Simulating Emergence In Biological Systems. Emergence can be seen throughout nature, and the observation of these natural systems has inspired computational models, to aid in existing problems in areas such as network- ing infrastructure. For example, Panait and Xxxx show how a colony of ants foraging for food can find optimal paths without any centralised control or knowledge, and they imitate this mechanism in an agent-based simulation [88]. In their simulator, each agent is situated in a grid, and moves towards one of two possible locations: a food source, or the colony’s nest. However, agents have no knowledge about the wider world, and they are unable to map a geographical space internally. Instead, agents must rely on pheromones that they leave behind themselves, and those that have been left behind by others, to guide them. Initially no pheromones exist, and so all agents randomly explore, but once an agent finds a food source it will begin to drop pheromones pointing to the food source as it returns to the colony’s nest. As more agents find the food source, they increase the pheromone levels in its vicinity, gradually forming a pheromone gradient, so that in the future agents can move directly to the food source, rather than locate it through exploration. Additionally, because these pheromones are continually updated, if the food source is moved the pheromone gradient will, over time, change to direct agents to the new location. The processes that the individual agents undertake (moving up pheromone gradients, updating pheromone levels according to the immediate environment, collecting food if it is available and dropping food if it is at the nest) are very simple, and the technique for interaction (dropping pheromones) is simpler still, but the actions that the ants take do not seem to have any direct relation to seeking out food. Instead, they move around according to the pheromones in their immediate vicinity, and when they come across food, they pick it up. Yet, what emerges is a completely new and coherent behaviour: the system as a whole seeks out a locally optimal path to and from food sources, and adapts this path if the location of the food source changes.
