This paper is concerned with methods for analysing patterns of conflict. We survey
dynamic games, differential games, and simulation as alternative ways of extending
the standard static economic model of conflict to study patterns of conflict dynamics,
giving examples for each type of model.
It turns out that computational requirements and theoretical difficulties impose
tight limits on what can be achieved using the first two approaches. In particular, we
appear to be forced to model the outcome of conflict as being decided in a single
final confrontation if we employ non-linear contest success functions.
A simulation study based on a new model of adaptive, boundedly rational decision
making, however, is shown not to be subject to this limitation. Plausible patterns of
conflict dynamics emerge, which we can link to both historical conflict and standard
tenets of military theory.
