Corpus ID: 211258813

APAC-Net: Alternating the Population and Agent Control via Two Neural Networks to Solve High-Dimensional Stochastic Mean Field Games

@article{Lin2020APACNetAT,
  title={APAC-Net: Alternating the Population and Agent Control via Two Neural Networks to Solve High-Dimensional Stochastic Mean Field Games},
  author={A. T. Lin and Samy Wu Fung and Wuchen Li and L. Nurbekyan and S. Osher},
  journal={ArXiv},
  year={2020},
  volume={abs/2002.10113}
}
  • A. T. Lin, Samy Wu Fung, +2 authors S. Osher
  • Published 2020
  • Mathematics, Computer Science
  • ArXiv
  • We present APAC-Net, an alternating population and agent control neural network for solving stochastic mean field games (MFGs). Our algorithm is geared toward high-dimensional instances MFGs that are beyond reach with existing solution methods. We achieve this in two steps. First, we take advantage of the underlying variational primal-dual structure that MFGs exhibit and phrase it as a convex-concave saddle point problem. Second, we parameterize the value and density functions by two neural… CONTINUE READING
    A mean field game inverse problem
    OT-Flow: Fast and Accurate Continuous Normalizing Flows via Optimal Transport
    3
    Controlling Propagation of epidemics via mean-field games.
    1
    A Simple Multiscale Method for Mean Field Games

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 32 REFERENCES
    A mean-field optimal control formulation of deep learning
    66
    Smoothness and Stability in GANs
    9
    Generative Adversarial Nets
    17094
    Connecting GANs and MFGs
    3
    FFJORD: Free-form Continuous Dynamics for Scalable Reversible Generative Models
    213
    Improved Training of Wasserstein GANs
    3215
    How to train your neural ODE
    15
    Variational Mean Field Games
    55
    The variational structure and time-periodic solutions for mean-field games systems
    7