Intelligent decision support multi agent system in irrigation water management

Abstract

Hydrology and water resources research have some interested in to look for new methodologies that help to address current and future water conflicts over freshwater. The aim ofthis thesis is applied multi-agent approaches in the water resource research as a valuable tool. Multi-agent systems that are frequently applied in various academic disciplines represent the system based on more or less autonomous and cognitive agents. In the first part of the thesis, the method is critically reviewed. Applications conducted in hydrology, water sciences and related areas are considered for this purpose. In addition, existing software systems for multi-agent modeling are discussed. Since the representation of the environment has proved itself one of the most important points for applications in hydrology and water resources research, recent developments in this field ofresearch are taken into account. In the second part of this thesis, a prototype of a multi-agent model of the water supply in the area of Mahaweli H System was developed using the Java Agent Development Framework (JADE) and consist of diverse methods to support in Irrigation Water Management. The model aims at exploring the way the irrigation population handles this situation and manages to satisfy its cultivation demand ofthe water. Thereby, the representation ofthe natural water resources is based on empirical data and hydrological assumptions. In here highly concern about the water control technique used in irrigation system named as the Bulk Water Allocation system. The system was evaluated using the existing data from the Irrigation Department. It has resulted that the water wastage by the system was 72.04% and the natural phenomena was 74.00%. This result concluded that the automated software component of water management functions using multi agent technology as an effective solution for eliminating limitations in decentralized communication.