Skip to Main content Skip to Navigation

Knowledge-based configuration : a contribution to generic modeling, evaluation and evolutionary optimization

Abstract : In a context of mass customization, the concurrent configuration of the product and its production process constitute an important industrial challenge: Numerous options or alternatives, numerous links or constraints and a need to optimize the choices made. This problem is called O-CPPC (Optimization of Concurrent Product and Process Configuration). We consider this problem as a CSP (Constraints Satisfaction Problem) and optimize it with evolutionary algorithms. A state of the art shows that: i) most studies are illustrated with examples specific to an industrial or academic case and not representative of the existing diversity; ii) a need to improve optimization performance in order to gain interactivity and face larger problems. In response to the first point, this thesis proposes a generic model of the O-CPPC problem. This generic model is used to generate a realistic benchmark for evaluating optimization algorithms. This benchmark is then used to analyze the performance of the CFB-EA evolutionary approach. One of the strengths of this approach is to quickly propose a Pareto front near the optimum. To answer the second point, an improvement of this method is proposed and evaluated. The idea is, from a first approximate Pareto front, to ask the user to choose an area of interest and to restrict the search for solutions only on this area. This improvement results in significant computing time savings.
Document type :
Complete list of metadatas

Cited literature [142 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Tuesday, December 17, 2019 - 5:33:07 PM
Last modification on : Wednesday, June 24, 2020 - 4:18:34 PM
Long-term archiving on: : Wednesday, March 18, 2020 - 7:55:41 PM


Version validated by the jury (STAR)


  • HAL Id : tel-02416606, version 1



Luis Garcés Monge. Knowledge-based configuration : a contribution to generic modeling, evaluation and evolutionary optimization. Other [cs.OH]. Ecole des Mines d'Albi-Carmaux, 2019. English. ⟨NNT : 2019EMAC0003⟩. ⟨tel-02416606⟩



Record views


Files downloads