Optimizing a Multi-Scale Cellular Structure for Urban Growth Modeling Using NSGA-II
Alaeimoghadam, Sanaz ; Karimi, Mohammad LU and Mansourian, Ali LU
(2026)
In Transactions in GIS
30(1).
- Abstract
In urban growth modeling methods, researchers typically determine an appropriate scale based on various factors such as the extent of the study area, the scale of available data, and the level of detail required. With rapid urbanization, cities and towns have grown dramatically, leading to the emergence of numerous megalopolitan areas worldwide. In these areas, which include small and large cities, slums, and rural regions with varying functional scales, it is recommended to apply multiple scales for urban growth modeling. Most existing multi-scale models focus on sensitivity analysis of the scale and pay less attention to scale integration. In this study, a new method is developed to integrate multiple scales for modeling built-up... (More)
In urban growth modeling methods, researchers typically determine an appropriate scale based on various factors such as the extent of the study area, the scale of available data, and the level of detail required. With rapid urbanization, cities and towns have grown dramatically, leading to the emergence of numerous megalopolitan areas worldwide. In these areas, which include small and large cities, slums, and rural regions with varying functional scales, it is recommended to apply multiple scales for urban growth modeling. Most existing multi-scale models focus on sensitivity analysis of the scale and pay less attention to scale integration. In this study, a new method is developed to integrate multiple scales for modeling built-up areas. Urban growth potential is estimated first. Since increasing the scale (by enlarging the pixel size) results in lower resolution but higher accuracy, the simulation accuracy at each position is calculated for each scale and then considering these two opposing objectives—maximizing accuracy and resolution—the NSGA-II algorithm is employed to integrate the scales. The output is an optimized multi-scale cellular structure, which is employed to simulate urban growth. Evaluation results show that the multi-scale cellular structure simulates built-up density with 4.62% higher accuracy compared to a single-scale map, while the map resolution is increased by 9%.
(Less)
- author
- Alaeimoghadam, Sanaz
; Karimi, Mohammad
LU
and Mansourian, Ali
LU
- organization
- publishing date
- 2026-02
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- built-up density, GIS, multi-objective optimization, multi-scale cellular structure, urban growth modeling, Geospatial Artificial Intelligence (GeoAI)
- in
- Transactions in GIS
- volume
- 30
- issue
- 1
- article number
- e70182
- publisher
- Wiley-Blackwell
- external identifiers
-
- scopus:105030124051
- ISSN
- 1361-1682
- DOI
- 10.1111/tgis.70182
- language
- English
- LU publication?
- yes
- additional info
- Publisher Copyright: © 2026 John Wiley & Sons Ltd.
- id
- 1d52a836-1148-4aa6-ba5f-22b389f4bfb8
- date added to LUP
- 2026-02-27 12:02:11
- date last changed
- 2026-03-13 13:22:41
@article{1d52a836-1148-4aa6-ba5f-22b389f4bfb8,
abstract = {{<p>In urban growth modeling methods, researchers typically determine an appropriate scale based on various factors such as the extent of the study area, the scale of available data, and the level of detail required. With rapid urbanization, cities and towns have grown dramatically, leading to the emergence of numerous megalopolitan areas worldwide. In these areas, which include small and large cities, slums, and rural regions with varying functional scales, it is recommended to apply multiple scales for urban growth modeling. Most existing multi-scale models focus on sensitivity analysis of the scale and pay less attention to scale integration. In this study, a new method is developed to integrate multiple scales for modeling built-up areas. Urban growth potential is estimated first. Since increasing the scale (by enlarging the pixel size) results in lower resolution but higher accuracy, the simulation accuracy at each position is calculated for each scale and then considering these two opposing objectives—maximizing accuracy and resolution—the NSGA-II algorithm is employed to integrate the scales. The output is an optimized multi-scale cellular structure, which is employed to simulate urban growth. Evaluation results show that the multi-scale cellular structure simulates built-up density with 4.62% higher accuracy compared to a single-scale map, while the map resolution is increased by 9%.</p>}},
author = {{Alaeimoghadam, Sanaz and Karimi, Mohammad and Mansourian, Ali}},
issn = {{1361-1682}},
keywords = {{built-up density; GIS; multi-objective optimization; multi-scale cellular structure; urban growth modeling; Geospatial Artificial Intelligence (GeoAI)}},
language = {{eng}},
number = {{1}},
publisher = {{Wiley-Blackwell}},
series = {{Transactions in GIS}},
title = {{Optimizing a Multi-Scale Cellular Structure for Urban Growth Modeling Using NSGA-II}},
url = {{http://dx.doi.org/10.1111/tgis.70182}},
doi = {{10.1111/tgis.70182}},
volume = {{30}},
year = {{2026}},
}