GPU-friendly, auto-differentiable LQR solver with JAX.

Replication Materials for "Very Simple Markov-Perfect Industry Dynamics: Theory"

A two-player dynamic game where deep Q-learning agents collaborate to find optimal solutions to a stochastic version of the graph coloring problem, featuring α-Rank for joint policy ranking and evaluation.

HARDPIG a class of credible policy games with heterogeneous agents

a challenging console game