This paper presents new mathematical programming-based planning models for the provision of affordable housing to low-income and moderate-income families by government and nongovernmental entities. These models address two key policy concerns of housing providers: setting priorities for investments among a variety of affordable housing programs, and choosing locations and configurations for particular affordable housing initiatives. This paper also incorporates elements of location-design models to address various programmatic and physical attributes associated with affordable housing. Computational results based on a case study demonstrate the potential of these models to generate affordable housing strategies that are flexible, meet affordable housing `gaps', and address social benefits and fairness concerns concurrently.