Skip to Main Content
The authors investigated an inverse problem in adaptive control. From an analysis of certain characteristics of an engineer's behavior they have been able to infer the existence, and possibly uniqueness, of a performance measure of a certain type. The existence of such a measure is significant because it implies that it is possible to objectify the predilection of the engineer. It is obviously far easier to discuss the relative merits of different criterion functionals than it is to compare such abstract entities as preference patterns.