As mobile devices are more and more equipped with different wireless radio interfaces, multi-homing becomes feasible for the sake of better quality of experience and robustness. From a networking perspective, multi-homing is challenging due to the heterogeneity of the accessed wireless networks. MOMO proposes the use of multi-overlays as a network virtualization technique to integrate and control different QoS, mobility management, pricing, privacy and security techniques of the involved wireless networks. Hereby, MOMO will provide a set of models and methods: (i) a context model relevant for wireless networking togehter with a mapping function to assign overlays to particular multivariate context data, (ii) analytical models for creating an appropriate number of multi-overlays considering the trade-off between perfect fit and overhead for overlay creation and management, (iv) an appropriate search algorithm for finding the best fitting overlay either based on centralized or based on decentralized approaches (or hybrid forms), and (iv) efficient adaptation and maintenance routines of multi-overlays. The models provided by MOMO will be targeted to the use case of wireless mesh networks.