Solving Permutation Flowshop Problem with Deep Reinforcement Learning

Permutation flowshop problem is a classic problem in combinatorial optimization. In this paper, we propose a deep reinforcement learning (DRL) model with heterogeneous network according to the different task characteristics of actor and critic in actor-critic model. The actor, acting as the policy network, is mainly responsible for strategy search and composed of LSTMs; the critic, acting as the value network, is mainly responsible for strategy evaluation and formed by the attention network. In order to increase the exploration ability of the model, we adopt the epsilon-greedy strategy to further improve the effectiveness of the model. The experimental results on multiple data sets show that our model achieves better performance on permutation flowshop problem.