Evaluating the complexity of UML class diagrams

Abstract

One of the main purposes of software engineering is to improve the quality of software products. As we all know if we want to have good and acceptable software in quality criteria we must spot it from the early phases of the development life cycle. One of the key artifacts in theoretical modeling phase is class diagrams in which their quality has an important impact on the quality of our system. If they have low quality they will lead to so many problems, for instance, the construction cost will be more than the estimated one, so if we measure the quality of class diagrams we can find out if they have low or high quality and then we try to eliminate the problem in those diagrams. One of the ways for evaluating the quality of UML class diagrams is to measure the complexity of those classes. In this thesis we tried to recommend and present a way of measuring the complexity of class diagrams with respect to the complexity metrics based on the relationship between classes. This method of measuring has many good assets and can measure the complexity of each class diagram independently by using the tool based on fuzzy logics.