A Data Driven Structural Decomposition of Dynamic Games via Best Response Maps

TL;DR

This paper presents a novel data-driven approach to simplify the computation of Nash equilibria in dynamic games by embedding best-response maps as feasibility constraints, improving efficiency and solution quality.

Contribution

It introduces a structural reduction method that removes nested optimization layers in dynamic games using offline-compiled best-response maps, enabling more efficient equilibrium computation.

Findings

The reduced problem solutions correspond to local Nash equilibria when the best-response operator is exact.

The learned surrogate provides approximately equilibrium-consistent solutions.

The method outperforms state-of-the-art solvers in solution quality and computational efficiency.

Abstract

Dynamic games are powerful tools to model multi-agent decision-making, yet computing Nash (generalized Nash) equilibria remains a central challenge in such settings. Complexity arises from tightly coupled optimality conditions, nested optimization structures, and poor numerical conditioning. Existing game-theoretic solvers address these challenges by directly solving the joint game, typically requiring explicit modeling of all agents' objective functions and constraints, while learning-based approaches often decouple interaction through prediction or policy approximation, sacrificing equilibrium consistency. This paper introduces a conceptually novel formulation for dynamic games by restructuring the equilibrium computation. Rather than solving a fully coupled game or decoupling agents through prediction or policy approximation, a data-driven structural reduction of the game is proposed that removes nested optimization layers and derivative coupling by embedding an offline-compiled best-response map as a feasibility constraint. Under standard regularity conditions, when the best-response operator is exact, any converged solution of the reduced problem corresponds to a local open-loop Nash (GNE) equilibrium of the original game; with a learned surrogate, the solution is approximately equilibrium-consistent up to the best-response approximation error. The proposed formulation is supported by mathematical proofs, accompanying a large-scale Monte Carlo study in a two-player open-loop dynamic game motivated by the autonomous racing problem. Comparisons are made against state-of-the-art joint game solvers, and results are reported on solution quality, computational cost, and constraint satisfaction.