A dynamic model for ride-matching problem in multi-hop ride-sharing system
summary
A dynamic matching model represents vehicles and passengers as intersecting time-expanded ellipsoids and computes an analytic midpoint and short axis to locate feasible matches, controlled by a detour ratio k (1.1–2.0). Implementation uses SUMO + TraCI and one hour of Haikou taxi data for calibration and a doubled-demand stress test. Results show higher transfer acceptance raises share ratio, cuts waiting delay and total vehicle kilometers for short trips, while longer trips remain hard to consolidate and smaller k shrinks search area for faster but weaker matches.
AI-written summary · about