TY - CONF

T1 - Determining the range of predictions for calibrated agent-based simulation models

AU - Shi, DongFang

AU - Brooks, Roger

PY - 2009

Y1 - 2009

N2 - Agent-based simulation is increasingly used to study systemsin many areas of business and science. Using agentbasedsimulation for prediction could be very valuable.However, these models usually have many parameterswhich are difficult to measure directly leading to uncertaintyas to the best values to use. It is very important toconsider the alternative feasible sets of parameter values,otherwise the modeling results could be misleading. Amethod is described to take account of this when makingpredictions and was applied to an agent-based word ofmouth consumer model. The result was a wide predictionrange indicating the limitations of using such a model forpredictions. The method could also be applied in othersituations.

AB - Agent-based simulation is increasingly used to study systemsin many areas of business and science. Using agentbasedsimulation for prediction could be very valuable.However, these models usually have many parameterswhich are difficult to measure directly leading to uncertaintyas to the best values to use. It is very important toconsider the alternative feasible sets of parameter values,otherwise the modeling results could be misleading. Amethod is described to take account of this when makingpredictions and was applied to an agent-based word ofmouth consumer model. The result was a wide predictionrange indicating the limitations of using such a model forpredictions. The method could also be applied in othersituations.

M3 - Conference paper

T2 - INFORMS Simulation Society Research Workshop Simulation: At the Interface of Modeling and Analysis

Y2 - 25 June 2009 through 27 June 2009

ER -