Optimized task assignment and predictive maintenance for industrial machines using Markov decision process

This paper considers a distributed decision-making approach for manufacturing task assignment and condition-based machine health maintenance. We consider information sharing between the task assignment and health management agents. The proposed design of the agents uses Markov decision processes. A key advantage of using a Markov decision process-based approach is the incorporation of uncertainty into the decision-making process. The paper provides detailed mathematical models along with the associated practical execution strategy. To demonstrate the effectiveness and practical applicability of our proposed approach, we have included a detailed numerical case study that is based on open-source milling machine tool degradation data. Our case study indicates that the proposed approach offers flexibility in terms of the selection of cost parameters, and it allows for offline computation and analysis of the decision-making policy. These features create an opportunity for future work on learning the cost parameters associated with our proposed model using artificial intelligence.