Please move this to a relevant Advanced topic if required (I can’t post there even though I’m using Advanced).
This is a suggestion for a new scheduler type:
Targets are sorted by Priority and RA (although the sorting doesn’t really matter, the below could apply to any type of priority order).
RoboTarget picks the first eligible target on the list and starts imaging.
The scheduler continues to run in the background, checking every 60s for eligible targets.
As soon as a target higher on the list becomes available and the sequence minimum duration has elapsed, the scheduler stops the current sequence and switches over (optionally waiting for the actual exposure to finish).
The imaging sequence starts again for the new target until another target higher on the list becomes eligible or, finishes and moves on to the next target down the list.
Targets lower down the list can’t interrupt those running above them. Only targets higher on the list are allowed to “skip ahead”.
This type of scheduler would allow higher priority targets to receive the time they deserve while still making the most use of the night.
For example if you have a top priority target, visible only for a couple of hours, among dozens of long running lower priority ones, the default scheduler may collect little if any data on it. Effectively making a high priority target, low priority.
Of course I’m not an expert on these kind of things and the devil is in the details, so feel free to ignore if any of this sounds naive.
I echo the needs of this scheduler. Ultimately Voyager Advanced show prioritize truly high priority targets. We had a discussion on this post here.
Just to add thoughts here, on your bullet #1, when Robotargets sorts targets, it should also be able to populate the start & end time for each one given the constraints. That way it’s simply arranging the targets based on start & end time, with priority considered. That way you don’t need to check it in the background.
Of course as long as the function is delieverd, either method works for me. Thanks!
Some of yours ideas simply not works, I understand its simple to think about a scheduler but things not work like you think. For example if you change target settings the previous calculation are not more valid.You may have added targets and with more higher priorities or simple remove one. Or for example between 10 target at same priority and all greater that the actual which one is to choose (usually by date time introduction like the default scheduler) but if one oldest with same priority become available ? I stop thing each 15 minutes … what you think to have in your HD at finish of nights in this case !!
I suggest all to be patience and leaves things running, AP is a pleasure not a rush ! If you live AP like a rush something is not gone in the right way.
If there is a way to have better things I will add more scheduler. The reason why I have foreseen that everyone can choose their own scheduler is precisely this, to give you the possibility to make mistakes independently and find the most congenial things to you. There is no wrong and no right solution!
I couldn’t agree more with Leo’s statement, namely “AP is a pleasure not rush”. Let’s be patient friends - Leo mentioned it already many times - more schedulers will come in due time - but meanwhile, let’s just enjoy the beauty of Advanced…
PS : by the way, Yizhou’s suggestion about populating the start & end time for each target is a good one !
Populating the start and end times sounds good. Note that targets could have multiple start and end times. For example, image Target A (normal priority) from 22:00 - 00:00, image Target B (high priority) until 02:00, then switch back to Target A for the remainder of the night.
And just to be clear, this is not about wanting to rush AP. These are only suggestions to allow higher priority targets to at least get a chance to collect data. Otherwise we wouldn’t need a priority constraint at all if patience was all that was required.
Voyager has been always very open to suggestions - look at the number of new features that have been implemented over the last couple of months, and I am not only referring to Advanced. I can’t speak on behalf of Leo, but for sure he’s been always very, very sensitive to suggestions .
So, I am sure that we will see sooner or later the suggestions that you raised
It’s fine Leo, I’ve gone back to a manual DragScript containing targets with start and end times. This allows the most of the night to be used while still collecting data on high priority targets. This does require some manual planning but for me works better. Just wish I could’ve trialed Advanced before buying it.
I am looking forward to the release of the priority scheduler as I think it should work well for my purposes. The big step forward for me in advanced is the ability to have the scheduler manage which filters to shoot based on the moon, previously I was managing that with multi sequence dragscripts, changing the sequence of scripts and times that each one ran night by night based on the moon phase.
With a priority based scheduler I will have to take on a bit of the load myself in setting priorities but it will allow me to limit the time to shoot each target closer to the meridian ot get the best skies, knowing that I can add a “Fill in” target that can be shot while the higher priority targets don’t qualify to be shot, and that they will start shooting more or less as soon as they meet their constraints.
For me advanced is great already and a “Priority” scheduler will make it even better.
I’m also running into the issue that some lower priority targets get picked (because none of the higher priority targets are available at that time) and then get imaged for 5 hours.
I could imagine two solutions:
Simple : Every 60 or 90 minutes, re-run the robotarget to recheck if higher priority have now become available
complex: (1) start with First priority target, check over duration of night when each become available and allocate on the basis of first time saved. So for example, you would have target X from 1.30 to 3.00 and target Y from 4.00 to 5.00. (2) Then go to High and do the same thing on remaining slots, (3) continue with next level.
Is not an issue this scheduler work in this way.
Use max sequence time for lower priority targets if you want to rescheduling and catch the higher priority target if available
.