The scientific target of Survey Strategy System (SSS) of LAMOST is to make an effective observation project, determine the observation time and arrange the observation process. Where to point to the position and how to allocate the fibers is the primary problem that SSS solves. In order to achieve the goal, mean-shift method is developed, which is proved by simulation and preliminary practice to be a feasible and effective method. Due to the process of making a valuable observation project for a survey observation with a large sky area, huge number of objects and long-term astronomical observation course, is more complicated and difficult. The restraint (transit restriction and focal plane and fiber location restriction) conditions should be considered carefully.
To exert the abilities of multi fibers of LAMOST as much as possible, the greatest density object-selection strategy is adopted in SSS. For an available sky area, a batch of observation object are stored in database and updated after every observe night. Looking for the max-density tile is major task of SSS. To achieve this goal, the maximum density algorithm, density gradient algorithm and the mean-shift algorithm are used in SSS.
Fiber allocation method
Given all the objects in a certain magnitude range in the field of view of the telescope and the location data of all the fiber cells, one can find the optimal solution to the problem in order to own a maximum match number so as to reach a high fiber utilization ratio. In order to reach this task, it is necessary to take focal design and fiber location restrain, sky light and flux standard observation restraints into account. We can use bipartite graph algorithm to shape the problem by describing objects and fiber as the two partite, the relation between a fiber and one of available object as one side connecting the two parts. Furthermore, these sides are marked with a PRI-related number generated by detailed algorithm. In this algorithm, the distance between the object and the fiber center, the priority and magnitude are all studied: the PRI stands for the observers demand on the object; in a single observation the dark or bright stars will be chosen with high priority; to avoid fiber collision, it is better to choose the stars that are closer to the fiber centers although it is not the necessary and sufficient condition for preventing fiber collision which is much more complex relating to the mechanical shape of every fiber cell and the arrangement of time of the overall situation. By using methods of bipartite graph, an optimal allocation will be found and additionally the function for manual allocating object for each fiber is provided.
The LEGUE survey is divided into three parts: the spheroid, the disk, and the anticenter.
We initially set out to solve the general spectroscopic survey target selection problem: starting with an input catalog of stars with any number of “observables”, define a general target selection algorithm that is capable of producing the desired distribution of targets. The assumption is that one begins with a large input catalog, where the number of sources is larger than the number of objects that can be observed with LAMOST. For every LAMOST field, a number of stars can be randomly selected as targets (based on how many fibers are available) among stars which are located in the field, and for which each was assigned a statistical weight.
Once the assignment probability for each star in the input catalog has been defined, a cumulative probability is calculated for each star in the list which consists of the sum of the probabilities of all stars in the list up to (and including) that particular star. A random number is then generated (using a uniform distribution from 0.0 to 1.0), and one star from the list is identified for which the random number is less than the cumulative probability, but greater than the cumulative probability of the previous element. This star is placed in the list of “selected” stars and removed from the sorted list of candidates for selection.
(For details, refer to http://adsabs.harvard.edu/abs/2012RAA....12..755C)
With the LAMOST, we will be able to complete, in a reasonable time frame (say, 5 years), surveys of galaxies and quasars which exceed existing surveys by a factor of 5∼10 in the number of objects and in the volume probed.
The target objects are selected in the 8000 square degrees in the North Galactic Cap (NGC) covered with the SDSS five band images and in the 3500 square degrees in the South Galactic Cap (SGC) with the imaging data expected to be provided by Pan Starrs. The u-band photometry necessary for the QSO selection in the SGC is planned to be taken by Chinese astronomers with the 2.3 meter Bob telescope at the Kitt Peak Observatory. The boundaries of the NGC and SGC sky areas are illustrated in Figure 1.
In order to study the environmental and evolutional properties of galaxies, we plan a galaxy survey down to an r band magnitude 19.5 to obtain high quality spectra for deter-mining physical properties of galaxies. Given the weather condition at the telescope site, we select about 2600 square degrees in SGC and some 800 square degrees in NGC, with the boundaries given in Figure 1. The selection of the sky areas is mainly determined by the availability of dark nights and by the availability of multiwavelength observations to a similar depth. This is the LAMOST Galaxy Deep Survey. For the areas of 8100 square degrees in the NGC and in the SGC other than that covered by the Deep Survey, we aim to observe redshift for all galaxies with r-band magnitude brighter than 19.0, which is the LAMOST Shallow Survey. The combination of the two surveys will yield a detailed picture of galaxy distributions from small to large scales. The structures on even larger scales will be probed by the LAMOST Early Massive Galaxy Survey which contains about 1 million massive galaxies with the idev brighter than 20.5. With these intrinsically bright galaxies,we can sample the large scale structures to redshiftz= 0.8. We will also include those massive galaxies with star formation, different from the Luminous RedGalaxies in the SDSS, to study the evolution of most massive galaxies, despite the fact that the number of these star forming EMGs is small. Our quasar survey will be a complete survey of quasars to i-band magnitudei = 20.5, which will produce some 600,000 quasars with 19.1< i <20.5 that the SDSS has not observed. In the meantime, we plan to combine the near-infrared imaging data of UKIDSS to uncover those dust obscured quasars that the SDSS could not find with their 5-band photometry data. From the quasar sample, we will have anaccurate determination of luminosity function and clustering for quasars atz≤3 and a new large sample of dust obscured quasars, thus providing a powerful database for testingthe unified model of AGN and for studying the co-evolution of supermassive black holes with their host galaxies.
(For details, refer to LAMOST Extra GAlactic Surveys—LEGAS, Proposed by the Working Group on LAMOST Extragalactic Surveys, June 11, 2009)