Bi-objective Optimization for Reducing Urban Network Sprawl

The urban population is steadily rising, with city growth and an increase in land use as its corollaries. As temperatures rise due to climate change, there is an urgent need for green cities. Addressing this requires the reallocation of portions of already artificialized soils for the purpose of greening a city. The possibility of repurposing some urban spaces, specifically by reducing street network sprawl, received much attention recently. In this work, the urban street network is modelled as a strongly connected directed graph. The problem can be formulated as a bi-objective optimization problem. Based on a particular property, several algorithms are proposed to reduce the number of arcs in the network while preserving the strong connectivity property. The solutions generated by these algorithms collectively build a Pareto front for a particular city. Using OpenStreetMap data, the proposed algorithms are applied to several French cities. The results demonstrate that a substantial number of symmetric arcs, from 15% to 50% in some cities, can be removed from the street network with a limited increase on the distances between locations in the city, between 5% to 15%. This work constitutes an additional contribution to the methods designed to mitigate the effects of climate change on urban areas.