ASCENT: a tensor-oriented approach to population-scale agent-based simulation
4 affiliations
- 1Department of Computer Science, University of Oxford
- 2Department of Pathology, Microbiology and Immunology, University of California
- 3School of Geography and the Environment, University of Oxford
- 4Institute for New Economic Thinking at the Oxford Martin School, University of Oxford
Abstract
Agent-based models (ABMs) offer a compelling framework for simulating complex systems but scale poorly with conventional software execution paradigms. When modelling millions of interacting agents with complex lifecycles and spatial dynamics, conventional ABMs encounter severe computational bottlenecks, e.g., from memory fragmentation and loop-based state updates. We introduce ASCENT (agent-based simulation using computationally efficient nested tensors), a tensor-oriented approach that reformulates agent state, interactions, and transitions as tensor programs, which enables population-level vectorised computation on commodity GPU hardware. ASCENT supports heterogeneous agent attributes, stochastic state transitions, spatially explicit movement, and machine-learning–driven behaviours, while substantially reducing the traditional trade-off between expressiveness and computational scale. Applied to modelling of gene drive mosquito dynamics on Príncipe Island, ASCENT can simulate millions of individual agents with full lifecycle, genetic inheritance, and spatial dispersal fidelity, demonstrating that a single strategically placed release achieves 95% allele prevalence within approximately 73 days. The emergent Allee effect threshold at ∼ 300 individuals further validates the model’s ecological fidelity. These results demonstrate that our tensor-oriented reformulation enables population-scale agent-based simulation without sacrificing behavioural richness.
Related
- Irrational health beliefs as barriers to health behaviors: development and validation of a brief scale in a Japanese population
- Population-scale analysis of prophages and defense systems reveals distinct evolutionary routes for virulence and antimicrobial resistance in Staphylococcus aureus
- Size Matters: A Large-Scale, Real-World Dataset of Smartphone-Based Obstructive Sleep Apnea Screening in the General Population
- Comprehensive comparison of six optimization algorithms for automated population pharmacokinetic model selection via two simulation studies and eleven real case studies
- Neural shape models enable cohort-scale simulation of knee cartilage mechanics
BibTeX
@article{mahmood2026ascent,
title = {{ASCENT: a tensor-oriented approach to population-scale agent-based simulation}},
author = {Imran Mahmood and Sureni Wickramasooriya and Richard M. Bailey and Gregory C. Lanzaro and Anisoara Calinescu and J. Doyne Farmer and Michael Wooldridge},
journal = {Scientific Reports},
year = {2026},
month = jul,
doi = {10.1038/s41598-026-61495-6},
url = {https://doi.org/10.1038/s41598-026-61495-6},
note = {CC BY},
}
RIS
TY - JOUR
TI - ASCENT: a tensor-oriented approach to population-scale agent-based simulation
AU - Imran Mahmood
AU - Sureni Wickramasooriya
AU - Richard M. Bailey
AU - Gregory C. Lanzaro
AU - Anisoara Calinescu
AU - J. Doyne Farmer
AU - Michael Wooldridge
PY - 2026
DA - 2026/07/01
JO - Scientific Reports
DO - 10.1038/s41598-026-61495-6
UR - https://doi.org/10.1038/s41598-026-61495-6
AB - Agent-based models (ABMs) offer a compelling framework for simulating complex systems but scale poorly with conventional software execution paradigms. When modelling millions of interacting agents with complex lifecycles and spatial dynamics, conventional ABMs encounter severe computational bottlenecks, e.g., from memory fragmentation and loop-based state updates. We introduce ASCENT (agent-based simulation using computationally efficient nested tensors), a tensor-oriented approach that reformulates agent state, interactions, and transitions as tensor programs, which enables population-level vectorised computation on commodity GPU hardware. ASCENT supports heterogeneous agent attributes, stochastic state transitions, spatially explicit movement, and machine-learning–driven behaviours, while substantially reducing the traditional trade-off between expressiveness and computational scale. Applied to modelling of gene drive mosquito dynamics on Príncipe Island, ASCENT can simulate millions of individual agents with full lifecycle, genetic inheritance, and spatial dispersal fidelity, demonstrating that a single strategically placed release achieves 95% allele prevalence within approximately 73 days. The emergent Allee effect threshold at ∼ 300 individuals further validates the model’s ecological fidelity. These results demonstrate that our tensor-oriented reformulation enables population-scale agent-based simulation without sacrificing behavioural richness.
SN - 2045-2322
ER -