Emergent behaviors of multi-objective swarms with applications in a dynamic underwater environment.

The allocation of resources between tasks within a swarm of agents can be difficult without a centralized controller. This problem is prevalent when designing a swarm of Autonomous Underwater Vehicles, in which underwater communication becomes challenging and a centralized controller cannot be used. In this thesis, a disjunctive fuzzy control system is used to solve the problem of resource management. Multi-objective, multi-state swarms are evolved with an offline learning algorithm to adapt to dynamic scenarios. Some of the emergent behaviors developed through the evolutionary algorithm are state-switching and recruitment techniques. In addition, the adaptability of swarms is tested by removing sensors from the system and re-evolving the swarm to allow it to compensate for its sensor loss. The concepts of a multi-objective, multi-state swarm are also applied to an underwater minefield mapping scenario, which is used to test the robustness of the swarm with respect to swarm size.