Modeling Genocide at the System and Agent Levels
Figure 7 . Average population spread versus threshold-level .
These results derive from the fact that the two groups are initialized with equal numbers of agents , and as their Threshold to Act level increases , less killing occurs , and the total populations move exponentially toward the initial condition of equal population counts . While this is in part due to the rule structure of the model , the exponential decline in the spread was unexpected . The implication of this result is that an increasing societal resistance to acting against those with a different identity has an exponentially beneficial effect as that resistance approaches a level at which no violent action is committed against other identity groups . This reinforces the point taken from this experiment ’ s first set of results ; policies geared toward incentives for people not to act against the other have a positive impact potential that increases exponentially as more people choose peace over violence .
SD Model Implementation
Next Modeling Steps
This section provides a brief overview of an implementation framework for a SD stock and flow model that simulates genocide . In brief , stocks are elements in the system that accumulates in time . Flows are the rates of change into and out of a stock . Different stocks may be connected to each other only through their flows . Stocks can be physical items , money , or even emotional states . To implement a running SD simulation , it is necessary to deviate in part from the causal loop diagram shown in Figure 1 . This allows derivation of stocks that are less challenging to implement and validate . Figure 8 shows the simplified causal loop diagram for this model developed using Vensim PLE software ( Version 6.3G ; Ventana Systems , Inc ., 2006 ). Having this diagram is the next step in the progression toward a running SD model . Here , Dominance of the In- Group increases with its percentage of the total population . As the In-Group ’ s
44