Dynamic Passenger Routing in Autonomous Ride-Sharing Systems
Abstract
The study proposes an advanced system for routing passengers in ride-sharing systems that work automatically, with the main goal of improving both the efficiency and the experience of people living in cities with fast changes. We do this by bringing together data about various aspects such as ride requests, traffic, vehicle state, and environmental information and using it to guide real-time and situation-aware routing. Thanks to the coordination of agents and edge computing, autonomous vehicles figure out routes and select passengers without having to depend too much on centralized control. An important part of the design is the ability to group passengers according to what is convenient for them and the vehicle, which saves time and makes the best use of each car in the fleet. Besides, intelligent routing policies are improved with reinforcement learning, which reviews traffic flows and the opinions of people using bus services. The framework makes it possible for high-priority rides to be served without reducing how efficiently the fleet operates. Many studies using a simulator show that using the new routing approach results in quicker passenger waits and shorter journeys, higher fleet use, and lower operational costs. The study plays a part in making ride-sharing autonomous by handling problems in real-time routing, teamwork among vehicles, and personalized support. The suggested idea forms the basis for networks that can provide support and convenience for people, saving energy and effectively handling urban traffic.