Flexible Affordable Housing Policy Design using Facility Location Models

Affordable and subsidized housing providers must design and implement housing strategies: where, when and with what types of housing to best meet the needs of low- and moderate-income households for affordable/low-cost permanent shelter. Recent research has proposed a multiobjective integer programming model for this purpose that jointly optimizes measures of net social benefit and equity and addresses limited variations in housing characteristics. In this paper, we extend the affordable housing planning model to better reflect current research and practice in affordable housing and better meet the needs of affordable housing providers. First, we allow for scale impacts in net social benefits and provision costs that result in a discontinuous, piecewise-linear net social benefit objective. Second, we define alternative nonlinear social equity objectives that better capture social concerns associated with increased class and residential integration. Computational results indicate that these affordable housing model variants are computationally tractable and provide substantial planning flexibility in both decision space and objective space. Planning models incorporating equity objectives that minimize the maximum negative impacts on potential destination communities and minimize variance in impacts on labor markets provide the greatest policy insights.