An agent-based model of counterinsurgency: Soft versus hard approaches

This article uses an agent-based simulation model to investigate the effectiveness of hard (attrition), balanced, and soft (hearts and minds) counterinsurgency approaches. The model involves a combination of irregular geography and social networks with four different types of agents: civilian, military, civil servant, and insurgent. After running 900 simulations (each with up to 3,650 iterations and 5,000 agents) in nine different experimental conditions, the authors reach the following conclusions. First, it is nearly impossible to eliminate an insurgent group if the popular support for the group is high, even when using a ‘hard’ approach. Second, a counterinsurgent government can eliminate an insurgent group only if it uses a ‘hard’ approach and when the popular support for the insurgency is low or medium. Third, a counterinsurgent government can decrease the popular support for the insurgency if it uses a ‘soft’ or a ‘balanced’ approach when the initial support is not already low. Fourth, although it is difficult to eliminate an insurgent group, it is possible to decrease the number of insurgents when the support level is low. The number of insurgents increases when the support level is high, regardless of the counterinsurgency approach.