Scholarlink

In this chapter, we propose incorporating the concept of “super-agentsSuper-Agent (SA)” as efficient and versatile techniques to enhance the agent-in-cell (AIC) model for pandemicsPandemics, the agent-based modelAgent Based Modeling (ABM) (ABM) introduced in the Agent-in-Cell Modeling of PandemicsPandemics in Chap. 5 in The COVID Information Commons—Research Insights from the Coronavirus. These super-agentsSuper-Agent (SA) enable the efficient simulation of more complex infection dynamics in urban environments, while reducing computational complexity, all while maintaining the granularity of individual-level interactions. We extend the AIC model by employing geospatial tessellation, specifically through the construction of Voronoi diagrams based on street network locations. Our approach demonstrates that flexibility inTessellations tessellations—particularly the introduction of different tessellationsTessellations for various types of points-of-interest (POI) visited by agents—achieves a better balance between accuracy and computational efficiency compared to standard tessellationsTessellations based on Census Block Groups (CBGs). Special attention is given to calibrating AIC using super-agentsSuper-Agent (SA), validating coarser tessellationsTessellations against their higher-resolution counterparts (considered as ground truth), and benchmarking the model against an existing open-source ABM. This flexible, hybrid approach optimizes both precision and performance in modeling infection spread in urban settings.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Agent-In-Cell Modeling of Pandemics: Harnessing Super-Agents for Predictive Modeling

  • 章节

  • 发表日期：

    2026年

查看全文
  • Amir Mohammad Esmaieeli Sikaroudi,
  • Alon Efrat,
  • Michael Chertkov

摘要

In this chapter, we propose incorporating the concept of “super-agentsSuper-Agent (SA)” as efficient and versatile techniques to enhance the agent-in-cell (AIC) model for pandemicsPandemics, the agent-based modelAgent Based Modeling (ABM) (ABM) introduced in the Agent-in-Cell Modeling of PandemicsPandemics in Chap. 5 in The COVID Information Commons—Research Insights from the Coronavirus. These super-agentsSuper-Agent (SA) enable the efficient simulation of more complex infection dynamics in urban environments, while reducing computational complexity, all while maintaining the granularity of individual-level interactions. We extend the AIC model by employing geospatial tessellation, specifically through the construction of Voronoi diagrams based on street network locations. Our approach demonstrates that flexibility inTessellations tessellations—particularly the introduction of different tessellationsTessellations for various types of points-of-interest (POI) visited by agents—achieves a better balance between accuracy and computational efficiency compared to standard tessellationsTessellations based on Census Block Groups (CBGs). Special attention is given to calibrating AIC using super-agentsSuper-Agent (SA), validating coarser tessellationsTessellations against their higher-resolution counterparts (considered as ground truth), and benchmarking the model against an existing open-source ABM. This flexible, hybrid approach optimizes both precision and performance in modeling infection spread in urban settings.