Multidisciplinary design optimization of transonic rotors, especially focusing on improving the aeroacoustic performance, has received little attention. This article employs an automatic design optimization approach to reduce the shock-associated tone noise and to increase the adiabatic efficiency of NASA rotor 37. An evolutionary algorithm for multi-objective trade-offs searching is used to obtain the noise-efficiency Pareto front. Three-dimensional Reynolds-averaged Navier–Stokes simulations are directly applied in the optimization loop for aeroacoustic and aerodynamic analysis. Significant noise reduction and efficiency increase have been achieved. Two optimized designs corresponding to the endpoints of the Pareto front have been analyzed comprehensively from the perspectives of rotor characteristics, noise performance, and flow and acoustic fields, to understand the mechanisms responsible for performance improvements.