Conference Paper
Using Feed Forward Neural Network to Solve Eigenvalue Problems
Table 5
Analytic and neural solution of Example
2.
| Input | Analytic solution | Output of suggested FFNN for different training algorithms | | | Trainlm | Trainbfg | Trainbr |
| 0.0 | 1 | 1.000000000000000 | 0.999999791497422 | 1.000000531336361 | 0.1 | 1.000100005000167 | 1.000144087759814 | 1.000100368418017 | 1.000097184294931 | 0.2 | 1.001601280682940 | 1.001601280682940 | 1.001498511767109 | 1.001607051239879 | 0.3 | 1.008132893753152 | 1.008132893753152 | 1.008132867669427 | 1.008128167440348 | 0.4 | 1.025930494190382 | 1.025930494190382 | 1.026019395339251 | 1.025927732110325 | 0.5 | 1.064494458917860 | 1.064590790583371 | 1.064494437575892 | 1.064506007546008 | 0.6 | 1.138372943065416 | 1.138372943065416 | 1.138272638536007 | 1.138359307576485 | 0.7 | 1.271376281592894 | 1.269464922443141 | 1.271376301269824 | 1.271384988973427 | 0.8 | 1.506215178528160 | 1.500405060697450 | 1.506214213098092 | 1.506211973532767 | 0.9 | 1.927261339283878 | 1.927261339283878 | 1.927192081328471 | 1.927261991541353 | 1.0 | 2.718281828459046 | 2.718281828459046 | 2.718280402057775 | 2.718281767743469 |
|
|