the themes of knowledge representation, model management and model execution. We advocate that appropriate knowledge representation and management tools can facilitate model integration and linking. We stress that a model development process adhering to industry standards and good practices, called “model engineering” is to be pursued. Finally, we focus on the requirements of the experimental frame, which can ensure the transparency and traceability of the execution of simulation scenarios and optimisation problems in complex integrated assessment studies.
Read also:
Filed Under (Book chapters) by I N Athanasiadis on 18-10-2006
In Environmental Modelling, Software and Decision Support: State of the art and new perspective, (A. Jakeman, A. Voinov, A. E. Rizzoli and S. Chen, ed.), pp. 205-228, 2008.
Abstract Over recent years a huge library of data mining algorithms has been developed to tackle a variety of problems in fields such as medical imaging and network traffic analysis. Many of these techniques are far more flexible than more classical modeling approaches and could be usefully applied to data-rich environmental problems. Certain techniques such as Artificial Neural Networks, Clustering, Case-Based Reasoning and more recently Bayesian Decision Networks have found application in environmental modeling while other methods, for example classification and association rule extraction, have not yet been taken up on any wide scale. We propose that these and other data mining techniques could be usefully applied to difficult problems in the field. The paper is a general introduction to Data Mining techniques for Environmental Scientists who may be interested in using them in their applications. So, special work is done on the contributions of DM techniques to environmental applications and on general guidelines of good practice in real world domains. Technical details on the specific DM techniques are not the focus of this paper, but to provide general ideas to a non expert user that reading them can decide which is the proper technique useful to solve his problem and specific references are provided
Read also:
Filed Under (Book chapters) by I N Athanasiadis on 01-06-2006
In More Puzzle Solving for Policy, (P. Valkering, B. Amelung, R. van der Brugge and J. Rotmans, ed.), pp. 208-217, Maastricht, The Netherlands, 2006.
Abstract In this paper we address the steps that precede the creation of a mathematical representation of real world phenomena. More specifically, we investigate first the formulation of the problem statement in an integrated assessment modelling study, seeking here to offer a more precise definition of problem formulation and examining its influence on later stages of modelling. We identify, through both numerous group discussions at the summer school and a survey of the Integrated Assessment literature, that although there is much discourse around the problem’s formulation, it has not been well defined. Subsequently, a more precise definition may aid in helping us to understand the steps researchers might take in formulating and solving a problem. Furthermore, problem formulation will be improved through additional classification and detailed analysis particularly with respect to stakeholder involvement and participatory approaches. These steps may be of value to not only emerging researchers and professionals such as has attended this current and future Summer Schools, but also giving credence to the observed lack of attention to problem formulation in the literature.
Read also:
Internet for peace
Environmental Modelling & Software
