Question about adjusting 'case' parameter in Poisson_2d class

[ccdzz]

Hello,

I'm working with the ls.ipynb file in the TransNet/PDE_Poisson_2d_annulus/ directory. In the third code block, I noticed the following line:

my_prob = Poisson_2d(case=1, data=data)

I appreciate that the author has provided a way to adjust the true solution and the right-hand side source term by changing the case value. This demonstrates the transferability of the method.

However, I've encountered an issue: when I change the case value to 2, 3, or 4, the resulting MSE (Mean Squared Error) is not as good as expected.

Questions:

Is this behavior expected? If not, what could be causing this discrepancy? Are there any recommended adjustments or additional steps I should take when changing the case value to maintain good MSE results? Any insights or suggestions on how to address this would be greatly appreciated. Thank you for your time and assistance!

[ccdzz]

Upon changing the case value to 2, I observed that the resulting MSE curve exhibits significantly higher values than expected. Despite my attempts to fine-tune the shape_vec parameter to improve accuracy, I was unable to achieve satisfactory results.