Single object observations: Large telescopes vs. multiple small telescopes
A significant fraction of large-telescope observing time is devoted to single-object spectroscopy and photometry. We compare the cost-effectiveness of different telescope architectures, including single seeing- and diffraction-limited telescopes and arrays of seeing-limited or diffraction-limited telescopes. We use the survey speed per unit cost as the performance metric. For single telescopes with equal collecting area, diffraction-limited operation becomes more cost-effective once seeing-limited observations enter the background-dominated regime, typically at visible magnitudes of $\sim20$. For unresolved single-object observations, arrays of seeing-limited telescopes are generally more cost-effective than a single seeing-limited telescope. When matched in survey speed, a single diffraction-limited telescope outperforms an array of small seeing-limited telescopes only for targets fainter than $\sim20$--$24\,$mag, while a single large diffraction-limited telescope becomes more cost-effective than an array of smaller diffraction-limited telescopes only for targets fainter than $\sim26$--$28\,$mag. We conclude with a discussion on aspects related to the minimum telescope size and maintenance of telescope arrays. These conclusions are sensitive to the specific assumptions and system parameters adopted for each architecture and should therefore be interpreted with appropriate caution. Our results motivate the development of telescope arrays such as MAST and LFAST and may inform the design of future optical facilities.