Intelligent supporting techniques for the maintenance of constraint-based configuration systems

Constraint-based systems like knowledge-based recommendation and configuration are well established technologies in many different product areas like cars, computers, notebooks, and financial services. Such systems reduce the number of valid products and configurations regarding the customers’ preferences. The relationship between product variables and customer questions is represented by constraints. Nowadays, constraint-based configuration systems represent volatile product assortments: Variables must be adapted, new product features lead to new questions for the customer, and / or the constraints must be updated. We call such scenarios maintenance tasks. In complex constraint-based configuration systems the maintenance task is time consuming and error prone. Previous research focused on the detection of conflicts, repair actions for the conflicts, and redundant constraints. In this paper we give an overview about these techniques and present new approaches like recommendation, well-formedness violation, simulation, and knowledge base verification for the support of knowledge engineers.