Tomas Balyo, Markus Iser

The constraint-based product configurator CAS Merlin has to solve an NP-complete problem after every user click. Therefore, in the worst case, each click may take exponential time (in the size of the product structure) to evaluate. Considering the high complexity of the products of our customers (even the smaller ones) this leads to astronomical values. Fortunately, in most cases we are very far away from the worst case and most configuration clicks are evaluated in a fraction of a second. Nevertheless, the actual computation time may be anywhere between a few milliseconds and billions of years. The goal of this thesis is to design and implement a way to improve the upper bound, i.e., to automatically estimate the complexity of configuration clicks for a given product ruleset. This would be very useful for a modeler, that could use this information to tweak the rules of their products in order to improve the configuration performance.