Abstract
Recent researchers active in the field of agent-based modeling have called for the alignment, or ‘docking,’ of models which simulate the same system using different techniques. Addressing this need, the present article details a systematic approach for docking models described by (nonlinear) ordinary differential equations with analogous models employing autonomous agents — i.e., agent-based models (ABMs). In particular, the approach is demonstrated by example for an epidemiological SEIR (Susceptible, Exposed, Infectious, Recovered) ODE model with a newly-developed agent-based model. The ABM is designed such that the assumptions present in the ODE model are matched by the actions of the ABM agents and the model. In addition, less-than-transparent coefficients present in the ODE model are examined via difference equations and then mapped to appropriate agent behavior. The result is very good agreement in comparisons made between ODE and ABM model-generated time-histories — i.e., successful alignment. It is anticipated that the systematic alignment approach described herein should be useful for aligning ODE and ABM models in other fields of study — e.g., Lanchester ODE combat models vice ABM combat models.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call