In the ROLF research project, Fraunhofer IIS and DATEV are investigating the potential of artificial intelligence (AI) in business management processes.
Decisions in business management processes are often complex and subject to a large number of dependencies. Reinforcement learning (RL), as a form of AI, is particularly suitable for these types of challenges. Here, an AI agent (software) learns to make the right decisions independently by interacting with its environment.
The project therefore aims to develop an AI based on RL that allows fast and efficient decisions to be made in highly complex situations. In this way, for example, managers in multidimensional business processes (such as personnel planning) can be relieved and supported in their decisions.
The main task of Fraunhofer IIS in the project is the methodical realization, i.e. the implementation and training of the actual agent that interacts with its environment. To this end, various reinforcement learning methods are being evaluated, corresponding algorithms implemented and interfaces for data transfer provided.
DATEV is responsible for project management and coordination. It defines the technical requirements based on practical experience and provides simulation environments. Application-related data is provided in the simulation environments for interaction with the AI agent. Depending on the use case, this can be, for example, texts, order situation or (machine) workloads.