1. Introduction
Rapid urban expansion onto agricultural lands constitutes a fundamental challenge to land use planning in Iran, particularly where economic growth and food security compete for limited spatial resources. The Isfahan Plain, surrounding the metropolis of Isfahan, exemplifies this persistent conflict. Over several decades, residential development and agricultural production have continually competed for fertile, accessible lands; however, a systematic framework for evaluating and resolving this spatial trade off has not yet been developed. Previous studies have either focused on monitoring past land use changes or assessed the suitability of a single land use type (typically urban development) without treating agricultural land as an independent, quantifiable alternative. Furthermore, although multi objective optimization algorithms such as NSGA II have been applied to land allocation problems, their application in Isfahan has remained confined to intra urban zoning, not to the urban–agricultural conflict at the landscape scale. The present study addresses these gaps by identifying conflict zones and offering Pareto optimal scenarios. By integrating remote sensing, GIS, fuzzy multi criteria decision making, and NSGA II, this research aims to: (1) map current land use and quantify urban and agricultural suitability; (2) derive a spatial conflict index from their overlay; and (3) solve the conflict through a three objective optimization model.

2. Methodology
This applied analytical study was conducted in the Isfahan Plain (50°45′–52°00′ E, 32°15′–33°00′ N), encompassing fertile agricultural lands surrounding the metropolis. The area was selected for its dual pressure: high accessibility to urban infrastructure and highly productive alluvial soils. The methodological framework comprised four sequential steps. First, land use mapping was performed using Landsat 8 and 9 OLI/OLI 2 imagery acquired for summer 2025 (June–August). Pre processing included cloud and shadow masking, followed by temporal median compositing. A supervised maximum likelihood classification distinguished built up areas from active agricultural lands, achieving 96.4% overall accuracy and a Kappa coefficient of 0.93, based on 500 independent validation points. Second, urban development suitability (UDSI) and cropland suitability (CSI) were independently weighted using a fuzzy AHP approach. UDSI criteria included distance from the metropolis, highways, existing urban areas, main roads, and slope. CSI criteria included distance from rivers, groundwater depth, soil organic carbon, slope, soil texture, and drainage. Criteria were standardized using linear fuzzy membership functions or categorical scoring, and weights were derived from pairwise comparison matrices (CR < 0.1). Final suitability maps were produced through weighted linear combination. Third, the spatial conflict index was calculated by cell by cell multiplication of normalized UDSI and CSI layers. The continuous index was classified into three equal frequency classes (low, moderate, high) using the 33rd and 66th percentiles as thresholds. Fourth, the conflict was addressed through a three objective optimization model implemented in R using the mco package for NSGA II. The original 30 m layers were aggregated to 250 m cells. Decision variables were continuous vectors [0,1] for all candidate non urban cells. New urban land demand was set at five percent of the existing urban area. The three objective functions were: (1) maximizing mean urban suitability; (2) minimizing mean cropland suitability lost; and (3) minimizing the count of active agricultural cells converted. The algorithm ran with a population size of 60 and 80 generations, using non dominated sorting and crowding distance. From the Pareto front, three scenarios were extracted: urban priority, agricultural conservation, and balanced.

3. Findings
Land use classification revealed that existing urban areas covered 34,971.63 hectares (349.72 km²), while active agricultural lands encompassed 63,924.24 hectares (639.24 km²) nearly twice the urban extent. High urban suitability zones were predominantly adjacent to the existing urban core and along major transportation corridors, whereas high agricultural suitability zones occupied the eastern and southern portions of the plain. The conflict index ranged from near zero to 0.999, with a mean of 0.41, indicating a widespread structural pattern. Approximately one third of the area (33.4%) fell into the high conflict category (mean 0.760, SD 0.130), primarily located in the peri urban ring around Isfahan. The NSGA II generated a well distributed Pareto front revealing striking asymmetry: the range of urban suitability was only 0.059 (0.796–0.855), whereas the range of agricultural loss was 3.25 (5.31–8.56) more than 50 times greater. This implies that decision makers face choices among markedly different agricultural damage levels for virtually the same urban suitability. Among the three scenarios, the balanced scenario reduced agricultural conversion from 918.72 hectares (urban priority) to 825.03 hectares (a decrease of 93.69 hectares (approximately 10%)) while urban suitability declined only marginally from 0.853 to 0.846 (<1%). The agricultural conservation scenario, despite achieving the lowest suitability loss (7.40), exhibited a larger converted area (849.96 hectares) than the balanced scenario. This paradox indicates that the loss index captures not merely quantity but also intrinsic quality; the conservation scenario sacrificed lower quality land to protect high value cores. Spatial interpretation confirmed that the urban priority scenario concentrated development adjacent to existing fabric, the conservation scenario redirected growth toward low productivity lands, and the balanced scenario integrated both logics.

4. Discussion and Conclusion
The findings reveal that urban–agricultural conflict in the Isfahan Plain is structurally pervasive, particularly across the peri-urban fringe, where shared locational advantages intensify competition for land. The asymmetric Pareto front shows that a balanced strategy sacrifices less than 1% of urban suitability while preserving nearly 94 hectares of active cropland. Importantly, agricultural conservation reduces qualitative losses by protecting high-value soils, demonstrating that land quality matters more than converted area alone. Among the three growth scenarios, the balanced approach offers a pragmatic compromise between spatial cohesion and agricultural protection. The integrated framework provides a robust basis for sustainable urban growth boundaries and is transferable to comparable regions.