Web-Based Intelligent Decision Support System for Real-Time Toll Plaza Management and AI-Driven Operational Optimization

Toll plaza operators urgently need clear situational awareness and actionable recommendations to address persistent congestion challenges that increase operational costs and environmental impact. This paper proposes a web-based decision support system that provides real-time analysis of toll plaza conditions, including average queue lengths, operational costs, and operational efficiency, with an intuitive traffic-light status indicator (red-yellow-green) for immediate situational understanding. The system delivers AI-driven recommendations for optimal lane configurations and adaptive barrier control strategies, enabling operators to understand "what's happening" and "what to do" in real-time decision-making scenarios. The web-based platform serves as an intelligent decision-support tool for toll plaza managers, providing the foundation for future automation, centralized monitoring, and proactive alerting capabilities. The system is built upon a hybrid optimization framework comprising two core components: (i) empirical modeling using real transaction data from Thailand’s Expressway Authority (EXAT), leveraging GRU neural networks with Optuna hyperparameter tuning to predict average queue lengths, and (ii) simulation-based scenario generation, allowing operators to explore configuration outcomes under varied, synthetic traffic conditions. Performance improvements such as a reduction of up to 98.6% in the average queue length, a 35.4% decrease in operational costs, and a 1,247 kg CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> daily reduction. This work contributes a deployment-ready intelligent system that transforms toll plaza management from reactive to proactive operations, representing a significant advancement in smart transportation infrastructure and digital mobility solutions.