Learning Distributed Equilibria in Linear-Quadratic Stochastic Differential Games: An $α$-Potential Approach
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912911994126336 |
|---|---|
| author | Plank, Philipp Zhang, Yufei |
| author_facet | Plank, Philipp Zhang, Yufei |
| contents | We analyze independent policy-gradient (PG) learning in $N$-player linear-quadratic (LQ) stochastic differential games. Each player employs a distributed policy that depends only on its own state and updates the policy independently using the gradient of its own objective. We establish global linear convergence of these methods to an equilibrium by showing that the LQ game admits an $α$-potential structure, with $α$ determined by the degree of pairwise interaction asymmetry. For pairwise-symmetric interactions, we construct an affine distributed equilibrium by minimizing the potential function and show that independent PG methods converge globally to this equilibrium, with complexity scaling linearly in the population size and logarithmically in the desired accuracy. For asymmetric interactions, we prove that independent projected PG algorithms converge linearly to an approximate equilibrium, with suboptimality proportional to the degree of asymmetry. Numerical experiments confirm the theoretical results across both symmetric and asymmetric interaction networks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_16555 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Learning Distributed Equilibria in Linear-Quadratic Stochastic Differential Games: An $α$-Potential Approach Plank, Philipp Zhang, Yufei Optimization and Control Machine Learning Probability 91A06, 91A14, 91A15, 91A43, 49N10, 68Q25 We analyze independent policy-gradient (PG) learning in $N$-player linear-quadratic (LQ) stochastic differential games. Each player employs a distributed policy that depends only on its own state and updates the policy independently using the gradient of its own objective. We establish global linear convergence of these methods to an equilibrium by showing that the LQ game admits an $α$-potential structure, with $α$ determined by the degree of pairwise interaction asymmetry. For pairwise-symmetric interactions, we construct an affine distributed equilibrium by minimizing the potential function and show that independent PG methods converge globally to this equilibrium, with complexity scaling linearly in the population size and logarithmically in the desired accuracy. For asymmetric interactions, we prove that independent projected PG algorithms converge linearly to an approximate equilibrium, with suboptimality proportional to the degree of asymmetry. Numerical experiments confirm the theoretical results across both symmetric and asymmetric interaction networks. |
| title | Learning Distributed Equilibria in Linear-Quadratic Stochastic Differential Games: An $α$-Potential Approach |
| topic | Optimization and Control Machine Learning Probability 91A06, 91A14, 91A15, 91A43, 49N10, 68Q25 |
| url | https://arxiv.org/abs/2602.16555 |