Options
Using structural equation-based metamodeling for agent-based models
Publikationstyp
Conference Paper
Date Issued
2018-01-04
Sprache
English
Author(s)
Institut
TORE-URI
Start Page
1372
End Page
1382
Citation
Proceedings - Winter Simulation Conference : 1372-1382 (2018-01-04)
Contribution to Conference
Publisher DOI
Scopus ID
Publisher
IEEE
ISBN
978-1-5386-3428-8
978-1-5386-3430-1
978-1-5386-3249-5
Trustworthy statistical modeling is an emerging challenge in agent-based modeling (ABM). However, typical characteristics of ABM such as the potential for high numbers of entities and parameters, interdependent relations between entities, several layers of effects, and emergent social phenomena challenge this process. In particular, aggregated outcomes emerging from individual agent interactions are, at least partly, difficult to measure. This might impede the statistical modeling process and thus the formulation of trustable conclusions. For this reason, we introduce structural equation modeling (SEM) as a promising statistical modeling method to analyze the behavior of ABMs. SEM allows for the estimation and evaluation of highly networked systems by explicating interactions between types of agents, measuring emergent phenomena, and identifying output patterns of simulation models. Overall, these contributions foster the credibility and trustworthiness of ABMs, and also support the communication and understanding of simulation models' behavior and their output.
DDC Class
005: Computer Programming, Programs, Data and Security