Rabe, M.; Dross, F.: A Reinforcement Learning Approach for a Decision Support System for Logistics Networks. In: Yilmaz, L.; Chan, W. K. V.; Moon, I.; Roeder, T. M. K.; Macal, C.; Rossetti, M. D. (Hrsg.): Proceedings of the 2015 Winter Simulation Conference. Piscataway: IEEE 2015, S. 2020-2032.

This paper presents the architecture and working principles of a Decision Support System (DSS) for logistics networks. The system relies on a data-driven discrete-event simulation model. A brief introduction to Reinforcement Learning (RL) and an explanation of the adoption of RL to the concepts of the DSS is given. An illustration of the realization is presented using a specific aspect of a logistics network. The logistics network is described in a data model which is represented by database tables. The tables are used to dynamically instantiate the simulation model. The authors describe how SQL queries can be used to model actions of an RL agent. A Data Warehouse can be used to measure Key Performance Indicators on the simulation output data of the simulation model, which can be used as a reward criterion for the RL agent. The paper presents a basis for the ongoing development of an RL agent.