Analytic Network Process for model elicitation in nation-building simulations

Discrete-event simulation models are the current main tools for nation-building analysis due to the complexity of the problem. However, it is difficult to study the input-output relations in these models because of the numerous parameters as well as the complicated model structure. Model elicitation can be considered as an approximate study of the significant effects of the model without spending excessive time in multiple replications of discrete event simulation. In this paper, Analytic Network Process (ANP) is applied to model elicitation in large-scale nation-building models. It can be used to study the factor significance and the inter-influence among different kinds of factors. This new approach is strictly based on the quantitative analysis instead of the expert knowledge, which greatly improves the traditional ANP method that uses possibly conflicting human expertise. The numeric results show that this methodology provides good solutions at significantly reduced computational time compared with traditional discrete-event simulations.