Our future environment will be managed by a multitude of different pervasive systems. A pervasive system consists of users and devices which cooperate to provide functionality to the users. The provision of functionality is realized by pervasive applications. A major characteristic of pervasive applications is their context-interactivity. On one hand, pervasive applications are context-aware and can adapt themselves to changing context. This ability enables them to provide their functionality in different configurations. On the other hand, pervasive applications have the ability to influence and change the context themselves. A context change can be caused implicitly as a side effect of employed resources or explicitly through the use of actuators. Due to the context-interactivity, problems are likely to occur when two or more applications are executed in the same physical space. Since applications share a common context and interact with it, they can have a direct impact on each other.The described problem is defined as an interference in this thesis. An interference is an application-produced context that impairs the functionality provision of another application.To manage interferences in pervasive systems, a coordination framework is presented. The framework detects interferences using a context model and information about how applications interact with the shared context. The resolution of an interference is achieved through a coordinated application adaptation. The idea is based on the assumption that an alternative application configuration may yield a different context interaction. Thus, the framework determines a configuration for each application such that the context interactions do not interfere. Once a configuration is found for each application, the framework instructs applications to instantiate the selected configuration, resolving the interference.The framework is unique due to three design decisions. At first, the framework is realized as a cross-system coordination layer in order to allow an integration of arbitrary systems. Secondly, the integration of applications can be achieved through the extension of existing systems while iii iv Abstract preserving their system characteristics. Thirdly, the framework supports a generic interface to integrate arbitrary resolution strategies in order to allow the customization of the framework to the needs of different pervasive systems. The thesis introduces the theoretical concepts of the framework, presents a prototypical implementation and evaluates the prototype and its implemented concepts through extensive measurements.