I think there are numerical instabilities being introducing for large numbers of grid points due to the division by $\Delta$ and $\Delta^2$ in the definitions.
I think I made all of the changes to the algebra document, though you should check very carefully in conjunction with #11 and then #13. After the algebra is verified, and before you do $13, you should convert the discretization code to match the algebra in the document.
Basically, the discretize_univariate_diffusion.m file just needs to match the new algebra for the A to return, and the checks in the test suite are modified to the new A.
I think there are numerical instabilities being introducing for large numbers of grid points due to the division by $\Delta$ and $\Delta^2$ in the definitions.
I think I made all of the changes to the algebra document, though you should check very carefully in conjunction with #11 and then #13. After the algebra is verified, and before you do $13, you should convert the discretization code to match the algebra in the document.
Basically, the discretize_univariate_diffusion.m file just needs to match the new algebra for the A to return, and the checks in the test suite are modified to the new A.