Check time-varying $r$

[jlperla]

Want the $r$ to change smoothly.

Check for some interior set of the $x$ values (e.g i in the middle half of the statespace). FInd the $$ \max_{1/4 I < i < 3/4 I}{v^{n+1}_i - v^n} $$ and see how that number changes with $n$

[sevhou]

I tried for time varying $u$ function, the value $\max{n,i} v{i}^{1}-v_{i}^{n}=7.4$, with tgrid 0 to 1, zgrid from 0.1 to 3, it's a strictly increasing function on t dimension.

[jlperla]

Are there any odd discontinuities on the final few points? Play around with this, maybe even doing a 3D plot of the value function over time.

You should be able to do that we a reshape command in matlab, and then surf or something like that.

On Mon, Dec 11, 2017 at 3:03 PM sevhou notifications@github.com wrote:

I tried for time varying $u$ function, the value $\max{n,i} v{i}^{1}-v_{i}^{n}=7.4$, with tgrid 0 to 1, zgrid from 0.1 to 3, it's a strictly increasing function on t dimension.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/econtoolkit/continuous_time_methods/issues/36#issuecomment-350887726, or mute the thread https://github.com/notifications/unsubscribe-auth/AMf-ITMYxmaPtgwXJEgyhJDsSGn8eDuaks5s_bTegaJpZM4Q-JWy .

[sevhou]

Yeah I've done that and also plot some 2-D graphs and check along t dimension and z dimension, I think it behaves smoothly.

[jlperla]

OK. Keep all of the code in the tests for later comparison. If things seem to be smooth for both the $r$ and the $u$ changes, that is good news. Make sure you are looking at examples where the changes are cranked up to be large enough to see relatively large swings.

On Mon, Dec 11, 2017 at 3:16 PM sevhou notifications@github.com wrote:

Yeah I've done that and also plot some 2-D graphs and check along t dimension and z dimension, I think it behaves smoothly.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/econtoolkit/continuous_time_methods/issues/36#issuecomment-350890445, or mute the thread https://github.com/notifications/unsubscribe-auth/AMf-ISncdTZjgs5x1nMwVnpniUk3pky3ks5s_bffgaJpZM4Q-JWy .