Chapter 24 - Configuration

https://doi.org/10.1016/S1574-6526(06)80028-3Get rights and content

Publisher Summary

This chapter explains that configuration is the task of composing a customized system out of generic components. This task is of universal concern as component-based systems are omnipresent in modern industry. Components are generic in nature and can be produced in mass, however are destined to support customized solutions. The available components for a given type of system are usually described in the form of a catalog. Each catalog item is a product type and describes the functional and technical characteristics of the component. For example, take a printer component whose function is printing. The functional characteristics include printing quality and support of colors. The technical characteristics include printing type and printing speed. Whereas the catalog describes the generic knowledge of components, a customer usually has specific requirements for the desired component-based system. The chapter also discussescomplex configuration problems encountered in the engineering and manufacturing departments of the computer and automotive industries, which need to complete customized sales orders by choosing suitable parts from huge catalogs. This situation has led to the development of configurators that do this completion automatically while respecting difficult technical constraints. However, configurators can also support the sales process and assist the user in choosing options while guaranteeing their compatibility.

Section snippets

A Whole Spectrum of Problems

Configuration has been an outgrowth of research on rule-based expert systems. John McDermott [41] used the term configuration for a specific form of a design task [9], where a system was assembled out of predefined components that are connected in predefinedways. Whereas more innovative design tasks often require a suitable modeling of the physical behavior of components [64], work on rule-based configurators focused on the functional aspects of components. Frayman and Mittal summarized those

Configuration Knowledge

This section introduces the knowledge that exists about components, namely component catalogs in sub-Section 24.2.1, component structure in sub-Section 24.2.2, and component constraints in sub-Section 24.2.2. It introduces the different ingredients of a component such as functions (features), attributes, subcomponents and connections, and resources and provides the basic vocabulary for formulating constraint models.

Constraint Models for Configuration

A configuration problem is defined by functional requirements and by a configuration model that describes the possible configurations according to the given configuration knowledge. The configuration models in the literature differ quite substantially in the way choices and constraints are represented, although they use the same product models at their origins. It is indeed possible to build different constraint models for the given configuration knowledge. It is even possible to use different

Problem Solving for Configuration

Whereas Constraint Programming is used to find a solution or an optimal solution of a CSP, configuration brings in new problem-solving tasks, such as the maintenance of global consistency in interactive configuration, the computation of an explanation if a given set of requirements cannot be satisfied, or the computation of a relaxation of those requirements. The automatic computation of configurations also has particularities. Solution search should produce functionally well-justified

Conclusion

This chapter has given an overview on constraint-based configuration while stressing the particularities of configuration. Configuration is not a classical application of ConstraintProgramming. The constraint networks are not static, but evolve during problem solving. As such, the formalism of constraint networks is not sufficient to specify a configuration problem and it is necessary to go up to a higher modeling level, which describes the configuration knowledge independent of a

Bibliography (65)

  • Rainer Weigel et al.

    Compiling constraint satisfaction problems

    Artificial Intelligence

    (1999)
  • Michel Aldanondo et al.

    Configurator and CAD: gathering the best of two worlds

  • Liliana Ardissono et al.

    A framework for the development of personalized, distributed web-based configuration systems

    AI Magazine

    (2003)
  • Virginia E. Barker et al.

    Expert systems for configuration at Digital: XCON and beyond

    Communications of the ACM

    (1989)
  • Christian Bessière et al.

    Propagate the right thing: how preferences can speed-up constraint solving

  • Craig Boutilier et al.

    Preference-based constrained optimization with CP-nets

    Computational Intelligence

    (2004)
  • James Bowen et al.

    Conditional existence of variables in generalised constraint networks

  • Gerhard Brewka

    Preferred subtheories: An extended logical framework for default reasoning

  • David C. Brown et al.

    Design problem solving: knowledge structures and control strategies

    (1989)
  • Roman Cunis et al.

    Plakon — an approach to domain-independent construction

  • Adnan Darwiche

    New advances in compiling CNF into decomposable negation normal form

  • Alexander Felfernig et al.

    UML as domain specific language for the construction of knowledge-based configurations systems

    International Journal of Software Engineering and Knowledge Engineering

    (2000)
  • Gerhard Fleischanderl et al.

    Configuring large systems using generative constraint satisfaction

    IEEE Intelligent Systems

    (1998)
  • Felix Frayman et al.

    COSSACK: A constraint-based expert system for configuration

  • Eugene C. Freuder

    Synthesizing constraint expressions

    Communications of the ACM

    (1978)
  • Eugene C. Freuder

    A sufficient condition for backtrack-free search

    Journal of the ACM

    (1982)
  • Eugene C. Freuder

    Eliminating interchangeable values in constraint satisfaction problems

  • Gerhard Friedrich

    Elimination of spurious explanations

  • Gerhard Friedrich et al.

    Consistency-based configuration

  • Felix Geller et al.

    Assumption-based pruning in conditional CSP

  • Michael R. Genesereth et al.

    Logical Foundations of Artificial Intelligence

    (1987)
  • Andreas Günter et al.

    Knowledge-based configuration: Survey and future directions

  • Cited by (13)

    • Tacton: Use of Tacton Configurator at FLSmidth

      2014, Knowledge-Based Configuration: From Research to Business Cases
    • Configuration Knowledge Representation and Reasoning

      2014, Knowledge-Based Configuration: From Research to Business Cases
    • Using On-The-Fly Model Checking to improve Constraint Programming for Dynamic Problems

      2023, Proceedings - International Conference on Tools with Artificial Intelligence, ICTAI
    View all citing articles on Scopus
    View full text