Problem. Housing is a major contributor to Switzerland's carbon footprint and energy consumption, but it is also a basic need. Research on climate change mitigation strategies has so far paid insufficient attention to households' preferences and their contribution to housing sustainability. Depicting residential preferences requires an understanding of the multilevel, context-specific, and interrelated determinants of the match between households and dwellings, which are made explicit in the residential mobility process. Goal. The goal of this thesis is to clarify the determinants of residential mobility and their interrelations to illuminate the effects of measures aimed at reducing housing's environmental footprint while meeting the needs of its households. Methods. This research employs a systems and transdisciplinary approach to explore, dig into and disentangle the complex relations governing the residential mobility process. It starts by conceptualising residential preferences and dwelling forms as the manifestations of the housing human and material subsystems, which are determined by the system's function. The applicability of this framework to residential mobility is explored by sequentially gathering and integrating qualitative and quantitative data on the tenants of three Swiss multifamily housing owners. This empirical basis is used to develop an agent-based model (ABM) of households' residential dynamics for the exploration of the emergent effects of housing sustainability measures. Results. Firstly, the present work shows that several functions coexist in the housing realm, each of which determines, for given elements of the societal and environmental supersystems (e.g. culture, location), different residential preferences and dwelling forms. Secondly, empirical investigations demonstrate that the notion of function offers analytical support to the study of Swiss tenants' residential mobility, as it permits the identification of linkages between the qualities of residential environments and residents' characteristics, their residential satisfaction, and triggers to move. In particular, functions help explain the observed preference for larger dwellings and the low propensity of shrinking households to relocate. Furthermore, this notion enables the observation of the effects of the exceptional circumstances posed by the COVID-19 on residential preferences, and the increased desire for a place for 'self-representation'. Lastly, using this empirical and context-specific knowledge, the thesis introduces the agent-based model ReMoTe-S. Simulation experiments show that measures to reduce space consumption are insufficient if not accompanied by an offer of dwellings that simultaneously meet and reshape households' preferences. They also reveal that these preferences cannot be fulfilled via the direct correspondence between desired and current housing characteristics, as housing functions result from several compromises between them. These results invite practitioners and citizens to collectively reimagine the way housing functions express themselves in dwelling forms with a lower environmental footprint. Conclusions. This thesis offers a platform for systemic co-inquiry that enables researchers and practitioners to navigate the complex interactions at play for the provision of sustainable housing. To limit human-induced global warming, the approaches and tools used must rapidly permeate practice and teaching in architecture.