Neville's delay problem.

[Peter230655]

@moorepants: I have played around with this for quite a while, without any success.
The attachment shows what I tried, but failed. I have no idea where my it went wrong.

25_12_17_Neville_s_delay_problem.pdf

[moorepants]

Hi Peter, I have not had time to look into this due to vacation and being sick. My suggestion is to discuss it with Neville.

[Peter230655]

Dear Jason, Hope you will feel better soonest! This has no rush, I just wanted to ‘write down’ my current status. I have kept Neville informed about my (lack of) progress, but in the past 1 – 2 weeks I have not heard back from him. Get better soon! Peter From: Jason K. Moore ***@***.***> Sent: Wednesday, December 17, 2025 8:40 AM To: csu-hmc/opty ***@***.***> Cc: Peter Stahlecker ***@***.***>; Author ***@***.***> Subject: Re: [csu-hmc/opty] Neville's delay problem. (Discussion #525) Hi Peter, I have not had time to look into this due to vacation and being sick. My suggestion is to discuss it with Neville. — Reply to this email directly, view it on GitHub <#525 (comment)> , or unsubscribe <https://github.com/notifications/unsubscribe-auth/AT5MQUVY22E5UF7RQAGJHID4CECFTAVCNFSM6AAAAACPI5BRV2VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTKMRXGYZDOMY> . You are receiving this because you authored the thread. <https://github.com/notifications/beacon/AT5MQUTREG2TF6TR5F6LSP34CECFTA5CNFSM6AAAAACPI5BRV2WGG33NNVSW45C7OR4XAZNRIRUXGY3VONZWS33OINXW23LFNZ2KUY3PNVWWK3TUL5UWJTQA5EMPC.gif> Message ID: ***@***.*** ***@***.***> >

[Peter230655]

Update 1.1.26
I have played around with this for quite a while.
Except for the simplest problem, like a particle moving from A to b on the x - axis, I have not managed to get it to converge.
It seems, that for more complicated systems, my 'derivatives' are better than just setting the derivatives = 0, either way does not work for me.
Intuitively, as opty looks at all points in time simultaneously, this should be a good problem for opty - but i do not manage to get it.

[Peter230655]

My successes are limited, see e.g. here:
https://pydy.org/pst-notebooks/examples/plot_car_on_racecourse_delay_opty.html#sphx-glr-examples-plot-car-on-racecourse-delay-opty-py
@moorepants Neville told me that he had talked to you before Christmas, you somehow solved the problem, and he'd revert. I have not heared anything since. As I have played around with this problem for some time, I am interested how it was solved. Thanks!

[moorepants]

I don't recall solving it! :)

[moorepants]

Ah, he did realize at some point he was trying to solve a problem harder than he actually needed for his thesis. So maybe that is what he's talking about.

[Peter230655]

Maybe that was it and I did not totally grasp the difference.
Still I think it is a very interesting problem, and I thought opty should be ideally suited as it looks at all time steps simultaneously. But I have not been very successful so far. I think, I got the derivatives needed for numerical trajectories right - but of course I do not know for sure.