Green space terrain design and plant habitat optimization based on genetic algorithms

Objective In response to the current problem of landscape terrain design relying on manual labor, low efficiency, and disconnection from ecological data analysis, this paper aims to develop an automated, data-driven terrain optimization method to accurately improve plant habitats.

Method This study developed an automated digital terrain optimization program based on a genetic algorithm, implemented on the Grasshopper platform. The optimization targeted a 5.57 ha site along the north bank of the Weihe River of northwestern China, focusing on mitigating issues such as arid soil conditions, simplified terrain configurations, and spatial homogenization of plant habitats. Afterwards, SWAT and Fragstats software were used for environmental multi-factor evaluation and landscape pattern index analysis to detect the changes in plant habitat after optimization.

Result The optimization process resulted in an increase in maximum terrain undulation by 0.36 m, a reduction in the average daily duration of direct incoming solar radiation by 0.36 h, and a decrease in global radiation by 11.53 (kW·h)/m². Additionally, average soil moisture content increased by 0.24 mm, while the diversity and homogeneity indices of plant habitats improved by 0.20 and 0.11, respectively.

Conclusion Research findings indicate that the genetic algorithm-based automated digital terrain optimization program can rapidly identify optimal solutions for terrain parameters and achieve key design objectives. These include enriching terrain configurations, reducing direct incoming solar radiation duration and global radiation, enhancing soil moisture content, and improving habitat diversity and homogeneity through targeted terrain interventions. This study contributes to advancing digital intelligence in terrain design and habitat-site design methodologies, enhancing the efficiency, precision and comprehensiveness of habitat optimization efforts.