Information Technology Reference
In-Depth Information
Table 1.
Real-time multitasking software test report
Failure Rate with
DS
Task
module
Test
Failure
B
i
Test
Time
T
i
Code
Number
N
i
P
i
Φ
i
1
0.1566
5
3
470
0.125988
2
0.1364
7
2
405
0.034529
3
0.099
5
2
396
0.09049
4
0.2227
38
10
1783
0.211092
5
0.0922
13
4
738
0.095572
6
0.0495
17
3
792
0.029602
7
0.0420
22
4
672
0.042815
8
0.0503
14
3
807
0.089472
9
0.0222
12
2
440
0.028256
10
0.0584
20
6
1169
0.091465
11
0.0501
19
3
1002
0.100794
12
0.0528
32
11
1267
0.043711
13
0.0282
21
3
848
0.057195
14
0.0309
16
4
928
0.054471
Table 2.
Analyzing result comparison
Result of
original neural
network
Relative error
of original
neural network
Result of
improved
neural network
Relative error of
improved neural
network
Result with DS
0.057196
0.061472
7.5%
0.058447
2.1%
0.054471
0.058793
7.9%
0.053373
1.7%
5 Conclusion
Characteristics of real-time determine that it laid high requirements on software reli-
ability. Based on analyzing on characteristics of real-time multitasking software and
highlight its distinguish with reliability with general software, a kind of neural net-
work analyzing model of real-time multitasking software was presented. The model
introduced running time of each task module into reliability model to meet features of
real-time multitasking software, which is more accurate than methods not according
to task module. It can analyze reliability of real-time multitasking software well and
solve shortcoming of large error of general model, which also has good scalability.
Example study shows its effectiveness and superiority.
Search WWH ::
Custom Search