An inexact model matching approach and its applications

  • Authors:

  • Hai Zhuge

  • Affiliations:

  • Key Lab of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences, 100080 Beijing, PR China

  • Venue:
  • Journal of Systems and Software
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

Matching between two model specifications is the key step of the repository-based model reuse. From the approximate specification point of view, this paper presents a quantified inexact matching theory for flexible model retrieval from large-scale model repositories. The theory specifies a model repository as two levels: a model level based on a multi-valued model specialization relationship and a fundamental function level based on a function specialization relationship. The matching degree between two models depends on their matching functions. The matching degree between two functions on a function specialization graph depends on the function-distance between them. A set of model specialization rules enables a new matching to be derived from the existing matchings. Embedded in an SQL-like command, the theory has been applied to a large-scale mathematical software model repository system. Users can use the command to retrieve the required models with an inexact query condition. Applications show that the approach is useful and tractable.