Comparison of non solvable problem solving principles issued from CSP and TRIZ

Sébastien Dubois (1) , Ivana Rasovska (1) , Roland de Guio (1)

1 LGeco - Laboratoire de Génie de la Conception

Inventive problems from many domains are usually problems we are not able to solve. This problem insolvability is often due to the incomplete or unmatched representation model of the problem that does not correspond to the given problem. In this paper, we introduce two problem solving theories for the solutionless problems: Constraint Satisfaction Problem (CSP) and dialectical based methods and models (TRIZ). It is an exploratory analysis of both theories in order to compare grounding approach and tools of both theories. Their potential complementarities will be defined in further objective to improve problem solving strategy for the inventive problems by matching the CSP and TRIZ solving principles. We consider that it will contribute to better understanding of non-solvable problems, i.e. to improve representation models of the problems and to make the problem solving more accurately.

Mots clés: Constraint Satisfaction Problem; Over-constrained systems; TRIZ

Domaines: Autre

Sébastien Dubois, Ivana Rasovska, Roland de Guio. Comparison of non solvable problem solving principles issued from CSP and TRIZ. IFIP 20th World Computer Congress (WCC 2008), Sep 2008, Milano, Italy. pp. 83-94; ISBN 978-0-387-09696-4, ⟨10.1007/978-0-387-09697-1_7⟩. ⟨hal-00341031⟩