Complex business processes are usually modeled and implemented by workflows or composite services (we call them entities) which contribute to a common goal. To achieve that, the different entities need to collaborate in the sense that the
execution of one entity may have certain dependencies on other entities. A dependency describes a relationship between entities, whereas an entity can be for instance a web service in a service composition or an activity in a workflow. Each dependency has specific features (e.g., uni- or bidirectional, inverse, disjoint) which describe the characteristics of the relationship between the entities or the relation to other dependencies.
To handle dependencies, it is necessary to have this knowledge available in explicit form, i.e. represented by a dependency model (DM). A DM captures important knowledge about dependencies between the entities of a business process. This paper presents and compares two approaches for modeling dependencies as a base for managing adaptations of complex business processes.