Proceedings of the Workshop on Configuration at ECAI 2012 (ConfWS'12) August 27, 2012 Montpellier, France Wolfgang Mayer and Patrick Albert, Editors Foreword In many cases competitiveness of modern products is defined by the degree of customization, i.e. the ability of a manufacturer to adapt a product according to customer requirements. Knowledge based configuration methods support the composition of complex systems from a set of adjustable components. However, there are two important prerequisites for a successful application of knowledge-based configuration in practice: (a) expressive knowledge representation languages, which are able to capture the complexity of various models of configurable products and (b) powerful reasoning methods which are capable of providing services such as solution search, optimization, diagnosis, etc. The Configuration Workshop aims to bring together industry representatives and researchers from various areas of AI to identify important configuration scenarios found in practice, exchange ideas and experiences and present original methods developed to solve configuration problems. The workshop continues the series of successful Configuration Workshops started at the AAAI'96 Fall Symposium and continued on IJCAI, AAAI, and ECAI since 1999. During this time the focus of the events broadened from configuration approaches applied to traditional products such as cars, digital cameras, PC, telecommunication switches or railway interlock systems to configuration of software and services available on the Web. In parallel, research in the field of constraint programming, description logic, non-monotonic reasoning, fundamentals of configuration modeling and so forth pushed the limits of configuration systems even further. The papers selected this year for presentation on the Configuration Workshop continue a recent trend in the research community and focus on modeling and solving of configuration problems. Wolfgang Mayer and Patrick Albert July 2012 Wolfgang Mayer and Patrick Albert, Editors. Proceedings of the Workshop on Configuration at ECAI 2012 (ConfWS'12), August 27, 2012, Montpellier, France. Workshop Organization Program Co-Chairs Wolfgang Mayer University of South Australia, Adelaide, Australia Patrick Albert IBM, France Program Committee Tomas Axling Tacton Systems AB, Sweden Claire Bagley Oracle Corporation, USA Conrad Drescher University of Oxford, UK Alexander Felfernig Technische Universit¨atGraz/Configworks, Austria Gerhard Friedrich AAU Klagenfurt, Austria Albert Haag SAP AG, Germany Alois Haselboeck Siemens AG, Austria Lothar Hotz Universit¨atHamburg, Germany Klas Orsvarn Tacton System AB, Sweden Markus Stumptner University of South Australia, Australia Barry O'Sullivan Cork Constraint Computation Centre, Ireland Juha Tiihonen Aalto University, Finland Wolfgang Mayer and Patrick Albert, Editors. Proceedings of the Workshop on Configuration at ECAI 2012 (ConfWS'12), August 27, 2012, Montpellier, France. Contents Papers Maintaining alternative values in constraint-based configuration...........................1 Caroline Becker and H´el`eneFargier K-Model { Structured Design of Configuration Models................................8 Axel Brinkop, Thorsten Krebs, and Hartmut Schlee Towards a Formalism of Configuration Properties Propagation........................... 15 David Fabian, Radek Maˇr´ık,and Tom´aˇsOberhuber Testing Object-Oriented Configurators With ASP.................................. 21 Andreas Falkner, Gottfried Schenner, Gerhard Friedrich, and Anna Ryabokon Towards Hybrid Techniques for Efficient Declarative Configuration......................... 27 Ingo Feinerer Unifying Software, Product Configuration: A Research Roadmap.......................... 31 Arnaud Hubaux, Dietmar Jannach, Conrad Drescher, Leonardo Murta, Tomi Mannisto, Krzysztof Czarnecki, Patrick Heymans, Tien Nguyen, and Markus Zanker An Improved Constraint Ordering Heuristics for Compiling Configuration Problems............... 36 Benjamin Matthes, Christoph Zengler, and Wolfgang K¨uchlin Concurrent configuration and planning problems: Some optimization experimental results............ 41 Paul Pitiot, Michel Aldanondo, Elise Vareilles, and Paul Gaborit Sales configurator capabilities to prevent product variety from backfiring...................... 47 Alessio Trentin, Elisa Perin, and Cipriano Forza Author Index 55 Wolfgang Mayer and Patrick Albert, Editors. Proceedings of the Workshop on Configuration at ECAI 2012 (ConfWS'12), August 27, 2012, Montpellier, France. Caroline Becker and H´el`ene Fargier 1 Maintaining alternative values in constraint-based configuration Caroline Becker and H´el`eneFargier 1 Abstract. Constraint programming techniques are widely components, options, or more generally by a set of attributes, used to model and solve interactive decision problems, an es- the values of which have to be chosen by the user. These pecially configuration problems. In this type of application, values must satisfy a finite set of configuration constraints the configurable product is described by means of a set of con- that encode the feasibility of the product, the compatibility straint bearing on the configuration variables. The user then between components, their availability, etc. interactively solves the CSP by assigning (and possibly, relax- Several extensions of the CSP paradigm have been pro- ing) the configuration variables according to her preferences. posed in order to handle the constraints-based definition of a The aim of the system is then to keep the domains of the other catalog or a range of products, and more specifically the def- variables consistent with these choices. Since maintaining of inition of configurable products. These extensions have been the global inverse consistency is generally not tractable, the motivated by difficulties and characteristics that are specific domains are instead filtered according to some level of local to the modeling and the handling of catalogs of configurable consistency, e.g. arc-consistency. products. Dynamic CSPs [13], for instance suit the problems In the present work, we aim at offering a more convenient where the existence of some optional variables depends on interaction by providing the user with possible alternative val- the value of another variable. Other extensions proposed by ues for each of the already assigned variables - i.e. the values the CSP community include composite CSPs [17], interactive that could replace the current one without leading to the vio- CSPs [10], hypothesis CSPs [1], generative constraint satis- lation of some constraint. We thus present the new concept of faction [19, 7], etc. alternative domains in a (possibly) partially assigned CSP. In this article, we do not deal with such representation prob- We propose a propagation algorithm that computes all the al- lems: we assume that the product range is specified by a clas- ternative domains in a single step. Its worst case complexity sical CSP. Instead, our work focuses on the human-computer is comparable with the one of the naive algorithm that would interaction. When configuring a product, the user specifies her run a full propagation for each variable, but its experimental requirements by interactively giving values to variables. Each efficiency is much better. time a new choice is made, the domains of the variables must be pruned so as to ensure that the values available for the fur- 1 Introduction ther variables can lead to a feasible product (i.e., a product satisfying all the initial configuration constraints): the aim of The Constraint Satisfaction Problem (CSP) formalism offers the system is to keep the domains of the other variables con- a powerful framework for representing a great variety of prob- sistent with these choices. Since the maintaining of the global lems, e.g. routing problems, resource allocation, frequency inverse consistency is generally not tractable, the domains are assignment, configuration problems, etc. The main task ad- rather filtered according to some level of local consistency, e.g. dressed by the algorithms is the determination of the consis- arc-consistency. In the present paper, we propose to make this tency of the CSP and/or the search for an (optimal) solution, interaction more user-friendly by showing not only (locally) and this is a difficult task: determining whether a CSP is con- consistent domains, but also what we call the alternative do- sistent is an NP-complete request. In the CSP community, mains of the assigned variables, i.e. the values that could re- the main research stream thus addresses this question, either place the one of the assigned variable without leading to the directly (looking for efficient complete algorithms) or getting violation of some constraint. around (studying the polynomial subclasses or proposing in- The structure of the present article is as follows: the prob- complete algorithms). lematics of alternative domains is described in the next Sec- But these algorithms do not help solving decision support tion. Section 3 then develop the basis of our algorithm. Our problems that are interactive in essence. For such problems, first experimental results are shown in Section 4. Proofs are the user herself is in charge of the choice of values for the gathered in Appendix. variables and the role of the system is not to solve a CSP, but to help the user in this task. Constraint-based product configuration [14, 18, 12, 19, 20] is a typical example of such 2 Background and Problematics problems: a configurable product is defined by a finite set of A CSP is classically defined