We investigate a novel satellite avoidance strategy to mitigate the impact of large commercial satellite constellations in low-Earth orbit on the Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST). We simulate the orbits of currently planned Starlink and OneWeb constellations (∼40,000 satellites) to test how effectively an upgraded Rubin scheduler algorithm can avoid them, and assess how the overall survey is affected. Given a reasonably accurate satellite orbit forecast, we find it is possible to adjust the scheduler algorithm to effectively avoid some satellites. Overall, sacrificing 10% of LSST observing time to avoid satellites reduces the fraction of LSST visits with streaks by a factor of 2. Whether such a mitigation will be required depends on the overall impact of streaks on science, which is not yet well quantified. This is due to a lack of adequate information about satellite brightness distributions as well as the impact of glints and low surface brightness residuals on alert purity and systematic errors in cosmological parameter estimation. A significant increase in the number of satellites or their brightness during Rubin Operations may make implementing this satellite avoidance strategy worthwhile.