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




Roach, Jon H.

Access rights

Worldwide access.
Access changed 9/17/15.

Journal Title

Journal ISSN

Volume Title



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.



Swarm intelligence., Emergent behaviors., Evolutionary computing., Multi-objective swarms., Disjunctive control.