Ir. J. E. ALBERDA,
On the importance of Decision theory for Geodesists
i. Introduction
The object of this paper is to give an outline of some elementary
aspects of decision theory. It should be seen as the result of a rather
superficial exploration of this relatively new approach, an explo
ration which made to find out if there was anything in it which could
be valuable to geodesists.
There seems to be no universally accepted meaning of the term
"decision theory", especially if one takes into account its translation
into different languages. It may be taken to mean the same as
"operations research" which can be just about any activity which
uses scientific methods to find a quantative basis for decisions. For
example, the recent book "The Theory of Decision-making" by
W. Sadowski [8] has the sub-title "An introduction to Operations
Research"; it covers applications of differential calculus, linear
programming, probabilistic methods, the theory of games, statistical
methods and dynamic programming. It seems that in British-
American usage the term decision theory is taken to mean more
specifically statistical decision theory, i.e. the theory founded by
A. Wald [io] which gave a new approach to statistics. The intro
ductory textbook "Elementary Decision Theory" by H. Chernoff
and L. E. Moses presents the decision-theory approach to elemen
tary statistical methodsthis paper is mainly inspired by ideas
presented in it.
Every language will have its own translation for "decision theory"
and "operations research", and each translation may have different
shades of meaning. The field of activity and its name in the Dutch
language were treated by W. Monhemius [6]. He gives an outline
and examples of different techniques and discusses the appro
priateness of the Dutch name, which literally translated into
English would be something like the "knowledge of decision
making". It is not surprising that he arrives at the conclusion that
the term does not fully cover the field of activity, and at the same
time suggests more than is actually meant by it. We shall not dwell
upon terminology, but rather follow Monhemius in restricting
ourselves to the point of view that decision theory is mainly
concerned with the taking of a choice between alternatives.
In each decision problem we can imagine we have to face an
opponent on whom the consequences of our actions are dependent.
We wish to take a decision whose consequences have certain
optimum properties. The opponent may be a malevolent intelligent
adversary who wants to do us as much harm as possible. This
Computing Centre of the Delft Geodetic Institute: