Coupling Multi-criteria Analysis And Machine Learning For Agent Based Group Decision Support: Spatial Localization

Abstract

The land use management context is known for its spatial complexity. It is a multidimensional problem influenced by several criteria of dissimilar importance. This kind of problem involves many decision-makers (individuals and institutions) with often conflictual preferences. The authors’ contribution consists of designing and developing a web intelligent multi-criteria group decision support system (WIM-GDSS), which combines four tools so that the shortcoming of one tool is complemented by the strength of the others. These tools are Multi-Agent System, Geographic Information System, Multi-Criteria Analysis methods (TOPSIS and AHP) and Machine Learning techniques (Linear Regression). The current study aims to assist decision-makers in choosing the most adequate alternative that best meets certain criteria. The chosen solution has to satisfy the majority of the involved decision-makers. In this perspective, WIM-GDSS will be enriched with a coordination protocol, allowing the agents to properly collaborate to find a compromise solution using multiple criteria analysis methods and prediction models.