In spite of the limited intelligence of each individual, ants always seem to find the most effective routes between their anthill and the most rewarding food sources.
Similar intelligence is applied to build efficient internal traffic systems, inside the hill. Large, meticulously well-organized communities, seemingly without any central planning or coordination.
So, what do they know or have that we humans don’t?

FRAGRANCE MARKERS
The secret is a smart biological communications system based on pheromones used as fragrance markers. The strongest fragrance indicates the most frequently travelled road. If and when one individual finds the way, everybody else follows.
AI professionals often use that very example to describe “swarm intelligence”, as opposed to a single brain controlling it all.
In a recent blog post, automation expert Pascal Bornet listed several other examples of swarm intelligence applications, including:
- Autonomous cars being alerted by other vehicles ahead of yours, adjusting to the traffic flow and ensuring a smooth traffic flow.
- Reporting of available parking slots, making better use of available space.
- Identifying not-yet-visible traffic lights ahead, warning about roadwork, icy roads, and so forth.
And that is only the beginning.

SWARM INTELLIGENCE AND IoT
When we think of tomorrow’s smart cities it’s not mainly about new cool cars and other vehicles. The real issue is the safe and efficient management of traffic, also three-dimensional to include drones and other connected airborne vehicles. Not to mention intersections, street lighting, and all kinds of services.
Intimately related to Artificial Intelligence (AI), Swarm Intelligence (SI) is widely used in conjunction with Internet of Things (IoT) to solve a wide range of challenges in complex, decentralized systems. Safe and efficient city traffic is only one example.
Scientists use the Ant Colony Optimization (ACO) algorithm to address complex global challenges. Similarly, Artificial Bee Colony (ABC) is a stochastic search technique that (among other things) can be used as a “fitness evaluator”. In one study it was applied to analyze the behavior of individual bees and categorize their behavior as “employed”, “onlookers”, or “scouts”.
The results of such studies have also been applied to solve a range of numerical and human-world problems related to, for example, machine learning for optimizing healthcare resources, training neural networks and signal processing applications.