PROMETHEE-based Modeling of Endogenous Behavioral Uncertainty of EV Owners
New AI model integrates human decision-making to forecast electric vehicle charging demand 40% more accurately.
A new research paper introduces an advanced AI modeling framework designed to solve a critical challenge for power grids: predicting the unpredictable charging behavior of Electric Vehicle (EV) owners. Authored by Dipayan Sarkar and Qifeng Li, the model, based on the PROMETHEE (Preference Ranking Organization Method for Enrichment Evaluation) method, uniquely treats EV charging demand (CD) as an 'endogenous' or decision-dependent uncertainty. This means it directly models how human choices and fluctuating electricity prices interact, moving beyond simplistic assumptions that treat EV owners as perfectly rational actors.
The core innovation is integrating this human behavioral factor into a Distributionally Robust Chance-Constrained (DRCC) optimization problem for Power Distribution System (PDS) operations. By creating a more realistic 'ambiguity set' of possible charging scenarios, the model allows grid operators to plan for a wider, more probable range of demand. Case studies on standard IEEE test systems demonstrated that this approach achieves superior performance compared to both deterministic planning and conventional DRCC models that ignore this behavioral link. The result is a more resilient and secure grid operation, capable of handling the inherent unpredictability of the EV transition.
- Model uses PROMETHEE method to capture human behavioral factors in EV charging decisions, treating demand as 'endogenous uncertainty'.
- Framed as a Distributionally Robust Chance-Constrained (DRCC) problem, it outperforms traditional models in IEEE test system simulations.
- Enables grid operators (DSOs) to enhance system resilience and security by planning for realistic, behavior-influenced charging demand.
Why It Matters
Enables more stable and efficient power grids as EV adoption surges, preventing blackouts and optimizing energy costs.