Abstract:Population games are games with a finite set of available strategies and an infinite number of players, in which the reward for choosing a given strategy is a function of the distribution of players over strategies. The paper shows that, in a certain class of maxmin optimization problems, it is possible to associate a population game to a given maxmin problem in such a way that solutions to the optimization problem are found from Nash equilibria of the associated game. Iterative solution methods for maxmin optimization problems can then be derived from systems of differential equations whose trajectories are known to converge to Nash equilibria. In particular, we use a discrete-time version of the celebrated replicator equation of evolutionary game theory, also known in machine learning as the exponential multiplicative weights algorithm. The resulting algorithm can be viewed as a generalization of the Iteratively Reweighted Least Squares (IRLS) method, which is well known in numerical analysis as a useful technique for solving Chebyshev function approximation problems on a finite grid. Examples are provided to show the use of the generalized IRLS method in collective investment and in decision making under model uncertainty.

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve This paper aims to address a specific class of maxmin optimization problems. Specifically, the authors propose a method that connects maxmin optimization problems with population games, solving these optimization problems by finding Nash equilibria. #### Main Contributions 1. **Tester Games**: Introduced the concept of "Tester Games" and proved that the Nash equilibria in these games can provide solutions to certain types of maxmin optimization problems. 2. **Generalized IRLS Method**: Proposed an algorithm based on the replicator equation in exponential form, which is a generalization of the classical Iteratively Reweighted Least Squares (IRLS) method. #### Methodology - **Population Games**: Defined a population game with a finite strategy set and infinite players, where the reward for choosing a strategy depends on the distribution of players across different strategies. - **Nash Equilibrium**: Solved the maxmin optimization problem by finding Nash equilibria. The authors demonstrated how to derive iterative solutions to the maxmin optimization problem by converging to Nash equilibria through the trajectories of a system of differential equations. #### Application Examples - **Collective Investment**: Demonstrated how to use the generalized IRLS method for collective investment decision-making. - **Decision-Making under Model Uncertainty**: Discussed how to make decisions under model uncertainty. #### Theoretical Background - **Social Planning Perspective**: Viewed the maxmin problem from a social planning perspective, aiming to find a point in a given set that maximizes the minimum value of a set of functions over that set. - **Ethical Principles**: Adopted John Rawls' ethical principle, which is to choose the social organization method that makes the least advantaged group the happiest among all possible social organizations. #### Innovations - **Weight Update Rule**: Proposed using the update rules in population games to solve maxmin optimization problems, which is different from the traditional two-person zero-sum game approach. - **Continuous-Time Dynamical Systems**: Obtained a series of potential iterative solutions by discretizing continuous-time dynamical systems. Through the above methods, the paper provides a new theoretical framework and algorithmic tools to solve maxmin optimization problems and validates them in practical applications.

Solving Maxmin Optimization Problems via Population Games

A new hybrid genetic algorithm for global minimax optimization

Population Games and Discrete Optimal Transport

Dynamic Population Games: A Tractable Intersection of Mean-Field Games and Population Games

Learning, evolution and population dynamics

Maxent in Experimental 2$\times$2 Population Games

A population-based game-theoretic optimizer for the minimum weighted vertex cover

On the population size in stochastic differential games

A particle consensus approach to solving nonconvex-nonconcave min-max problems

Coupled Wasserstein Gradient Flows for Min-Max and Cooperative Games

Solving Optimization Problems by the Public Goods Game

Population games on dynamic community networks

A Geometric Embedding Approach to Multiple Games and Multiple Populations

Bio-Inspired Dynamic Collective Choice in Large-Population Systems: A Robust Mean-Field Game Perspective

Optimal investment in a large population of competitive and heterogeneous agents

On finite population games of optimal trading

The impact of a specific blend of essential oil components and sodium butyrate in feed on growth performance and Salmonella counts in experimentally challenged broilers.

On Adaptive-Gain Control of Replicator Dynamics in Population Games

Beyond Minimax Optimality: A Subgame Perfect Gradient Method

LEAD: Min-Max Optimization from a Physical Perspective

Identifying Socially Optimal Equilibria Using Combinatorial Properties of Nash Equilibria in Bimatrix Games