Adaptive Time-of-Use Pricing for Heterogeneous Electric Vehicle Charging: A Behaviour-Aware Monte Carlo and Multi-Objective Optimization Framework

Abstract

The growing adoption of electric vehicles (EVs) offers valuable demand flexibility, yet it poses significant operational challenges for distribution grids. Uncoordinated charging can exacerbate evening peaks, while traditional time-of-use (TOU) tariffs often inadvertently create secondary peaks by clustering charging activity during cheap, off-peak windows. To address this, this study proposes a user-behaviour-aware EV charging framework driven by adaptive peak-valley tariffs. Unlike conventional methods, our approach accounts for diverse EV characteristics—including arrival and departure schedules, initial and required states of charge, and individual price sensitivities. Because external charging data is not assumed, we employ Monte Carlo simulations to generate a synthetic EV population alongside a price-response model that captures varying consumer behaviours. We frame this charging guidance as a multi-objective optimization problem: minimizing both overall user costs and grid fluctuations while strictly meeting individual charging needs. By integrating an adaptive tariff mechanism, the model actively prevents harmful charging synchronization during off-peak hours. Implemented in Python and validated against multiple benchmark scenarios, this reproducible methodology successfully quantifies improvements in grid stability, cost efficiency, and user satisfaction.