Distributed Intelligence

The DIS will work to complement the evolutionary optimisation of the EvE indirectly. An agent, when finding similarity with another agent based upon their descriptions, will be able to migrate. It will target the migration to a Habitat where it will have the opportunity to be useful, based upon the migration and usage history of the similar agent. In biological terms, it is equivalent to providing each individual with a relatively sophisticated mechanism of social interaction. It should allow niches to be fulfilled faster, so that the Habitat clusters will reach their ?climax communities? sooner, thereby significantly improving the speed of the ?succession? process. Also, communities will be able to adapt faster to changing conditions.

As the EvE increases in size (number of users), the total number of agents (services) available globally will become increasingly large. For the evolutionary optimisation to continue to work efficiently as the EvE expands, optimal subsets of the power-set of agents available globally will be required at the Habitats. The migration probabilities between the Habitats will work to achieve this in a `passive' manner. Although the mechanism is effective, the migration of agents will be relatively slow and not strongly directed. It allows the agents, based primarily upon success at their current location, to spread in the correct general direction within the Habitat network. The distributed intelligence will work in a more ?active? manner allowing the agents immediate highly targeted migration to specific Habitats, independent of success at the current location, instead of the generally directed migration only after success at the current location. This will help to optimise the subset of possible agent-sets found at the Habitats, which are then used in the evolutionary optimisation processes to find applications (combinations of agents) to user requests.