
294 Optimization: Algorithms and Applications
Clearly, the minimum of the two terms is
12 2
0
1
1
−
≥
x
x
Thus,
f
x
xs
2
1
12
3
11
2
12 2
3
12 2
*
,( ) =
−
= −min
Therefore,
f s x
x
1 1 1
1
2 3
11
2
12 2
3
*
,( ) = +
−
=min 22 12 2 12
1 1
x x+ − =
Thus, the maximum value of the function is 12 and occurs at
x
1
0
=
.
Substituting the value of x
1
in one of the constraint equations gives
x
.
11.3 Probabilistic Dynamic Programming
In deterministic dynamic programming, the state and decisions of the pres-
ent stage completely determine the state of the next stage. In probabilistic
dynamic programming, the state of the next stage is determined with some
proba ...