Interactive Distributed Configuration Peter Tiedemann, Tarik Hadzic, Thomas Stuart Henney, and Henrik Reif Andersen Computational Logic and Algorithms Group, IT University of Copenhagen, Rued Langgaards Vej 7, DK-2300 Copenhagen S, Denmark {petert, tarik, tsh, hra}@itu.dk Abstract. Interactive configuration is the process of assisting a user in selecting values for parameters that respect given constraints. In- spired by the increasing demand for the real-time configuration in Supply Chain Management, we apply a compilation approach to the problem of interactive distributed configuration where the user options depend on constraints fragmented over a number of different locations connected through a network. We formalize the problem, suggest a solution ap- proach based on an asynchronous compilation scheme, and perform experimental verification. 1 Introduction An interactive configuration problem is an application of Constraint Satisfac- tion Problems (CSP) where a user is assisted in interactively assigning values to variables by a software tool. This configurator assists the user by displaying the valid choices for each unassigned variable in what are called valid domains computations. In many application domains user options depend on constraints fragmented over a number of different locations connected through a network. This paper attempts to extend the functionality provided by centralized config- uration to distributed configuration while taking advantage of the distribution of the problem when possible. One of the current solution approaches to interac- tive configuration is to divide the computational effort into an offline and online phase. First compile the set of all valid solutions offline into a compact symbolic representation based on Binary Decision Diagrams (BDD). In the on-line phase we can deliver valid domains computation in time bounded polynomially in the size of the BDD representation[1,2,3]. 2 Interactive Distributed Configuration One application domain where constraint information is fragmented is in Sup- ply Chain Management [4], where the user options depend on many interrelated businesses. The dominating relationship in a supply chain is the supplier pro- ducing items offered to consumers. We model these relationships with a supply chain graph (SCG), which is a directed graph G(N ,A) where every node N i ∈N F. Benhamou (Ed.): CP 2006, LNCS 4204, pp. 761–765, 2006. c  Springer-Verlag Berlin Heidelberg 2006