Publications

Abstract : Weather forecasting is a challenging task carried out by meteorological stations throughout the globe. Traditionally, meteorologists use Numerical Weather Prediction (NWP) Models for weather prediction. This classical approach attempts to model the fluid and thermal dynamic systems for grid-point time series prediction based on boundary meteorological data. NWP solves a large system of nonlinear mathematical equations to provide forecasts. However, accurate weather prediction is challenging because of climate change and global warming. This paper proposes a novel approach for weather forecasting using a Multi Agent System (MAS). The proposed approach incorporates a deep neural network model within MAS with a hybrid Artificial Neural Network (ANN) algorithm to recognize the static and dynamic weather conditions. It uses ensemble prediction to account for indeterminism in weather conditions. Implementation challenges and advantages of MAS compared to NWP approach are discussed in this paper.