14 2. VALUE AND RATIONALITY: TRADITIONAL ENGINEERING DEFINITION
us, what engineers needed was to identify which group category the system belongs to
(System Identification). en, they will know what parameters they must take care of. ere-
fore, functions and cost can be evaluated quantitatively and such rationality is mathematical
rationality.
BOUNDED RATIONALITY AND SATISFICING
But Herbert Simon, Nobel Laurate in economics, pointed out that such rationality is
bounded [1]. Increasing size and complexity of the system causes computational complexity
and we cannot apply mathematical rational approaches. He pointed out that optimization is
possible only when mathematical rationality can be applied. Beyond the bounds of rationality,
we have no other choice than to settle for satisfaction. So, he proposed Satisficing (Satisfy C
Suffice) [2].
INDIVIDUAL PRODUCTS
Why value can be evaluated quantitatively, and optimum result (value) can be obtained in tradi-
tional engineering, is because engineers used to work on individual products. e systems they
worked on are not too large or complex. erefore, they can reduce the number of degrees of
freedom and solve the problem along the line of mathematical rationality.
GLOBAL OPTIMIZATION—ANOTHER SATISFICING
We must remember that even in the field of optimization, when it comes to global optimiza-
tion, we are not really optimizing mathematically. What we do is nothing other than satisficing
(Fig. 2.1)
SA Simulated Annealing
Global Optimization
Fluctuations
Figure 2.1: Global optimization.
For example, simulated annealing repeats searches many times. So, we believe the highest
one we find after many searches is the highest (optimum). But, the higher peak might come up
in the next search.