Using a teleo-reactive approach in building self-managing systems

Abstract

The Teleo-Reactive programming style has been demonstrated as an effective and robust way to design the logical processes of autonomous agents. This robustness is largely due to an approach to achieve goals and deal with unexpected events which is much more like the human thought process than traditional programming techniques, as it works at a higher, more abstract level. We find that this approach is as effective in producing much higher level autonomous applications, providing the same benefits as the technique does for lower level robotics and autonomous agents. The technique embraces unexpected change, accepting that events and errors are inevitable and instead of attempting to preempt these possibilities, simply moves the current state back or forward through the program accordingly.We have built a framework to assist in engineering applications using this technique. Using this framework, we develop self-managing programmes that demonstrate our claims.