Modelling PM10 Crisis Peaks Using Multi-Agent based Simulation: Application to Annaba City, North-East Algeria

TL;DR

This paper presents a multi-agent system simulation integrating dispersion models and neural networks to evaluate and compare strategies for controlling PM10 pollution peaks in Annaba, Algeria.

Contribution

It introduces a novel multi-agent simulation framework combining dispersion modeling and neural networks for air pollution control policy assessment.

Findings

MAS effectively simulates different control strategies

The approach helps identify the most efficient pollution reduction policies

Simulation results support policy decision-making in urban air quality management

Abstract

The paper describes a MAS (multi-agent system) simulation approach for controlling PM10 (Particulate Matter) crisis peaks. A dispersion model is used with an Artificial Neural Network (ANN) to predict the PM10 concentration level. The dispersion and ANN models are integrated into a MAS system. PM10 source controllers are modelled as software agents. The MAS is composed of agents that cooperate with each other for reducing their emissions and control the air pollution peaks. Different control strategies are simulated and compared using data from Annaba (North-East Algeria). The simulator helps to compare and assess the efficiency of policies to control peaks in PM10.