Flexible Scheduling for Semester 96A

The desiderata for any flexible scheduling scheme are relatively simple:

* To maximize the scientific return of the telescope globally — that is, to maximize the number of programs which are actually completed, in particular the highest priority programme.

* To continue to provide hands-on training for students and post-docs.

The difficulty of inventing a scheme satisfying even these modest requirements has delayed any serious attempt at flexible scheduling until now. With the imminent arrival of SCUBA and RxW — both high-frequency instruments — the need for an effective flexible scheduling system has become even more pressing. SCUBA and RxW will be highly competitive, and we envisage that many of the most highly-rated proposals will request time on one or both of these instruments.

With this in mind, the TAGs and ITAC have been working with the JCMT Director and his staff to try to define a workable system to begin in semester 96B.

At the present time, with 3 new instruments pending delivery either in semester 95B or in 96A it is difficult to implement significant changes to the scheduling technique and hence only minor attempts at both flexible and serviced mode observing will take place in these semesters. Certain flexible blocks have been created in the schedule for semester 95B and several allocations have been designated to be done in serviced mode by JCMT staff. The Netherlands community have taken the initiative and decided that ALL their allocated programmes will be attempted in a completely flexible fashion throughout the semester. The observations will be undertaken by Remo Tilanus and/or Fred Baas with occasional assistance from Netherlands observers who may visit the island for longer periods of time. The result of this experiment will be extremely useful in determining the direction of future JCMT allocations.

Further details on all aspects of JCMT scheduling should be referred to Graeme Watt at the JAC.

For semester 96A, it is unlikely that a significant fraction of the observing allocation will be setup in a serviced mode other than low-frequency backup programmes for SCUBA commissioning and for further commissioning of RxW. The community are advised that the SCUBA and RxW teams will require large quantities of stable, high-frequency weather in order to fully commission their instruments and therefore there will be ample opportunity for low-frequency (A-band and some B- band) applications to be completed.

In future semesters, the ITAC would like to allocate approximately half of the available JCMT time (after the E&C, UH and DDT shifts have been removed) to prioritized queue scheduling.

This gives each proposal a priority which is dependent on scientific merit, and then asks the PI to identify a minimum set of observing conditions (ie: tau(225 GHz), seeing, elevation, system temperature, etc). The highest priority proposal for which all these constraints are satisfied is executed. This scheme has the advantage that the highest-rated science always gets done with the highest priority.

The main responsibility for the conditions under which a programme is done lies with the Principal Investigator. The PI decides on the go/nogo criteria. This may extend as fas as the preparation of ICL control files if the PI so desires.

The removal of the 'backup' part of a regular application persuades applicants to seek a real allocation for low-frequency time if desired.

Sufficiently highly-rated proposals would be permitted to opt-out from queue scheduling, provided high- frequency proposals had a PATT-approved low-frequency backup. The backup could be either another accepted proposal, or a proposal which was not allocated time, but was nevertheless approved for this purpose by the appropriate national TAG. Opted-out proposals would be encouraged to form consortia and seek block scheduling to maximize their productivity. It is expected that most of these opt-out proposals will be primarily highly-rated low-frequency applications used for training students, research staff, etc.

In the first instance, we would anticipate queues could be created on a national basis, with support predominantly from national staff. We would hope however, that the system would evolve towards a single integrated queue in due course. The percentage of time allocated to queue scheduling will also be expected to evolve with experience.

There is no axe to grind here. We are definitely not prejudging the relative merits of high- vs low- frequency observing — proposals will continue to be ranked according to scientific merit by the TAGs as before. After that, the single task of the telescope staff and the TAGs is to maximize the scientific return from the telescope. Doubtless there will be teething problems with queued-mode observing, and we certainly expect that its operation will evolve with time. (Constructive) feedback from the users will be very important in striking the necessary balances. Your input is very welcome now — you can make your views known to any TAG member, Ian Robson or Graeme Watt. But do remember that we are trying to juggle many different interests and that it will not be possible to satisfy everyone. Please give the experiment a fair try!

Rachael Padman, MRAO (UK Rep for the ITAC) &

Graeme Watt, JAC (ITAC Tech Sec)