The Peace Game: a data-driven evaluation of a software-based model of the effects of modern conflict on populations

Abstract

Over the years wargames have been used by decision makers and operational experts to gain insight into not only how an operation could expect to unfold, but also in unveiling gaps in capabilities of a plan or concept of operation. While most of these games focus on war, the Peace Game focuses on helping planners gain insight as to how Security Force Assistance (SFA) and Security Cooperation (SC) operations may unfold. The Peace Game attempts to model population behavior in a specific region. In the past, movement of the masses has been used as an indicator of regional stability. This thesis concentrates on assessing the migration algorithm within the Peace Game. While the current algorithm does a satisfactory job of simulating migration, it can be made better. The suggested improvements in the Peace Game focus on three models (two types of models). There are two regression models; one that models population deaths over time and another that models displaced persons over time. The third model is roughly based on a compounding interest model. This model is known as the Population Attrition Model (PAM) and is applied to both deaths and displaced persons. All of the models are grounded in an analysis of historical data from over 40 conflicts. All of the models are an improvement in simulating the reality of population migration in regions of conflict.