Since two decades, the developments in the information technologies have changed the computer science paradigms. The arrival of new approaches such as the pervasive computing create new technologies, which improved the quality of the interactions with the I.T. systems, However, the broad utilization of the pervasive computing and the development of smart spaces are facing several challenges. The large number of hardware and software components and their heterogeneous natures, contributes to the complexity of deploying and managing these environments, resulting in high costs. This thesis contributes to the software management of the smart spaces by reducing the complexity of the management tasks. Our proposal consists of an autonomic approach to software management based on the autonomic pervasive computing. As part of this work, we propose a solution using the environment’s contextual information to determine the software distribution among the environment’s devices, which best fits the needs of the software, environments and the interaction modalities and preferences of the environment’s users. The proposed solution was implemented and evaluated using a series of tests and software organization scenarios
