Step 5: Running the scheduler
As discussed in a previous section, the scheduler may
be unable to insert some observation requests in the schedule (this is referred to
as a failed insertion in the screen output and in the log file). Such cases may
often be avoided or turned into successful insertions with the help of the scheduling tips
listed below. Note that the various recommendations may conflict with each other to some extent;
some judgment may be necessary to obtain the best results.
- Re-run the scheduler to obtain better schedules. If you are not satisfied with the
schedule you generated, you may try to modify the list of observation requests based on
the recommendations given below, and then re-run the scheduler, trying to obtain a more
satisfactory schedule. If necessary, this refinement process may be repeated more than once.
As you acquire practice using the scheduler, you should be more and more successful at
preparing good schedules in a single iteration or in a few iterations.
- List your higher priority targets near the top of the list of observation requests.
This will ensure that these targets get the best observing windows, and will also maximize
the probability that they will be inserted in the scheduler.
- Requests for long series of observations of a given target should appear near the top of
the list of observation requests. It is often easier to insert several short series of
observations among the exposures belonging to a long, previously inserted series, than it
is to insert a long series in a schedule already populated with several short series.
- Requests for targets near the ends of your observable declination range should appear
near the top of the list of observation requests. Such targets usually have very short
observation windows, which may easily be taken by previously scheduled targets. This applies
to far southern (northern) targets to observed from the northern (southern) hemisphere.
- Request successive exposures for the fastest moving targets. If you request
a sequence of exposures of a fast moving target and specify the keyword mot
(which means that the time interval between exposures should be such that the target
moves a few pixels between one exposure and the next), the request may fail because
the target will move too many pixels in the finite time it takes to download an image
and perform a short slew to correct the pointing of the telescope for the target's motion.
This problem may often be solved by specifying the keyword suc (successive exposures)
instead of mot.
- Move observation requests which could not be inserted toward the top of the list of observation
requests. If an observation request could not be inserted because its observing window
has been taken by previously scheduled requests, moving it toward the top of the list of
observation requests may increase its probability of being scheduled, possibly at the expense
of failing to insert other requests which could previously be scheduled. In these cases, one
would have to decide which target is more important to keep in the schedule.
- Relax observational constraints of observation requests which could not be inserted.
Constraints such as minimum altitude, maximum |hour angle|, and moon avoidance radius may
be too restrictive, making it impossible to schedule a given request. For example, if at the
beginning of the observing run the target is west of the meridian at altitude 30 deg, an
observation request specifying a minimum altitude of 35 deg will not be inserted. If you
think that images taken at at altitudes of at least 25 deg would still be acceptable, you
might relax the minimum altitude constraint to 25 deg, possibly increasing the probability of
scheduling the observation request. Another example would be a far southern target to be
observed from a station in the northern hemisphere. If the minimum altitude is set too high,
the request will not be scheduled because the object never rises above the specified minimum
- Sequences of exposures of very slowly moving targets. If you wish to obtain a sequence
of exposures of a very slowly moving target with an interval between exposures which is long
enough to clearly show the target's motion (it should move at least 2-3 pixels between
exposures), the request may not be scheduled because the total time spanned by the observations
would have to be longer than the object's observing window in the night in question. For
extremely slow targets (such as trans-Neptunian objects near their stationary point), you may
consider taking one exposure per night and then blink the exposures from successive nights
to detect the object's motion.
- Targets which rise in the east toward the end of the observing run. Such targets should
be listed in (approximate) order of decreasing right ascension (that is, easternmost targets
first). The easternmost targets are the ones at lowest elongation, and therefore have the
shortest observing windows, so they should be scheduled first to prevent other requests from
"stealing" their observing windows.
- Targets which set in the west near the beginning of the observing run. Such targets should
be listed in (approximate) order of increasing right ascension (that is, westernmost targets first).
The westernmost targets are the ones at lowest elongation, and therefore have the shortest
observing windows, so they should be scheduled first to prevent other requests from
"stealing" their observing windows.
- Fill "holes" in the schedule with targets at appropriate right ascensions and declinations.
If after scheduling your most important targets, you find "holes" (periods of idle telescope time),
you might select additional targets from the list of potential targets. The chosen targets
should be visible during the time periods corresponding to the "holes", and their observations
should have at least some chance of fitting in the available "holes". For example, a 10-min
"hole" would not be enough to schedule a sequence of three 2-min exposures spaced 10 min apart.
Single exposures of fixed targets and sequences of a few short successive exposures of a fast
moving asteroid or comet are often suitable to fill up small holes in the schedule.
- User quotas. If you are a telescope manager and have assigned quotas of telescope time
to the various users, you may use the information given in the telescope usage summary at
the beginning of the schedule summary file to ensure that no user will exceed his/her quota.
If some user exceeds the assigned quota, you may delete some of his/her observation requests
to make the amount of telescope or exposure time used fit the quota.
Previous: Schedule output (continued)
Next: Starting the scheduled observations
© 1999-2004 Paulo Holvorcem