Mixing Formal Methods, Machine Learning, and Human Interaction Through an Autonomics Framework

Autonomic approaches aim to manage large complex systems by enabling self-adaptation in response to the changing state of these systems. As the size and complexity of systems increases, autonomics helps conceal that complexity and provides a higher level, more abstract view to human managers of these systems. One such autonomic approach is the Rainbow autonomics framework, developed at Carnegie Mellon University. This paper describes our use of Rainbow in various applications including supervisory control of unmanned systems and management of operator task assignment. It also describes enhancements made to the version of Rainbow utilized.