Executable product models - The intelligent way

Executable product models are an alternative approach in business process management with the objective to increase the flexibility during the execution of business processes in the service industry. This approach uses a special product model based on information dependencies comprising a compact representation of the set of all possible execution paths. To improve this approach and to take advantage from the flexibility provided we combine the multi-agent system with relational reinforcement learning and a genetic algorithm. We give an insight in the challenges of this approach and show how the efficiency is increased significantly.