Discrimination Criteria for Modeling Association, Aggregation, and Composition in UML Class Diagrams

Abstract

A class diagram in the Unified Modeling Language (UML) is a powerful tool to model the structure of a system by means of classes, their attributes and operations, the relationships between them, and some additional elements. In this type of models, associations between classes play a crucial role and are determining in the subsequent stages of the development. UML provides three main types of associations for class diagrams in order to distinguish several scenarios: simple associations, shared aggregations, and composite aggregations. Several interpretations of what these three types of relationships mean and how they reflect usual situations in business or software modeling have been proposed, which causes some misunderstandings and makes it difficult to achieve the common interpretation which would be desirable in a unified language. This work intends to extract the key factors which help us decide what type of relationship best describes the domain we try to model. By means of a set of criteria which characterize them, the main differences and similarities are extracted, not only from a semantic point of view, but also considering the implications they have in the structure of the model, and the constraints which appear when using certain additional elements, such as association classes and self-associations. Furthermore, a series of examples covering a wide range of applications is provided to illustrate the distinction and make it possible to extract analogies.