Abstract : One of today's challenges in the framework of ubiquitous computing concerns the design of ambient systems including sensors, smart-phones, interconnected objects, computers, etc. The major difficulty is to propose a compositional adaptation which aims to integrate new features that were not foreseen in the design, remove or exchange entities that are no longer available in a given context. In order to provide help to overcome this difficulty, a new approach based on the definition of strategies validated using discrete-event simulation is proposed. Such strategies make it possible to take into account conflicts and compositional adaptation of components in ambient systems. These are defined and validate using a discrete-event formalism to be integrated into a prototyping and dynamic execution environment for ambient intelligence applications. The proposed solution allows the designers of ambient systems to define the optimum matching of all components to each other. One pedagogical example is presented (switch-lamp system) as a proof of the proposed approach.