Equilibrium Dynamics in Market Games with Exchangeable and Divisible Resources
José Correa, Tobias Harks, Anja Schedel, José Verschae
Abstract
We study a market game with n ≥ 2 players competing over m ≥ 1 divisible resources of different finite capacities. Resources are traded via the proportional sharing mechanism, where players are price-anticipating, meaning that they can influence the prices with their bids. Additionally, each player has an initial endowment of the resources which are sold at market prices. Although the players’ total profit functions may be discontinuous in the bids, we prove existence and uniqueness of pure Nash equilibria of the resulting market game. Then, we study a discrete dynamic arising from repeatedly taking the (unique) equilibrium resource allocation as initial endowments for the next market game. We prove that the total utility value of the dynamic converges to either an optimal allocation value (maximizing total utility over the allocation space) or to a restricted optimal allocation value, where the restriction is defined by fixing some tight resources which are exclusively allocated to a single player. As a corollary, it follows that for strictly concave utility functions, the aggregated allocation vector of the dynamic converges to the unique (possibly restricted) optimal aggregated allocation, and for linear utility functions, we even get convergence of the dynamic to a (possibly restricted) optimal solution in the (non-aggregated) original allocation space.
Ask about this paper
Ask your agent about it.
Lune has read the top-tier papers around this one, so every answer names the papers it rests on.
Your agent calls
Lunesearch_papers
Free to start. No credit card required.
Terminal
Install the CLIlune papers get 854d58d4-80ab-4574-bf3a-7412957d4486Related papers
- Asynchronous Proportional Response Dynamics: Convergence in Markets with Adversarial SchedulingYoav Kolumbus, Menahem Levy, Noam NisanNeurIPS 2023 · 9 citations
- Online Market Equilibrium with Application to Fair DivisionYuan Gao, Alex Peysakhovich, Christian KroerNeurIPS 2021 · 35 citations
- Revisiting the Bertrand Paradox via Equilibrium Analysis of No-regret LearnersARNAB MAITI, Junyan Liu, Kevin Jamieson, Lillian RatliffICML 2026
- PPA-Game: Characterizing and Learning Competitive Dynamics Among Online Content CreatorsRenzhe Xu, Haotian Wang, Xingxuan Zhang, Bo Li et al.KDD 2025 · 1 citation
- Approximations for Indivisible Concave Allocations with Applications to Nash Welfare MaximizationNathaniel Kell, Kevin SunAAAI 2023
