Researchers have developed a novel online calibration framework for agent-based models (ABMs), a process that is typically computationally intensive. This new method integrates black-box optimization with stochastic batch optimization, a technique previously underutilized in ABM calibration. By treating each simulation iteration as a stochastic batch, the framework allows for incremental parameter updates and the co-evolution of model parameters and agent behaviors within a single simulation run. Tested on the MATSim transport model for Switzerland, the approach successfully aligned mode shares with observed data, demonstrating its efficiency and potential for real-time ABM calibration. AI
IMPACT This new calibration method could accelerate the development and deployment of complex agent-based models in various fields, including transportation and urban planning.
RANK_REASON Academic paper detailing a new methodology for agent-based model calibration. [lever_c_demoted from research: ic=1 ai=0.7]
Read on arXiv cs.MA (Multiagent) →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
