DIAG_DSIZE

[wtam]

Dear Didier,

Is the example below the correct way of using diag_dsize?
drift 10 10
diag_dsize3 11 5
drift 10 10
diag_size 11 0 0 20
diag_dsize3 11 5

Is the above the same as this example below?
drift 10 10
diag_size 11 0 0 15
drift 10 10
diag_size 11 0 0 20

Thanks and best regards,
Wai-Ming

[Didier]

Dear Wai-Ming,

The fisrt example seems good, but size3 concerns phase spread and it's not true for the second one !

Regards,

Didier

[wtam]

Dear Didier,

If I understand the user manual correctly, my second example is also for phase spread. Please see attached snapshot.

Best regards,
Wai-Ming

Screenshot 2021-10-11 122505.png (56.85 KiB) Viewed 1540 times

[wtam]

Dear Didier,

My understanding to my first example is that:
drift 10 10
diag_dsize3 11 5
drift 10 10
diag_size 11 0 0 20
diag_dsize3 11 5
I am demanding a phase spread dP(z=20)=20 at z=20. At the same time I am demanding the difference in phase spread to be dP(z=20) - dP(z=10) = 5. So equivalently, I am demanding dP(z=10)=15.

The above is exactly the same as my second example:
drift 10 10
diag_size 11 0 0 15
drift 10 10
diag_size 11 0 0 20

My questions are:
Are the two example the same, or if I have misunderstood something?
If they are the same, which one is better to converge the optimization solver?

Thanks and best regards,
Wai-Ming

[Didier]

Dear wtam,

Yes, that seems similar, now did you simulate these two ways to check ?
In both cases, solver is the same.

Regards,

Didier