@article{Balhas2026, 
author = {Kassem Balhas and Mohammad Karimi and Parastoo Pilehforooshha},
title = {A new multi-level neighborhood parcel-based cellular automata model for urban land use allocation},
year = {2026},
journal = {Geo-Spatial Information Science},
volume = {29},
number = {4},
pages = {2511-2528},
keywords = {Land use allocation, cellular automata, parcel-based CA, land use interactions, multi-level neighborhood effect},
url = {https://www.sciopen.com/article/10.1080/10095020.2025.2544959},
doi = {10.1080/10095020.2025.2544959},
abstract = {The neighborhood effect is a fundamental component of land use allocation, reflecting the influence of surrounding units on a central unit. Over the past decade, urban land use allocation models have increasingly relied on Cellular Automata (CA) frameworks to simulate these effects. However, the detailed and complex structure of parcel maps poses challenges when employing large neighborhood sizes in CA models. To address this issue, a new Multi-Level Parcel-based CA (MPCA) model is proposed, incorporating a hierarchical neighborhood structure at the parcel, block, and sector levels. This structure generalizes parcels at the block and sector levels while preserving detailed spatial data at the parcel level. The integrated multi-level neighborhood effect is determined in two steps: first, neighborhood effects are calculated independently using distances of 200 m, 1000 m, and the entire study area for the parcel, block, and sector levels, respectively; second, these effects are aggregated to produce a single integrated value for each parcel. For evaluation, the complete land use allocation process is implemented within a structured framework. With an overall accuracy of 0.868, a Kappa index of 0.715, and a Figure of Merit (FoM) of 0.66, the results demonstrate that incorporating multi-level neighborhoods significantly enhances allocation accuracy compared to conventional single-level models. This approach improves the ability to capture broader spatial interactions while maintaining computational efficiency, offering a more robust and realistic solution for urban land use allocation.}
}