2

I made a code using python language to solve partial differential equations My supervisor told me to check the results from the numerical model against the analytical results He didnt tell me the reason and said that I have to figure it out by myself

Any ideas what is the reason of checking the numerical model results against the analytical results ? Best Regards

davidlowryduda
  • 91,687
CDplayer
  • 25
  • 5
  • I have removed the [tag:analytic-number-theory] tag from this post, as I believe this has nothing to do with analytic number theory. If I am wrong, then I encourage you to edit your post to make it clear how number theory relates, and to include that tag in again. – davidlowryduda Jun 28 '18 at 19:22
  • Maybe your supervisor wanted to hint that you should have searched for "analytical solutions" all along. – Christian Blatter Jun 28 '18 at 19:28
  • @ChristianBlatter the analytical solution does exist and I was basing my model on the equations they used. But when I was comparing the results at 20 points , four of them didnt match which I dont know how to justify – CDplayer Jun 28 '18 at 19:46

1 Answers1

2

When analytic results are possible, it's a great way to test your numeric model. [Especially for a supervisor, since automated checking reduces the amount of work that your supervisor has to do]. This gives more confidence in later applications of the model against PDE where there aren't analytic results.

In practice every numerical tool should be tested against a wide variety of situations where the answer is known to indicate that the tool is functioning as intended.

davidlowryduda
  • 91,687
  • Thank you for your answer, would you please expand more. There are analytic results and I have to obtain exactly or close enough results to the analytic one. But I still dont understand why I have to check against the analytic results? @davidlowryduda – CDplayer Jun 28 '18 at 19:28