Double Deep-Q Learning-Based Output Tracking of Probabilistic Boolean Control Networks

Abstract

In this article, a reinforcement learning (RL)-based scalable technique is presented to control the probabilistic Boolean control networks (PBCNs). In particular, a double deep-Q network (DDQN) approach is firstly proposed to address the output tracking problem of PBCNs, and optimal state feedback controllers are obtained such that the output of PBCNs tracks a constant as well as a time-varying reference signal. The presented method is model-free and offers scalability, thereby provides an efficient way to control large-scale PBCNs that are a natural choice to model gene regulatory networks (GRNs). Finally, three PBCN models of GRNs including a 16-gene and 28-gene networks are considered to verify the presented results.

Keywords

• Q learning%22%2C%22default_operator%22%3A%22AND%22%7D%7D%7D"> Double deep-<italic xmlns:ali="http://www.niso.org/schemas/ali/1.0/" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">Q</italic> learning
• model-free technique
• output tracking
• probabilistic Boolean control networks
• scalability