We consider the problem of representing collective behavior of large popula- tions and predicting the evolution of a population distribution over a discrete state space. A discrete time mean field game (MFG) is motivated as an interpretable model founded on game theory for understanding the aggregate effect of individ- ual actions and predicting the temporal evolution of population distributions. We achieve a synthesis of MFG and Markov decision processes (MDP) by showing that a special MFG is reducible to an MDP. This enables us to broaden the scope of mean field game theory and infer MFG models of large real-world systems via deep inverse reinforcement learning. Our method learns both the reward function and forward dynamics of an MFG from real data, and we report the first empirical test of a mean field game model of a real-world social media population.
more »
« less
Model-Free Mean-Field Reinforcement Learning: Mean-Field MDP and Mean-Field Q-Learning
We develop a general reinforcement learning framework for mean field control (MFC) problems. Such problems arise for instance as the limit of collaborative multi-agent control problems when the number of agents is very large. The asymptotic problem can be phrased as the optimal control of a non-linear dynamics. This can also be viewed as a Markov decision process (MDP) but the key difference with the usual RL setup is that the dynamics and the reward now depend on the state's probability distribution itself. Alternatively, it can be recast as a MDP on the Wasserstein space of measures. In this work, we introduce generic model-free algorithms based on the state-action value function at the mean field level and we prove convergence for a prototypical Q-learning method. We then implement an actor-critic method and report numerical results on two archetypal problems: a finite space model motivated by a cyber security application and a continuous space model motivated by an application to swarm motion.
more »
« less
- Award ID(s):
- 1716673
- NSF-PAR ID:
- 10169099
- Date Published:
- Journal Name:
- ArXivorg
- ISSN:
- 2331-8422
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
more » « less
Mean field games (MFG) and mean field control (MFC) are critical classes of multiagent models for the efficient analysis of massive populations of interacting agents. Their areas of application span topics in economics, finance, game theory, industrial engineering, crowd motion, and more. In this paper, we provide a flexible machine learning framework for the numerical solution of potential MFG and MFC models. State-of-the-art numerical methods for solving such problems utilize spatial discretization that leads to a curse of dimensionality. We approximately solve high-dimensional problems by combining Lagrangian and Eulerian viewpoints and leveraging recent advances from machine learning. More precisely, we work with a Lagrangian formulation of the problem and enforce the underlying Hamilton–Jacobi–Bellman (HJB) equation that is derived from the Eulerian formulation. Finally, a tailored neural network parameterization of the MFG/MFC solution helps us avoid any spatial discretization. Our numerical results include the approximate solution of 100-dimensional instances of optimal transport and crowd motion problems on a standard work station and a validation using a Eulerian solver in two dimensions. These results open the door to much-anticipated applications of MFG and MFC models that are beyond reach with existing numerical methods.
-
In this paper, we investigate how the self-synchronization property of a swarm of Kuramoto oscillators can be controlled and exploited to achieve target densities and target phase coherence. In the limit of an infinite number of oscillators, the collective dynamics of the agents’ density is described by a mean-field model in the form of a nonlocal PDE, where the nonlocality arises from the synchronization mechanism. In this mean-field setting, we introduce two space-time dependent control inputs to affect the density of the oscillators: an angular velocity field that corresponds to a state feedback law for individual agents, and a control parameter that modulates the strength of agent interactions over space and time, i.e., a multiplicative control with respect to the integral nonlocal term. We frame the density tracking problem as a PDE-constrained optimization problem. The controlled synchronization and phase-locking are measured with classical polar order metrics. After establishing the mass conservation property of the mean-field model and bounds on its nonlocal term, a system of first-order necessary conditions for optimality is recovered using a Lagrangian method. The optimality system, comprising a nonlocal PDE for the state dynamics equation, the respective nonlocal adjoint dynamics, and the Euler equation, is solved iteratively following a standard Optimize-then-Discretize approach and an efficient numerical solver based on spectral methods. We demonstrate our approach for each of the two control inputs in simulation.more » « less
-
We investigate reinforcement learning for mean field control problems in discrete time, which can be viewed as Markov decision processes for a large number of exchangeable agents interacting in a mean field manner. Such problems arise, for instance when a large number of robots communicate through a central unit dispatching the optimal policy computed by minimizing the overall social cost. An approximate solution is obtained by learning the optimal policy of a generic agent interacting with the statistical distribution of the states of the other agents. We prove rigorously the convergence of exact and model-free policy gradient methods in a mean-field linear-quadratic setting. We also provide graphical evidence of the convergence based on implementations of our algorithms.more » « less
-
The theory and application of mean field games has grown significantly since its origins less than two decades ago. This paper considers a special class in which the game is cooperative, and the cost includes a control penalty defined by Kullback-Leibler divergence, as commonly used in reinforcement learning and other fields. Its use as a control cost or regularizer is often preferred because this leads to an attractive solution. This paper considers a particular control paradigm called Kullback-Leibler Quadratic (KLQ) optimal control, and arrives at the following conclusions: 1. in application to distributed control of electric loads, a new modeling technique is introduced to obtain a simple Markov model for each load (the `agent' in mean field theory). 2. It is argued that the optimality equations may be solved using Monte-Carlo techniques---a specialized version of stochastic gradient descent (SGD). 3. The use of averaging minimizes the asymptotic covariance in the SGD algorithm; the form of the optimal covariance is identified for the first time.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- The DOI is not currently available.
