Logo do repositório
 
Publicação

Consensus among autonomous multi agent systems in public health: a simulation study on decision making under data uncertainty

authorProfile.emailniko.hems@gmail.com
datacite.subject.fosCiências Sociais::Economia e Gestão
dc.contributor.advisorZejnilovic, Leid
dc.contributor.authorHems, Niko
dc.date.accessioned2026-02-20T15:27:01Z
dc.date.available2026-02-20T15:27:01Z
dc.date.issued2025-06-25
dc.date.submitted2025-05-21
dc.description.abstractThis study tested whether autonomous AI agents can agree on public health actions using incomplete or noisy outbreak data. A multi-agent system based on large language models (via Lang Chain) enabled structured, role-based discussions. Simulations showed that while consensus is possible, it’s rare. Agreement was mainly driven by the depth of reasoning—shorter data-grounded arguments led to better outcomes. Surprisingly, missing or noisy data had limited effect on consensus. However, complete data slightly improved results, especially when manipulated. Overall, AI agents can support complex health decisions, but their effectiveness hinges on their ability to reason clearly with available information.eng
dc.identifier.tid204129370
dc.identifier.urihttp://hdl.handle.net/10362/200539
dc.language.isoeng
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/
dc.subjectPublic health
dc.subjectDecision making
dc.subjectAI agents
dc.subjectEpidemiological simulation
dc.subjectCOVID-19 outbreak data
dc.subjectHealth policy modeling
dc.titleConsensus among autonomous multi agent systems in public health: a simulation study on decision making under data uncertaintyeng
dc.typemaster thesis
dspace.entity.typePublication
thesis.degree.nameA Work Project, presented as part of the requirements for the Award of a Master’s degree in Management from the Nova School of Business and Economics

Ficheiros

Principais
A mostrar 1 - 1 de 1
Miniatura indisponível
Nome:
Individual_Research_58116_.pdf
Tamanho:
555.18 KB
Formato:
Adobe Portable Document Format
Licença
A mostrar 1 - 1 de 1
Miniatura indisponível
Nome:
license.txt
Tamanho:
348 B
Formato:
Item-specific license agreed upon to submission
Descrição: