Benchmarking Multi-Agent-Based Decision Support Systems with Current Technologies

Abstract

This study benchmarks Multi-Agent-Based Decision Support Systems (MABDSS) against existing weather forecast systems. The study compares classical, deep learning, and fuzzy algorithm-based decision support systems using a meteorological dataset. Weather prediction is complicated and involves several data sources and analytical methods. Using multi-agent systems to coordinate and cooperate autonomous agents to increase forecasting accuracy and decision-making efficiency seems promising. MABDSS is compared to deep learning models and fuzzy algorithm-based weather prediction systems in this article. Our study established an MABDSS framework that uses intelligent agents to collaborate, evaluate meteorological data, anticipate weather conditions, and assist decision-making. This architecture was compared to deep learning models like CNNs and LSTM networks, which have excelled in time-series forecasting. We also tested fuzzy algorithm-based systems, which can manage data uncertainty and imprecision. Performance criteria including prediction accuracy, computing efficiency, and resilience are compared. Though deep learning models excel in prediction accuracy, MABDSS offers a balanced performance with considerable benefits in flexibility and real-time decision-making. However, fuzzy algorithms handle uncertain data well, making them beneficial in high-variability situations. This study shows the pros and cons of each strategy, revealing their weather forecast potential. The results show that combining multi-agent systems, deep learning, and fuzzy algorithms might yield a hybrid solution that highlights each technology's capabilities. This study provides a benchmarking framework and illuminates the potential of advanced decision support systems in weather prediction.