Skip to main content

DAC: The Double Actor−Critic Architecture for Learning Options

Shangtong Zhang and Shimon Whiteson


We reformulate the option framework as two parallel augmented MDPs. Under this novel formulation, all policy optimization algorithms can be used off the shelf to learn intra-option policies, option termination conditions, and a master policy over options. We apply an actor-critic algorithm on each augmented MDP, yielding the Double Actor-Critic (\AC) architecture. Furthermore, we show that, when state-value functions are used as critics, one critic can be expressed in terms of the other, and hence only one critic is necessary. We conduct an empirical study on challenging robot simulation tasks. In a transfer learning setting, \AC outperforms both its hierarchy-free counterpart and previous gradient-based option learning algorithms.

Book Title
NeurIPS 2019: Proceedings of the Thirty−third Annual Conference on Neural Information Processing Systems