Combinatorial Optimization for Urban Planning: Strategic Demolition of Abandoned Houses in Baltimore, MD

In 2017, Baltimore City was awarded $75 million earmarked for the targeted demolition of a portion of its 16,000 vacant and abandoned buildings. Selecting an optimal set of demolition targets is difficult given that the cost per demolition is not independent of the overall demolition pattern; like many older cities, Baltimore's abandoned housing includes a large number of attached row houses, which require the construction of retaining walls when a demolished home abuts a non-demolished home. In this paper, we present a method by which planners can use integer linear programming to identify optimal demolition targets for a number of potential objectives. The simplest objective, demolishing the maximum number of houses for a specific budget, is compared with more complex functions that attempt to proxy improved quality of life resulting from the demolitions. The results of different objective functions are then assessed in terms of equity and efficiency using the spatial distribution of proposed targets as a point of comparison.