Distributionally robust location-inventory planning of commercial electric vehicle battery swapping stations with joint chance constraints

The electrification of commercial electric vehicle fleets is constrained by long charging times and insufficient infrastructure. Battery swapping stations (BSSs) offer a promising alternative, yet their strategic deployment is complicated by uncertain spatial demand and stringent service requirements. In this paper, the commercial electric vehicle BSS location-inventory problem is investigated through a two-stage distributionally robust optimization (DRO) framework with a joint chance constraint. Using moment-based ambiguity sets, we derive a tractable second-order cone programming reformulation via a conditional value-at-risk approximation and solve it with a tailored outer approximation algorithm. Numerical experiments on standard transportation networks, together with benchmark tests against commercial solver DICOPT, confirm the computational tractability and scalability of the proposed approach. Sensitivity analyses further reveal how ambiguity, transportation cost, inventory capacity, construction cost, and reliability requirements shape battery inventory deployment and network configuration. Finally, out-of-sample comparisons with stochastic programming and robust optimization benchmarks show that the DRO framework provides a more balanced trade-off between cost and reliability, maintaining strong network-wide service performance under multiple demand distributions.