s à lot of 
| order to obtain 
ficant models. 
not much 
(| period between 
tawa. 
| access to a 
only resumed 
roverned by the 
'ormation for 
r the conviction 
lereby increase 
nd to 
; will be 
In paragraph 2, 
ı the problem 
ed again in 
yi as probably 
f ihe original 
ih work is 
)ssibly lead 
sooperation of 
it, has to be 
reneral 
111 be necessary 
relop means to 
The problem analysis will form the basis for two further developments: 
a) A systematic planning procedure. In the long run this might be a computer 
programme for optimisation. In the immediate future it will at least consist 
of a complete list of considerations to be used for systematic planning. 
b) A comprehensive computer programme forQ.R. 
The objectives to be achieved by such a computer programme for Q.R.; ae intended 
by the ITC, are manyfold: 
-Evaluation of alternative process components in view of various applicatigns 
and circumstances. 
-Establishment of general rules and relationships with respect to various productis 
and circumstances. 
-Establishment of general information concerning the cost performance ratio of 
different types of products under various circumstances. 
-Identification and quantification of trade-off situations and break-even points. 
-Identification and quantification of bottle-necks in various procedures. 
-As a COR ROQUenCO of these bottle necks, identification and establishment of 
specifications for desirable techaalonisal developmehts. 
—Sensitivity investigations of parameters and the various categories of influencing 
factors. 
-Investigation into the stochastic characteristics of various influencing factors 
and their consequences on the total system. 
— And, last but not least, such an O.R. programme should also represent a valuable 
aid for educational and consulting activities. 
Finally, mention might be made of the fact that, although the absolute cost values 
derived by such an approach might not be very accurate (due to fluctuations and 
uncertainties in the input data used), the relative values can however be expected 
to be more accurate, and therefore more universally valid. This is of course what 
really counts with respect to optimisation.