After the successful Gravity Recovery and Climate Experiment (GRACE) mission, which ended in 2017, GRACE Follow-On was launched in 2018. Together, these missions have provided almost two decades of near-continuous information on the Earth's time-varying gravity field. Single pairs of satellites like GRACE and GRACE-FO are inherently limited in their spatio-temporal coverage, spatially to a few hundred kilometers and temporally to a roughly monthly resolution. In order to increase the global spatio-temporal resolution and therefore allow for the determination of submonthly time-varying gravity field events, a constellation of GRACE-type pairs has been studied. 
Small satellite instrumentation is becoming increasingly affordable, reliable and precise. This will soon allow a constellation of GRACE-type small satellites to be deployed. In this work we investigate the performance of such a constellation for different numbers of satellite pairs using simulation studies which include different orbital con¦gurations. We design constellations using an evolutionary algorithm approach to optimize spatial and temporal resolution given a set of N satellite pairs. This allows us to evaluate the improved spatio-temporal performance, and thus the science return, that might be gained from such future mission architectures.