Reinforcement learning for combined production-maintenance and quality control of a manufacturing system with deterioration failures

•The addressed research problem is heavily correlated with real-world industry problems.•Joint control policies are derived dynamically by means of Reinforcement Learning.•The derived joint policies can optimize successfully the cost-effectiveness of stochastic manufacturing systems.•The performance of integrated joint policies is proven to be superior to that of ad-hoc control policies, such as Kanban.•The framework ensures intelligently the high quality of final items and the operability of manufacturing facilities.