A distributed approach solving partially flexible job-shop scheduling problem with a Q-learning effect

The advent of new technologies transforms the manufacturing paradigm and facilitates the emergence of "smart" factories. The increased flexibility of modern production tools increases the scheduling complexity. In this work, we propose to deal with Partially Flexible Job-shop Scheduling Problem using a heterarchical approach based on intelligent products. According to its manufacturing process, an Intelligent Product (IP) requests a set of services from Service Providers. The IP collects data to precise the current scheduling context. Using this context and applying a reinforcement Q-learning approach, the Intelligent Product chooses and applies the most suitable Machine Selection Rule and Dispatching Rule to deal with complex scheduling problems.