This paper describes a software architecture which supports the design of h
ierarchical controllers that provide facilities for adaptation, supervision
and task planning. It details how this form of functional hierarchy differ
s from the structural hierarchy also inherent within a complex control syst
em. Then, both forms of hierarchy are combined in a single design notation
and development methodology. The system utilises intelligent control techni
ques (neuro-fuzzy and genetic optimisation) for controlling a cryogenic pla
nt used for superconductor testing by cooling the test samples to temperatu
res below 100 degrees K. The system supports the design of a hierarchical c
ontroller that provides facilities for adaptation supervision and task plan
ning. Simulation results are presented.