What do the Southwest Airlines boarding process and the video game Halo have in common?
They both rely on swarm intelligence to improve their experience.
Swarm intelligence describes the behavior of a population of simple agents whose aggregate behavior exhibits intelligence unknown to the individual agents. Groups exhibiting swarm intelligence have no central leader but rather members interact with each other based solely on information they have locally. Examples in nature include ant colonies, flocks of birds, schools of fish, and bacterial growth.
“Ants are not smart. But colonies are smart. So what’s amazing about ants is that in the aggregate, all of these inept creatures accomplish amazing feats as colonies. In an ant colony, there’s nobody in charge. There are no managers. There is nobody telling anybody what to do. The queen does not give rules. She just sits there and lays eggs.”
Similar behavior exists in herds of caribou that migrate across the Arctic coastal plan to a specific calving ground even though it is unlikely any of the individual animals know where they are going. Or in a vast school of fish that can simultaneously react to a prey and collectively change direction in an instant – seemingly as if it were a single fish. The intelligence of the swarm appears greater than the sum of the members.
According to the book Smart Swarm, Southwest Airlines used a simulation based on swarm intelligence to determine whether its open seating policy was more efficient than other airlines’ assigned seating. The digital ants were given only a few simple rules: find an open seat, wait if the path was blocked, ask other ants to move if they were in the way, etc. The resulting simulation showed that open and assigned seating were roughly equally as efficient, convincing Southwest to keep its longstanding policy. Instead, Southwest added an optional early bird fee which allows travelers to check in earlier to get a better place in line.
Recently I had the chance to describe swarm intelligence and the Southwest Airlines boarding process to an avid video gamer I met. He countered that the popular video game Halo likely uses swarm intelligence as well. From what I understood, there is a species in Halo known as the Hunters which are made up of thousands of tiny, orange worm-like creatures. On their own, the individual creatures can’t do much but when banded together they dramatically increase their strength, intelligence, and maneuverability. True to the concept of swarm intelligence, the Hunters have no leaders which make them very difficult to deal with.
Early in my career I worked at a company called MasPar which relied on thousands of processors to conduct massive amounts of coordinated computations in parallel. Recently the concept of massively parallel processing has regained popularity as relatively inexpensive computers and large of amounts of memory can be combined to solve problems not previously able to be tackled. SAP HANA is an example of such a system which has been used to reduce traffic jams, analyze cancer genomes, and better model energy demands.
I’m personally looking forward to seeing SAP HANA used to create swarm forecasting models. These models will support better predictive scientific models and allow proactive guidance for business processes. We may finally stop just looking in the rear view mirror.
Follow me on twitter @jbecher
This blog was originally posted on Manage By Walking Around.