MobiWave: Dispatch-Oriented Graph Wavelets and Drift-Guided Selective Optimization for Autonomous Fleet Rebalancing

MobiWave is a new framework for autonomous fleet rebalancing that combines graph wavelets with drift-aware selective optimization.

Its multi-scale graph wavelet module helps separate useful spatial signals from overlapping traffic patterns to improve dispatch representation.

Drift-guided layer-selective optimization addresses performance decay under mobility drift by updating only the model layers most affected.

On real-world and simulated datasets, MobiWave outperformed prior methods in rebalance decisions and adaptation under changing traffic conditions.