 |
|
Theses
Gestion de la complexité dans un logiciel destiné à la simulation numérique multiphysique
Abstract : The development of numerical simulation software, particularly those involving the résolution of multiphysics problems, implies a significant increase in the complexity of the data structures progressively with the enrichments. The control of this complexity requires a well-organized and extensible structure. These successive enrichments introduce more features interesting for the end-user but, vert/ often, at the price of a rising difficulty in use.
This thesis tackles the problems of managing the structure and the complexity in a multiphysics numerical simulation application, in applying the combination of the object-oriented programming and business rules with the help of Prolog, a logic programming language.
We start with the realization a data model containing abstract classes which facilitate the héritage and the expansion of this data structure. The goal is to permit the developers to add a new data model or to modify an existing model, without calling into question the existing ones. Such a data model requires a checking engine which ensures that all the orders and constructions of objects are valid. We propose to carry, out these checks using a rules interpréter and not in an algorithmic way as actually dope. This rudes interpréter, based on logic programming language, gives the possibility of carrying out not only the checking, of course, but also the proposition and the explanation.
This approach makes it possible on the one hand to reduce the task of the developer and on the other hand to facilitate the opérations of the end-user.
This thesis tackles the problems of managing the structure and the complexity in a multiphysics numerical simulation application, in applying the combination of the object-oriented programming and business rules with the help of Prolog, a logic programming language.
We start with the realization a data model containing abstract classes which facilitate the héritage and the expansion of this data structure. The goal is to permit the developers to add a new data model or to modify an existing model, without calling into question the existing ones. Such a data model requires a checking engine which ensures that all the orders and constructions of objects are valid. We propose to carry, out these checks using a rules interpréter and not in an algorithmic way as actually dope. This rudes interpréter, based on logic programming language, gives the possibility of carrying out not only the checking, of course, but also the proposition and the explanation.
This approach makes it possible on the one hand to reduce the task of the developer and on the other hand to facilitate the opérations of the end-user.
Cited literature [30 references]
https://tel.archives-ouvertes.fr/tel-00163799
Contributor : Sylvie Garcia <>
Submitted on : Wednesday, July 18, 2007 - 3:33:25 PM
Last modification on : Thursday, November 19, 2020 - 1:00:25 PM
Long-term archiving on: : Thursday, April 8, 2010 - 7:29:24 PM
Contributor : Sylvie Garcia <>
Submitted on : Wednesday, July 18, 2007 - 3:33:25 PM
Last modification on : Thursday, November 19, 2020 - 1:00:25 PM
Long-term archiving on: : Thursday, April 8, 2010 - 7:29:24 PM