Development of a multi-agent-based simulation system for air quality analysis

Abstract

The implementation of distributed collaborative intelligence can improve the efficiency of agent-based environmental systems (for environmental networks) in the context of solving emergency situations in real time, such as the critical situation of severe air pollution with an important impact in a given region. In this paper it is presented a distributed multi-agent system for air quality monitoring and analysis (AQDMAS) whose first version was implemented in Zeus, a Java-based toolkit for intelligent agents' development. The study is focused on the possibilities of the computer network distribution of the software agents. The purpose of the system is to verify the air quality standard fulfilment, and if, in case some exceedances occur, to inform the supervisor in order to take a decision. For this version there were designed seven intelligent agents placed in a local computer network in the Petroleum-Gas University of Ploiesti campus in Romania. The results revealed the effectiveness of the proposed monitoring and analyzing technique and the possibility of autonomous built-in simulation of tracking the air quality evolution with distributed resources.