Computing Supply Function Equilibria via Spline Approximations

TL;DR

This paper introduces a spline-based numerical method for efficiently computing supply function equilibria in markets, especially electricity markets, improving accuracy and reducing computational effort.

Contribution

It develops a novel spline approximation approach for solving the differential equations defining supply function equilibria, enhancing existing algorithms' efficiency and usability.

Findings

The spline-based algorithm converges asymptotically to true equilibria.

The method significantly reduces computational time compared to previous approaches.

Numerical examples demonstrate improved accuracy and ease of use.

Abstract

The supply function equilibrium (SFE) is a model for competition in markets where each firm offers a schedule of prices and quantities to face demand uncertainty, and has been successfully applied to wholesale electricity markets. However, characterizing the SFE is difficult, both analytically and numerically. In this paper, we first present a specialized algorithm for capacity constrained asymmetric duopoly markets with affine costs. We show that solving the first order conditions (a system of differential equations) using spline approximations is equivalent to solving a least squares problem, which makes the algorithm highly efficient. We also propose using splines as a way to improve a recently introduced general algorithm, so that the equilibrium can be found more easily and faster with less user intervention. We show asymptotic convergence of the approximations to the true equilibria for both algorithms, and illustrate their performance with numerical examples.