{"title":"Rethinking the Discrete-Continuous Choice Problem by Uncovering Asymmetry in High-Resolution Urban Modeling","authors":"Jiaxin Zhang, Eric J. Miller","doi":"10.1111/gean.70065","DOIUrl":null,"url":null,"abstract":"<p>A primary bottleneck in integrated urban simulation is the Discrete-Continuous Choice problem within Land Use Supply modeling, specifically, the challenge of simultaneously determining building type and floorspace capacity. This study develops a high-resolution, parcel-level, multi-stage econometric framework in which the bidirectional interactions between these decisions can be estimated rather than assumed. Using a large-scale dataset assembled for the Greater Toronto-Hamilton Area from 2015 to 2023, the framework incorporates real-world zoning constraints to simulate market behavior. This granular approach also uncovers the behavioral heterogeneity across retail, office, industrial, and mixed-use developers through their unique spatial preferences and patterns of inter-sectoral competition. The results reveal an asymmetric relationship between the two decisions. Expected floorspace significantly affects the choice of development type, but only when specified with alternative-specific parameters. In contrast, the reverse effect is not statistically significant in the floorspace models. Although this asymmetry may be specific to the sectors, region, and period examined, it suggests that bidirectional dependence should not be assumed by default. When the Discrete-to-Continuous relationship is weak or not statistically supported, a simpler sequential specification may offer lower modeling complexity. The proposed framework provides a direct way to examine which structure is more appropriate for a given application.</p>","PeriodicalId":12533,"journal":{"name":"Geographical Analysis","volume":"59 1","pages":""},"PeriodicalIF":2.4000,"publicationDate":"2026-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1111/gean.70065","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Geographical Analysis","FirstCategoryId":"89","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1111/gean.70065","RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"GEOGRAPHY","Score":null,"Total":0}
引用次数: 0
Abstract
A primary bottleneck in integrated urban simulation is the Discrete-Continuous Choice problem within Land Use Supply modeling, specifically, the challenge of simultaneously determining building type and floorspace capacity. This study develops a high-resolution, parcel-level, multi-stage econometric framework in which the bidirectional interactions between these decisions can be estimated rather than assumed. Using a large-scale dataset assembled for the Greater Toronto-Hamilton Area from 2015 to 2023, the framework incorporates real-world zoning constraints to simulate market behavior. This granular approach also uncovers the behavioral heterogeneity across retail, office, industrial, and mixed-use developers through their unique spatial preferences and patterns of inter-sectoral competition. The results reveal an asymmetric relationship between the two decisions. Expected floorspace significantly affects the choice of development type, but only when specified with alternative-specific parameters. In contrast, the reverse effect is not statistically significant in the floorspace models. Although this asymmetry may be specific to the sectors, region, and period examined, it suggests that bidirectional dependence should not be assumed by default. When the Discrete-to-Continuous relationship is weak or not statistically supported, a simpler sequential specification may offer lower modeling complexity. The proposed framework provides a direct way to examine which structure is more appropriate for a given application.
期刊介绍:
First in its specialty area and one of the most frequently cited publications in geography, Geographical Analysis has, since 1969, presented significant advances in geographical theory, model building, and quantitative methods to geographers and scholars in a wide spectrum of related fields. Traditionally, mathematical and nonmathematical articulations of geographical theory, and statements and discussions of the analytic paradigm are published in the journal. Spatial data analyses and spatial econometrics and statistics are strongly represented.