Swarm intelligence has proven particularly useful over the past two decades in identifying more efficient methods for computer engineering, machine learning and describing social media phenomena. Several years ago at a FPRI conference on al Qaeda foreign fighters I noted that I thought Ant Colony Optimization, an efficiency method by which ants find their food, might be an effective modeling system for analyzing, understanding and ultimately disrupting foreign fighter recruitment pipelines to places like Iraq and Afghanistan. Today, I believe it is time to re-examine the utility of this methodology to help analyze and potentially mitigate the largest foreign fighter mobilization since the original Afghan jihad during the 1980s -– the Syrian revolution.
Ant Colony Optimization, often referred to as ACO, explains the method by which ants efficiently find their food. Mauricio Perretto and Hector Silverio Lopes in their article “Reconstruction of phylogenetic trees using the ant colony optimization paradigm” explain:
“Real ants, when searching for food, can find such resources without visual feedback (they are practically blind), and they can adapt to changes in the environment, optimizing the path between the nest and the food source. This fact is the result of stigmergy, which involves positive feedback, given by the continuous deposit of a chemical substance, known as pheromone.”
In their Figure 2 (seen below), Perretto and Lopes diagram how ant pheromone trails that evaporate over time help guide ants along the most efficient route between their nest and a food source. Ants initially start off traveling in random directions laying pheromone trails that evaporate over time. Other ants follow these pheromone trails. When food is found, ants then travel back to the nest again laying a pheromone trail over their original path. The most efficient path to the food thus becomes overlaid with denser layers of pheromones. Ants sense the more densely laid and efficient trails and select them over the less dense and evaporating paths to food sources. This process results in the most efficient path being reinforced by the subsequent pheromone trails of followers who confirm the route from nest to food.
Further research on ACO has also shown that ants not only optimize their routes but select the best food source available when they are presented routes to different food sources. When presented with either less abundant but more nutritious honey and more abundant but less nutritious sugar, ants can optimize their routes to show a preference for the more nutritious food source of honey.
How do foreign fighters find their way to jihadi battlefields and what does that have to do with ants?
At first glance, it’s seemingly amazing that jihadi foreign fighters can mobilize volunteers across the globe. As noted in books such as Lawrence Wright’s The Looming Tower, foreign fighter recruits from as far away as Kansas made their way to the Afghan battlefields of the 1980s simply through “advertisement” flyers and audio broadcasts. However, with the expansive use of the Internet by jihadists, the recruitment process to battlefields like Iraq, Afghanistan after September 11, 2001 and now Syria has occurred more quickly and transparently as compared to the original Afghan jihad of the 1980s. Today, while the networks supporting jihadi recruitment to battlefields remain a bit opaque, the migration of fighters and their activities in jihadi campaigns are quite visible for all to see on social media. After watching these jihadi migrations, I believe that ACO models can be used to understand how and where foreign fighter pipelines will flourish in support of extremist conflicts.
Here is an example.
Much like ants seeking food, foreign fighter recruits often set out blindly seeking a successful path to a jihadi campaign. The initial foreign fighter recruits to a battlefield like Syria likely set out in small numbers of five or less with the most zealous Western recruits setting out on solo missions. These initial recruits test a variety of different routes and methods en route to places like Syria. Many recruits from North Africa and the Middle East travel old networks utilized during previous jihadi conflicts while other recruits blindly seek out their own path. During these travels to join jihadi groups, foreign fighters — eager to brag about their accomplishments — communicate back to peers at home via email, phone, social media and jihadi web forums. Instead of excreting pheromones like ants, foreign fighters lay digital trails marking their routes — virtual breadcrumbs. With successful migration and integration into places like Syria, foreign fighter recruits broadcast their jihadi adventures. Those recruits that broadcast most frequently and demonstrate the most successful route to Syria further encourage fence-sitters at home that jihadi dreams can be fulfilled creating an exponential recruitment pace — the Minneapolis, Minnesota recruitment of several Somali Americans to al-Shabaab around the 2008 time period may be one example. Meanwhile, those recruits that set out for jihad and are heard from less frequently and appear to take a longer time or fail in reaching their destination likely dissuade other foreign fighters at home from following their path — an example of stunted recruitment might be the failed attempts of five Americans from Alexandria, Virginia that traveled to Pakistan in 2009.
Like ant pheromones, foreign fighter digital signals fade over time as their relevance can be crowded out by the emergence of new alternative jihadi campaigns and the negative experiences of foreign fighters treated poorly while fulfilling their jihadi fantasy. For example, fickle foreign fighters prefer to join campaigns on the rise rather than those in decline. This past winter, jihadi Internet forums briefly lit up with calls to support al-Qaeda in the Islamic Maghreb (AQIM) and affiliated groups in repelling the French intervention into Mali. However, AQIM and its affiliates were quickly on the run at a time when the Syrian revolution continued to pick up steam. Seeing a better opportunity in Syria, like ants preferring honey to sugar, foreign fighters likely followed the well documented digital trail to join groups like al-Nusra.
In another context, the negative experiences of foreign fighters also likely influence the choice of emerging recruits. I would estimate the trials and tribulations of Omar Hammami, the American foreign fighter to al-Shabaab who has since been betrayed and hunted by his terror group, have stunted foreign fighter flow into Somalia. Similarly, in the Sinjar records, foreign fighters who traveled to Iraq through the coordination of a facilitator named “Loua’aie” noted they found difficulties dealing with him. In both cases, these negative experiences demonstrate how a digital foreign fighter pheromone trail can evaporate and eliminate less than optimal facilitation routes.
What could be accomplished by using ACO to model foreign fighter flows?
ACO modeling, if successful in its application to foreign fighter flows, could yield several benefits. First, this modeling could be used in early warning to detect when jihadi campaigns have reached a point where recruitment and facilitation of extremists have broken a critical threshold suggesting future strategic implications –- like when the Syrian revolution went from a handful of foreign fighters from neighboring countries to a global cause bringing in thousands of fighters from dozens of countries. Second, ACO may be able to quickly identify which routes have become the most optimal for foreign fighters and help counterterrorism strategists focus their interdiction and disruption efforts. Third, ACO would ideally determine those attributes of highly efficient foreign fighter recruitment pipelines. Knowing what facilitation attributes result in more efficient recruitment and facilitation will assist counterterrorism efforts by identifying key focal points that both signal the formation of efficient routes and become decisive points for tackling terror networks.