Strategically robust urban planning? A demonstration of concept

Abstract

Planning for the future is inherently risky. In most systems, exogenous driving forces affect any strategy's performance. Uncertainty about the state of those driving forces requires strategies that perform well in the face of a range of possible, even improbable future conditions. This study formalizes the relationship of different methods proposed in the literature for rigorously exploring possible futures and then develops and applies the computational technique of scenario discovery to the policy option of a subsidy for low-income households in downtown Lisbon. The work demonstrates one way in which urban models can be applied to identify robust urban development strategies. Using the UrbanSim model, we offer the first known example of applying computational scenario-discovery techniques to the urban realm. We construct scenarios from combinations of values for presumed exogenous variablespopulation growth rate, employment growth rate, gas prices, and construction costs -using a Latin Hypercube Sample (LHS) experimental design. We then data mine the resulting alternative futures to identify scenarios in which an example policy fails to achieve its goals. This demonstration of concept aims to lead to new practical application of integrated urban models in a way that quantitatively tests the strategic robustness of urban interventions.

The value of scenario discovery in land-use modeling: An automated vehicle test case

“…This sets the groundwork for adaptive policy approaches (Walker et al, 2013). One hundred scenarios are sampled using the LHS, which is comparable with other scenario discovery experiments in terms of the density of the sample in the multidimensional space (Swartz & Zegras, 2013). 5…”

“…Milkovits et al (2019) developed the Travel Model Improvement Program Exploratory Modeling and Analysis Tool (TMIP-EMAT), a travel demand modeloriented extension of the original Exploratory Modeling and Workbench (Kwakkel, 2017), and deployed TMIP-EMAT using the Greater Buffalo-Niagara Regional Transportation Council regional travel demand. Concerning land use, Swartz and Zegras (2013) provided a demonstration of concept of land-use modeling using UrbanSim to examine future growth in Lisbon, Portugal. This paper is just the second instance, to my knowledge, of exploratory modeling of land-use outcomes, and the first to compare scenario discovery results to more limited sampling approaches.…”

“…I thus examine scenario discovery and robust decision-making, a technique for using simulation models to explore the performance of policy options across uncertain scenarios (Lempert et al, 2006). Scenario discovery has limited deployment in the regional planning context, including a handful of transportation applications (Lempert et al, 2020;Milkovits et al, 2019), and even fewer in Land Use Transportation Integrated (LUTI) modeling (Swartz & Zegras, 2013). At a larger scale scenario discovery has been applied as part of an integrated assessment model that included a cellular automata model land use/land cover change model, a different approach to simulating land-use change than explored in this paper The primary goal of this paper is to demonstrate the application of scenario discovery for land-use modeling to the urban planning community.…”