Coordinated Scheduling of Distribution Networks and PEDF Microgrids via a Stackelberg Game with Wasserstein-DRO Reserve Requirements

High photovoltaic penetration increases reserve needs in active distribution networks, while PEDF microgrid flexibility is seldom coordinated as a reserve product. This study develops a complete-information Stackelberg framework in which a distribution system operator posts bounded prices for energy exchange, reactive support, and directional reserve. A Wasserstein ambiguity set centered on the empirical source–load error distribution supports a worst-case conditional value-at-risk calculation of hourly requirements. PEDF reserve offers are constrained by available battery energy, flexible-load headroom, recoverable photovoltaic curtailment, shared converter capacity, and activated feeder states. KKT conditions and Big-M linearization yield a mixed-integer second-order cone program. On a modified IEEE 33-bus feeder, joint PEDF and external reserve procurement reduces system cost by 2.455% and aggregate PEDF net cost by 10.402%, with zero reserve shortages and a minimum voltage of 0.951284 p.u. Fixed-ratio, SAA-CVaR, Box-RO, and WDRO-CVaR comparisons reveal an economy–reliability tradeoff. Rolling out-of-sample tests identify 0.1 as the smallest tested radius multiplier meeting a 5% violation target in all three training windows. An IEEE 69-bus extension reduces system cost by 4.233%, and all 144 nonlinear AC power-flow states converge. The results support one-interval steady-state reserve coordination under the tested assumptions.