Effectual template banks for upcoming compact binary searches in Advanced-LIGO and Virgo data

Abstract

Recent discoveries of gravitational wave (GW) signals from astrophysical compact binary systems of neutron stars and black holes have firmly established them as prime sources for advanced GW detectors. Theoretical templates of expected signals from such systems have been used to filter the detector data using the \emph{matched filtering} technique. An efficient grid over the parameter space at a fixed minimal match has a direct impact on improving the computational efficiency of these searches. We present the construction of three-dimensional template banks (in component masses and an effective spin parameter) by incorporating several new optimizations to the hybrid geometric-random template placement algorithm that we proposed recently. These optimizations allow us to create more efficient template banks in future compact binary searches by shrinking the hybrid banks by $\sim 34\%$ in comparison to the basic algorithm. Such optimized banks are also found to be $\sim 22\%$ smaller than the optimized stochastic bank constructed over a nominal range of parameters. We also construct an explicit hybrid template bank with parameters identical to the `uber-bank' used in the recently-concluded CBC searches in the second observation run of the Advanced LIGO and Virgo detectors. We demonstrate a reduction of more than 53,000 templates over the stochastic template bank at a near-identical coverage as determined by fitting factor studies. A computationally efficient technique for semi-numerical calculation of the parameter space metric, applicable for aligned-spin waveform family, is also outlined. The resulting hybrid template banks can be generated much faster in comparison to the stochastic banks, and are ready to be used in the upcoming observation runs of Advanced LIGO and Virgo.