A Review of Fleet Rebalancing and Energy Management in Shared Electric Mobility

Abstract

The continuous expansion of shared electric mobility systems has accentuated systemic tensions between spatio-temporal fleet rebalancing and battery degradation constraints across modern urban transit networks. This critical review examines contemporary methodologies bridging multi-agent deep reinforcement learning dispatching, electrochemical remaining useful life forecasting, and distributed cyber-physical infrastructures. Current investigations demonstrate that decentralized algorithmic rebalancing enhances vehicular utilization and mitigates urban supply-demand imbalances to some extent; however, persistent charging infrastructure bottlenecks and non-stationary commuter dynamics reveal lingering operational trade-offs. Furthermore, reported dispatching efficiencies and degradation attenuation may partially stem from stylized synthetic mobility simulations or selective agent boundary definitions rather than authentic metropolitan equilibrium. Considering these intersecting behavioral patterns, electrochemical wear kinetics, and platform governance requirements, this leads us to further thinking that fleet optimization cannot be formulated as a purely kinematic repositioning problem. Rather, intelligent shared mobility represents a complex socio-technical negotiation among real-time spatial accessibility, electrochemical battery longevity, and regional energy grid capacities. Further research is needed to formulate adaptive, multi-scale dispatching architectures that reconcile dynamic urban transit demands with rigorous component degradation mechanics under non-stationary environmental and electrical stresses.