In order for us to carry out in-depth experiments of a multi-agent system in nature we introduce a new classification into the world of distributed artificial intelligence, known as Swarm Systems. Swarm Systems sketch agents within the multi-agent system as individual entities, each with their own behaviour, scalability and flexiblity within the proposed environment. These individuals work concurrently to manipulate the environment according to their needs.
Though the scope of Swarm systems is very broad, it can be narrowed down to two major types: Homogenous and Heterogenous Swarm Systems.
This is a multi-agent system, where all individual agents possess the same level of capability, and all of them carry out similar tasks to achieve one common goal. In the real world, this can be depicted by a closer observation of the way ants work towards gathering food. The colony of ants can be classified as a homogenous swarm system, where the ants are sketched as agents that work towards achieving a common objective of collecting food. The system is homogenous, as all ants have the same level of capability: they can all sense, pick food, and carry it.
This is a multi-agent system, where the agents are all assigned different capabilities. This means that an agent can only carry out its own work and is not as flexible as the agents in the homogenous system; however, it could be argued that it allows agents within the environment to carry out multiple tasks. An example of agents in this system are birds.