A Methodology to Develop Agent-Based Models for Policy Support Via Qualitative Inquiry

V. Nespeca*, M. Comes, F.M. Brazier

*Corresponding author for this work

Research output: Contribution to journalArticleScientificpeer-review

68 Downloads (Pure)

Abstract

Qualitative research is a powerful means to capture human interactions and behavior. Although there are different methodologies to develop models based on qualitative research, a methodology is missing that enables to strike a balance between the comparability across cases provided by methodologies that rely on a common and context-independent framework and the flexibility to study any policy problem provided by methodologies that focus on capturing a case study without relying on a common framework. Additionally, a rigorous methodology is missing that enables the development of both theoretical and empirical models for supporting policy formulation and evaluation with respect to a specific policy problem. In this article, the authors propose a methodology targeting these gaps for ABMs in two stages. First, a novel conceptual framework centered on a particular policy problem is developed based on existing theories and qualitative insights from one or more case studies. Second, empirical or theoretical ABMs are developed based on the conceptual framework and generic models. This methodology is illustrated by an example application for disaster information management in Jakarta, resulting in an empirical descriptive agent-based model.

Original languageEnglish
Article number10
Number of pages40
JournalJournal of Artificial Societies and Social Simulation
Volume26
Issue number1
DOIs
Publication statusPublished - 2023

Keywords

  • Research Design
  • Simulation Methodology
  • Empirical Agent-Based Models
  • Information Diffusion
  • Information Management
  • Crisis Management

Fingerprint

Dive into the research topics of 'A Methodology to Develop Agent-Based Models for Policy Support Via Qualitative Inquiry'. Together they form a unique fingerprint.

Cite this