A Multi-Algorithm Framework for Multi-Scenario Simulation of Urban Built-Up Expansion in ASEAN Under Shared Socioeconomic Pathways

This study develops an integrated multi-algorithm spatial simulation framework for large-scale urban expansion analysis by combining a Back Propagation Artificial Neural Network (BP-ANN), the Future Land Use Simulation (FLUS) model, Cellular Automata (CA), the standard deviational ellipse, Neighborhood Aggregation Analysis, and landscape metrics within a unified data flow. Using the 2020 land use pattern as the baseline, BP-ANN estimates land use suitability probabilities from multiple driving factors; FLUS-CA allocates land under SSP1–SSP5 through adaptive inertia, neighborhood effects, and conversion costs; and spatial statistics and landscape metrics diagnose expansion magnitude, local aggregation, direction, built-up morphology, and cross-country differences. Urban built-up areas expand continuously under all scenarios, increasing by 38.92–49.33% by 2050 relative to 2020. Areas of high local aggregation intensity exhibit pronounced spatial differentiation, while the regional pattern maintains a northwest–southeast orientation, with scenario-dependent differences in centroid location and spatial dispersion. Edge density generally declines as built-up land increases, indicating greater built-up contiguity at the regional scale; this pattern should not be interpreted as evidence of reduced ecological landscape fragmentation. New built-up land encroaches on farmland and coastal ecological spaces and displays marked cross-country heterogeneity. The framework connects driving-factor identification, scenario configuration, spatial allocation, and pattern diagnosis, providing methodological support for cross-national multi-scenario simulation, urban planning, ecological conservation, and transboundary spatial collaborative governance in ASEAN.