Basic steps for the development of decision support systems

Abstract

There is a growing desire to develop effective and efficient computational methods and tools that facilitate environmental analysis, evaluation and problem solving. Environmental problems of interest may include concerns as apparently dissimilar as revitalization of contaminated land, and effective management of inland and coastal waters. The approach to effective problem solving in both of these examples can involve the development of what are commonly called Decision Support Systems (DSSs). Standard DSSs might be characterized as computational systems that provide access to a wealth of information pertaining to a specific problem. The types of information that might be available include information content, maps, and data. This information can be contained in databases and geographic information systems (GIS). Access is often provided through interfaces to queries that ease the task of sifting through the often large amounts of information available. These DSSs facilitate some numerical analysis (e.g., overlays of data on GIS images, rudimentary statistical analysis of data), but usually only indirectly affect evaluation and problem solving. Currently, DSSs of this form are the most common. However, an option exists to incorporate evaluation and problem solving directly into a DSS by using statistical decision tools such as sensitivity analysis and multi-criteria decision analysis. These systems may be thought of as decision analysis (MCDA) support systems. Development of a DSS requires consideration of both the problem to be solved and the computational tools that are appropriate or needed. In terms of the problem, important components include: definition of objectives; links to the legislative or regulatory context; model structuring including identification of, and relationships between, parameters; cost factors; and value judgments. These should encompass environmental, economic and socio-political concerns. This is the standard approach to performing decision analysis using MCDA tailored specifically to environmental problem solving. A further consideration is how to gather and present case studies, once the DSS is developed. Computational issues that are faced include: database management (e.g., information, data, GIS); analysis tools (statistics, fate and transport modeling, risk assessment, MCDA); visualization of the problem; presentation of results; document production; feedback mechanisms; help; and advice. The user interface to each of these components, the navigation through these components, and degree of openness of each component of the DSS must also be considered. Openness, including communication and stakeholder involvement, is very important for maintaining transparency and defensibility in all aspects of the DSS. © 2009 Springer Science+Business Media, LLC.