Objective This study aims to leverage the advancement of computer algorithms to provide scientific analytical methods for spatial simulation and offer a more objective approach to reflecting the natural operational laws of interdependencies between elements in the real physical world for landscape architecture design. Multi-source data, with its high-density distribution characteristics, reflects urban features and spatial patterns at a macro scale. This study employs agent-based modeling to simulate spatial information, exploring the spatial function distribution of planning sites under urban spatial function penetration, thereby guiding the structural morphology of the planning site.

Methods This research adopts a spatial network analysis design method using slime mold intelligence based on point of interest (POI) data mapping. Firstly, POI data and urban basic information data within the Yizhuang New City area of Beijing were obtained through the Gaode Map application programming interface (API) to construct the spatial network foundation of the research area. The POI data were categorized into eight types: residential, hotel accommodation, road traffic, education and culture, medical care, business office, commercial services, and green space tourism. Secondly, a slime mold agent model was constructed using the NetLogo platform, defining the behavioral rules of cytoplasm agents, nuclei agents, and patch agents to simulate the growth behavior of slime molds. By adjusting model parameters, the foraging paths of slime molds in the planning site were simulated, and their spatial function distribution was analyzed. The simulation environment included setting particle growth points, food locations, and environmental obstacles based on multi-source data analysis; six particle growth points were set, representing the main entry points to the site, while food locations were placed at major building entrances and preserved landscape nodes. The environmental obstacles consisted of planned buildings and river channels. Finally, based on multi-source data analysis, particle growth points, food locations, and environmental obstacles were set for spatial function analysis and optimization.

Results The research results indicate that the design method of spatial network analysis using slime mold intelligence based on POI data mapping is highly feasible in the landscape design of small and medium-sized sites. The agent-based model simulation of slime mold growth behavior effectively reflects the infiltration results of surrounding functional information into the site, forming a spatial function zoning map with path generation. The slime mold agent model's path simulation results can provide important references for path connection, functional layout, and crowd-gathering patterns in site design, aligning closely with actual site usage requirements. Specifically, the simulation demonstrated how different functional areas and their respective uses could be distributed across the site, allowing for a more informed and responsive design process. Additionally, the simulation results show that the slime mold agent model can quickly construct foraging networks in complex areas through its efficient organization and structural resilience and optimize path selection according to behavioral objectives.

Conclusion Agent-based model analysis, aided by extensive algorithms, considers the basic spatial information of the site and can simulate individual behavior patterns through distributed computation rules, providing a new paradigm for landscape architecture design thinking. The agent-based model not only effectively simulates the self-organizing behavior of urban spaces but also enhances the match between design and actual site conditions by integrating multi-source data. This method provides a scientific analytical approach for landscape design from macro to micro scales, helping to form spatial structure optimization schemes based on bottom-up design concepts. The combined application of agent-based models (ABM) and multi-source data in the future will contribute to the paradigm shift in urban spatial planning, offering new theoretical and practical support for the design of urban park systems. The application of slime mold intelligence through agent-based modeling provides a robust framework for analyzing and designing urban spaces. This approach allows for the simulation of complex spatial interactions and the emergence of spatial patterns that reflect the dynamic nature of urban environments. By leveraging multi-source data, this method offers a comprehensive tool for landscape architects to design functional, responsive, and sustainable urban spaces. The findings underscore the potential of agent-based models to transform urban landscape design by providing a detailed and data-driven understanding of spatial dynamics and functional distribution, ultimately leading to more informed and effective design decisions. This expanded approach underscores the feasibility of using slime mold intelligence in landscape design, demonstrating its potential to reshape traditional design methodologies. The integration of POI data with agent-based modeling not only provides a high level of accuracy in reflecting real-world conditions but also enhances the adaptability and sustainability of landscape architecture design. As cities continue to evolve and expand, the methodologies explored in this study will become increasingly relevant, providing landscape architects with the tools necessary to create urban environments that are both functional and harmonious with their natural and social contexts. The study paves the way for future research and application in urban planning, emphasizing the importance of data-driven, adaptive design processes in creating resilient and livable urban spaces.