Why is stigmergy a good platform for swarm intelligence?

We find universal coordination throughout the natural world. Dynamic environments enable indirect interactions for the production of unified objectives. Let’s explore the four principles of stigmergic collaboration and their application in the design of intelligent systems.

customer interaction
Thinkstock

Have you ever examined the path on which you find yourself standing today along with all the antecedent steps that brought you to the current moment? The modern path of artificial intelligence is a mix of cognitive science, psychology and dreams.

Is there a secret to how organisms collaborate? Methods to coordinate aren’t only found in science labs and the basements of research buildings. Universal coordination mechanisms can be found in many places, if we look. Environmental traces, mass collaboration and group interactions have much in common.

The secret to complete organization

Stigmergy derives from the Greek words στίγμα stigma meaning "mark or sign" and ἔργον ergon meaning "work or action.” Stigmergy is the universal coordination mechanism: a consensus mechanism of indirect coordination within an environment among agents or actions. While we don’t fully understand all the interactions of self-organizing organisms, the concept of self-organization is found in both robotics and social insects.

Remy Chauvin (1956) conducted the earliest work on stigmergy-based coordination in the biological sciences. However, the foundation of stigmergy was envisioned by Pierre-Paul Grasse in 1967, making use of his 1950s research on termites. The idea is that an agent’s actions leave signs in the environment, signs that it and other agents sense and which determine and incite subsequent actions. Stigmergy is also used within artificial intelligence for the study of swarming patterns of independent actors that use signals to communicate.

A better understanding of stigmergy and sociometry (a quantitative method for measuring social relationships) and group dynamics offers new insights into the world of multi-agent coordination, which is the essence of swarm intelligence.

The essence of stigmergy is that traces left within an environment — the result of an action — stimulate the performance of a future action. This combined positive and negative feedback loop enhances a mutual awareness, fostering the coordination of activity without the need for planning, control and communication.

Ants use pheromones. People use wikis. Wasps use secretions. These multi-agent coordination mechanisms function because agents exchange information within a dynamic environment. The agents modify their environment, which triggers a future response.

When open-source systems blossom from five users to 50,000 users, we might find our answers buried in the evolution of group work. Stigmergic collaboration has four distinct principles:

  1. Collaboration depends upon communication, and communication is a network phenomenon.
  2. Collaboration is inherently composed of two primary components — social negotiation and creative output — without either of which collaboration cannot take place.
  3. Collaboration in small groups (roughly 2 to 25) relies upon social negotiation to evolve and guide its process and creative output.
  4. Collaboration in large groups (roughly 26 to n) is enabled by stigmergy.

Collaboration with consensus

Stigmergic collaboration is when agents or individuals work without explicit knowledge of others. Adding a block to a blockchain isn’t controlled by a central function; it’s organic. Editing or changing a wiki page relies on a shared pool of content for mass collaboration and consumption.

Stigmergic interactions are coordinations of activities that, over time, use decentralized control. Primitive rules guide the orchestration of activity. There are no instructions, and there’s a self-awareness for actions and the sharing of information.

How can stigmergic principles be used in your mobile designs? How does the communication and messaging of self-organizing systems improve your IT landscape? How do unstable systems evolve into stable states in which order and organization are the norms? You can apply these theories to your technological environment.

Innovators are using these theories to design interactions that don’t presently exist within conventional artificial intelligence environments. Future artificial intelligence systems will be designed with an awareness of stigmergic collaborations.

This article is published as part of the IDG Contributor Network. Want to Join?

NEW! Download the Winter 2018 digital edition of CIO magazine