Object-oriented modelling as a Decision-Making tool in agriculturally overexploited Karstic aquifers

Abstract

Object-Oriented Bayesian Networks (OOBNs) utilise the power of Object-Oriented Programming (OOP) and offer a novel approach to the problems of integrated water management. This paper describes the building of an OOBN Decision Support System (DSS) that allows complex domains to be described in terms of inter-related objects. Thus, the DSS structure is able to represent an accurate reflection of a complex real-world water system made for an aquifer that has been used as an example of a successful application. In this research, conventional Bayesian Networks (BNs) are used to describe the probabilistic relationships between variables (objects) within each network. A network is a group of objects that can be described as a class. Different classes can possess similar sets of objects and be linked through other networks having common variables. Classes inherit commonly used states and behavior from other classes in a hierarchicalway. This model of networks represents a participatory DSS for helping water managers.