Decision making in an environment with unknown parameters

Decision making is the core of any control system. The driving metaphor for a traditional decision maker is the existence of an explicit memory base and its ability to analyze and express any situation that the controlled plant experiences, by a set of mutually disjoint rules. The disjoint nature of the rules reflects the fact that in engineering world we tend to partition the problems to smaller tasks, deal with each individual task separately, and then reassemble the overall solution. Given the ever increasing need for mechanical systems operating in environments with unknown parameters, and the fact that certain situations in real-time do not lend themselves easily to this method of partitioning, we are proposing a decision making scheme which develops its own rule base as it goes along and confronts the problems in their entirety, thus, functioning to a great extent like the human brain