Developing teleo-reactive autonomic solutions
James Hawthorne, Richard John Anthony · Greenwich Academic Literature Archive (University of Greenwich) · 2010
The increasing size and complexity of modern software and the difficulty for humans to maintain these systems has meant there is a demand for software to manage themselves, thereby reducing the need for human intervention. Autonomic systems can have many self-management characteristics in order to solve issues which would normally require a human to manage, maintain or correct. These characteristics include the ability to heal, configure, optimize and protect themselves. We have taken the Teleo-Reactive (T-R) technique, developed for control of autonomous agents, and designed a framework around T-R to support development of much higher level software. We argue that a T-R approach is a simpler method of developing autonomic systems than many other techniques. Although T-R systems are relatively easy to understand and can produce small and compact code structures, the design process can be difficult. In this paper we highlight the problems and develop some strategies for designing T-R programs.