EXOTIC: An Exact, Optimistic, Tree-Based Algorithm for Min-Max Optimization

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Maheshwari, Chinmay, Pimpalkhare, Chinmay, Chatterjee, Debasish
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866914593061732352
author Maheshwari, Chinmay
Pimpalkhare, Chinmay
Chatterjee, Debasish
author_facet Maheshwari, Chinmay
Pimpalkhare, Chinmay
Chatterjee, Debasish
contents Min-max optimization arises in many domains such as game theory, adversarial machine learning, etc. For these problems, gradient-based methods are well understood and enjoy strong guarantees. However, in the absence of convexity or concavity, existing approaches study convergence to an approximate saddle point or first-order stationary points, which may be arbitrarily far from global optima. In this work, we present an algorithmic framework for computing the global minimax value in convex--non-concave and non-convex--concave min-max optimization. For convex--non-concave min-max problems, we use a reformulation that transforms the problem into a non-concave--convex max-min optimization problem with suitably defined feasible sets and objective function. This reformulation can be viewed as an extension of Sion's minimax theorem to the convex--non-concave setting. We then introduce EXOTIC -- an Exact, Optimistic, Tree-based algorithm for solving the reformulated max-min problem. EXOTIC combines an iterative convex optimization solver for the inner minimization with an optimistic hierarchical tree search for the outer maximization, inspired by StroquOOL~\cite{bartlett2019simple}. Unlike StroquOOL, which assumes stochastic zero-mean noisy evaluations, EXOTIC handles deterministic, biased, and budget-dependent evaluation errors arising from finite-time solutions of the inner convex subproblems. We establish an upper bound on its optimality gap. The same framework also applies to non-convex--concave min-max optimization. Empirically, EXOTIC outperforms gradient-based methods on popular benchmarks from the literature. Finally, we demonstrate the utility of EXOTIC by computing security strategies in multi-player games with three or more players -- a computationally challenging task that, to our knowledge, no prior method solves exactly.
format Preprint
id arxiv_https___arxiv_org_abs_2508_12479
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle EXOTIC: An Exact, Optimistic, Tree-Based Algorithm for Min-Max Optimization
Maheshwari, Chinmay
Pimpalkhare, Chinmay
Chatterjee, Debasish
Optimization and Control
Artificial Intelligence
Computer Science and Game Theory
Multiagent Systems
General Economics
Economics
90C26, 90C47, 68Q32, 91A06, 65K05
Min-max optimization arises in many domains such as game theory, adversarial machine learning, etc. For these problems, gradient-based methods are well understood and enjoy strong guarantees. However, in the absence of convexity or concavity, existing approaches study convergence to an approximate saddle point or first-order stationary points, which may be arbitrarily far from global optima. In this work, we present an algorithmic framework for computing the global minimax value in convex--non-concave and non-convex--concave min-max optimization. For convex--non-concave min-max problems, we use a reformulation that transforms the problem into a non-concave--convex max-min optimization problem with suitably defined feasible sets and objective function. This reformulation can be viewed as an extension of Sion's minimax theorem to the convex--non-concave setting. We then introduce EXOTIC -- an Exact, Optimistic, Tree-based algorithm for solving the reformulated max-min problem. EXOTIC combines an iterative convex optimization solver for the inner minimization with an optimistic hierarchical tree search for the outer maximization, inspired by StroquOOL~\cite{bartlett2019simple}. Unlike StroquOOL, which assumes stochastic zero-mean noisy evaluations, EXOTIC handles deterministic, biased, and budget-dependent evaluation errors arising from finite-time solutions of the inner convex subproblems. We establish an upper bound on its optimality gap. The same framework also applies to non-convex--concave min-max optimization. Empirically, EXOTIC outperforms gradient-based methods on popular benchmarks from the literature. Finally, we demonstrate the utility of EXOTIC by computing security strategies in multi-player games with three or more players -- a computationally challenging task that, to our knowledge, no prior method solves exactly.
title EXOTIC: An Exact, Optimistic, Tree-Based Algorithm for Min-Max Optimization
topic Optimization and Control
Artificial Intelligence
Computer Science and Game Theory
Multiagent Systems
General Economics
Economics
90C26, 90C47, 68Q32, 91A06, 65K05
url https://arxiv.org/abs/2508.12479