EUROPT 2025
Abstract Submission

374. Bi-Level Optimization for Electricity Pricing: Links with the Rochet-Choné Model

Invited abstract in session TC-7: Mixed-Integer Bilevel Optimization, stream Bilevel and multilevel optimization.

Tuesday, 14:00-16:00
Room: B100/5015

Authors (first author is the speaker)

1. Abdellah Bulaich Mehamdi
CMAP, École Polytechnique
2. Wim van Ackooij
OSIRIS, EDF R&D
3. Luce Brotcorne
INRIA
4. Stephane Gaubert
CMAP, Ecole polytechnique, INRIA
5. Quentin Jacquet
OSIRIS, EDF

Abstract

Pricing strategies in electricity markets are crucial for optimizing supply and demand dynamics, ensuring that both producers and clients make effective decisions. These strategies help to manage grid stability, to encourage efficient energy use, and to promote competition among suppliers, ultimately benefiting clients through more competitive prices and better options.

In this talk, we model the pricing problem using a Principal-Agent framework under adverse selection: the Rochet-Choné model. This model addresses pricing in a monopolistic market under reservation constraints and captures the interaction between the provider and consumers. The goal is to determine tariffs that balance profitability with social welfare. To solve this problem, we discretize the client space and solve the corresponding bi-level problem using the value function approach.

Additionally, we introduce quantization methods to improve the representation of the provider’s offer. Using a menu of pricing contracts derived from the Rochet-Choné model, we show how to identify a subset that best represents the full menu. First we model the problem as a mixed-integer bi-level optimization problem. Next, we describe a greedy heuristic to compute the reduced menu, leveraging tools such as the Newton polygon and Fenchel conjugacy. Numerical experiments on real-life data from EDF highlight the method’s effectiveness.

Keywords

Status: accepted

Back to the list of papers