Beschreibung
Due to continuously changing requirements in production and logistics, future material flow systems have to provide solutions with increased flexibility and adaptivity. Especially material handling systems in warehousing and distribution are characterized by much shorter processing times than external transportation of goods. In this contribution a reinforcement learning method for dispatching strategies in intralogistics is presented. This method is implemented in a simulation model for a cross-docking distribution center and evaluated in comparison with common heuristic dispatching strategies. The developed method generates for a user-configurable setup a corresponding decision table for controlling the behavior of the implicated material handling equipment in the regarded application scenario. Each action is chosen based on the current operational situation. This leads to an improvement of performance indicators compared with the examined heuristic strategies.