AI-Driven Dynamic Multimodal Transport Demand Forecasting And Optimization In Pune

2 Dec
Posted on Author

Authors: Fatima Mohsin Inamdar, Pratyay Patel, Prathmesh Tiwari, Pranav Singh Langeh, Prathamesh Thotwe, Pratik Pawar, Pratitya Wankhade

Abstract: Urban mobility in cities like Pune is increasingly challenged by congestion, fluctuating travel demand, and inefficient multimodal transport systems. This paper proposes an AI-driven system for dynamic multimodal transport demand forecasting and optimization, utilizing advanced machine learning models and web-based technologies. We develop a predictive platform that integrates traffic data to estimate demand, determine optimal routes, forecast arrival times, and suggest accurate fares across various modes of transport. Using machine learning frameworks such as PyTorch, XGBoost, and Random Forests, alongside a Flask backend and a HTML, CSS, JavaScript, ReactJS and Bootstrap frontend, the system offers real-time insights to both commuters and transport operators. Our approach aims to alleviate urban transport issues, improve commuter experience, and contribute to smart city initiatives.

DOI: http://doi.org/10.5281/zenodo.17785838