Paper Title
Reinforcement Learning Performance for Complex Job-Shop Scheduling Problems in Manufacturing System

Job-shop scheduling has become an important research issue due to complex planning in different services and industries. In an industrial environment, there is lack of defined order for multiple jobs execution, which results into larger number of expected solutions for job-shop scheduling problems. To address such problems, recent developed machine learning method, called deep reinforcing learning which learn directly by interacting the environment and experience, is used. In this paper, we present an overview of previous literature for various solutions of complex job shop scheduling using reinforcement learning. We also proposed an exemplary solution by employing deep Q-Network (DQN) method that makes a schedule by assigning numerous job operations to the multiple machine. The proposed method automatically acquires knowledge of available resources and proactively assigns operations to these resources on account of minimizing the waiting time of operations and idle time of machines. Keywords - Deep reinforcement learning, Q-leaning, smart manufacturing, production scheduling, job-shop scheduling