The Goal-oriented Requirement Language (GRL), part of URN, is supported by jUCMNav for modeling goals and other intentional concepts. Currently, two major categories of evaluation approaches, qualitative and quantitative, exist and are supported by this tool. We focus on the quantitative evaluation approach in this project, i.e., contribution, satisfaction, and importance levels will be represented with integer values. GRL model evaluation in jUCMNav is currently using a bottom-up approach that allows users to ask "What if?" questions. The demand for a top-down or a more generic algorithm that supports ad-hoc queries in GRL models is growing, as these are necessary to answer “Is it possible?" and other optimization questions.
A generic quantitative propagation algorithm is presented in this project. The implementation of the algorithm uses JaCoP, a constraint solver library written in Java, to compute a solution to a set of equations obtained from a given GRL model. The supported type of links between the intentional elements are AND decompositions, OR decompositions, dependencies, and contributions. This project even enables circular dependency in jUCMNav and is a much more generic algorithm compared to bottom-up and even top-down approaches as it allows for GRL intentional elements to be initialized anywhere in the model (not just roots and leaves).
- 24 May 2011
- This is a Masters project report (2 semesters), not a thesis per se.
- Please feel free to discuss this article directly on this page. Constructive comments are welcomed! Please sign your TWiki name.
| Title || Generic Propagation Algorithm for Goal Models |
| Authors || Hao Luo |
| Type || Thesis |
| Conference/Journal Title || |
| Volume/Number || |
| Editors || |
| Publisher || SITE, University of Ottawa |
| Month || May |
| Year || 2011 |
| Pages || 34 |
| DOI || |
| Keywords || GRL semantics, propagation algorithms, constraint-oriented solving, JaCOP, jUCMNav |