A hierarchical robust scheduling framework for electric vehicle aggregators in coupled spot and ancillary service markets

This paper proposes a bi-level scheduling framework integrating probability-interval hybrid uncertainty modeling with robust optimization to address the co-existing challenges of electric vehicle (EV) user behavior and electricity market uncertainties. Targeting electric vehicle aggregators (EVAs) in the electricity spot market (ESM) and frequency regulation ancillary service market (FRASM), the framework hierarchically manages multi-market coordination. At the upper level, Analytic Hierarchy Process (AHP) and probability density functions (PDF) characterize stochastic EV behaviors (initial state-of-charge, arrival/departure times, multi-rate charging), with Monte Carlo (MC) sampling and Mini-Batch K-means clustering generating representative scenarios. The lower level develops a two-stage robust optimization model with battery degradation costs, where interval uncertainty sets (IPUs) bound market price/demand fluctuations, and compressed binary scenario matrices inform first stage bidding decisions. To solve this large-scale model efficiently, nonlinear constraints are linearized via Big M and McCormick envelopes, while the Column and Constraint Generation (C&CG) algorithm enables rapid worst-case optimization. Case studies validate the framework's effectiveness in enhancing economic returns and robustness for multi-market EVA operations.