Information Technology Reference
In-Depth Information
Table 5.2
The Web Services
p
0
and
p
1
22
Denote and Their QoS
Attributes
Availability (after
taking logarithm)
Places
Web service name
Cost
p
0
Sales management service
10
0.99 (
0.01)
p
1
Logistics service
50
0.84 (
0.17)
p
2
Order management service
20
0.92 (
0.08)
p
3
Customer payment service
50
0.89 (
0.11)
p
4
Transportation management service
35
0.97 (
0.03)
p
5
0.95 (
0.05)
Warehouse management service
60
p
6
Invoice management service
40
0.89 (
0.11)
p
13
Railage registry service
10
0.70 (
0.36)
p
14
Airfreight registry service
100
0.96 (
0.04)
p
15
Product inquiry service
(application area indexed)
25
0.89 (
0.11)
p
16
Product inquiry service (provider indexed)
20
0.92 (
0.08)
p
18
Credit validation service
60
0.95 (
0.05)
p
19
Remittance service
30
0.98 (
0.02)
p
20
Bank 3 account service
28
0.94 (
0.06)
p
21
Bank 2 account service
22
0.83 (
0.19)
p
22
Bank 1 account service
25
0.88 (
0.13)
L
รพ
L
p
7
12
;
17
Dummy service
In order to make our configuration adapted to the dynamic environ-
ment, we can perform sensitivity analysis [131] after the optimal
configuration is found. Generally, in the standard linear programming
problem, if there is any change in the values of coefficient matrix (i.e.,
[
E
-
A
]), the right-hand side vector (i.e.,
M
0
), or objective function
coefficients (i.e.,
V
i
and
w
i
), the optimal solution is likely to change.
However, when the perturbations are within a certain range, the current
optimal solution may remain unchanged. This invariance of the optimal
solution is a desirable property that helps significantly reduce the
computational complexity when perturbations occur and/or we are
uncertain about the exact values of coefficients. In the SFC problem,
the coefficient matrix and the right-hand side vector often stay
unchanged, while the objective function coefficients
V
i
often fluctuate
