Geoscience Reference
In-Depth Information
The matrix of this game has a saddle-shaped point in the case of following
conditions: M a (0)
N Rb (0)exp[d b N Rb (0)], M b (0)
N Ra (0)exp[d a N Ra (0)] and that it
takes place in the real system.
Thus for both systems it is advantageous during the
first step to provide the
protective elements with a small effectiveness and to increase their effectiveness
during the second step. This is natural, since it is better to lose a greater number of
protective elements during the
first step, thus securing a reliable defense of the
working elements during the second step of the game.
From Eq. ( 10.53 ) it follows that the loses of a-and b-elements by system H and
N, respectively will amount to:
Q 1 ¼ exp 1 þ d 1a d 2a
f
ð
Þ N Ra ðÞþ d 2a M b ðÞ exp d 1a N Ra ðÞ
½
g= d 2a ;
ð 10
:
54 Þ
Q 2 ¼ exp 1 þ d 1b d 2b
f
ð
Þ N Rb ðÞþ d 2b M a ðÞ exp d 1b N Rb
½
g=
d 2b :
In order to determine the optimum value of the effectiveness of the R a -and R b -
elements for each step of the game, we must
nd:
min
ð d 1a ; d 2a Þ
Q 1 ð d 1a ;
d 2a Þ and min
ð d 1b ; d 2b Þ
Q 2 ð d 1b ;
d 2b Þ
From Eq. ( 10.54 ) we obtain:
@
Q 1 =@
d 1a ¼ 1 þ 2d a d 1a
ð
Þ
f
2N Ra ðÞ M b ðÞ
ð 10
:
55 Þ
½
1 þ N Ra 2d a d 1a
ð
Þ
exp d 1a N Ra ðÞ
½
g ¼ 0
;
@
Q 2 =@
d 1b ¼ 1 þ 2d b d 1b
f
1 þ N Rb ðÞ 2d b d 1b
ð
Þ
2N Rb ðÞ M a ðÞ
ð 10
:
56 Þ
½
ð
Þ
exp d 1b N Rb ðÞ
½
g;
From Eq. ( 10.55 ) it is evident that if the equation:
f
ð
2d a d 1a
Þ= d a d 1a
ð
Þ
g exp d 1a N Ra ðÞ
½
¼ 0 : 5N Ra ðÞ= M b ðÞ
ð 10 : 57 Þ
has a real root 0 d 1a 2d a then system H, by utilizing this root for its own
optimum strategy, can guarantee on the average losses of elements not exceeding
Q 1 ¼½e ð 2d a d 1a Þ 1 . If Eq. ( 10.57 ) does not have a root in the interval [0, 2d a ],
the optimum strategy is then determined either by the root of Eq. ( 10.55 )orby
d 1a ¼ 0. In particular, when M b (0) = N ra (0), then d 1a ¼ 0. Similar calculations are
realized for Eq. ( 10.56 ).
Search WWH ::




Custom Search