Vibratory rollers expose operators to a coupled vibration environment generated by terrain irregularity, drum excitation, frame motion, and low-frequency resonances of the cab and seat. This critical review examines the evolution of vibration-isolation optimization for vibratory rollers from passive rubber and hydraulic cab mounts to drum isolation, hydro-pneumatic and semi-active systems, negative-stiffness structures (NSS), quasi-zero-stiffness (QZS) seat suspensions, and intelligent control. A focused evidence base of 30 publications is synthesized, combining a long-term research line on single- and double-drum vibratory rollers with recent SCIE studies on nonlinear isolation and data-driven control. The reviewed work shows that passive mounts remain simple and reliable but are strongly limited by the conflict between load support and low-frequency isolation. Hydraulic and hydro-pneumatic mounts improve damping and resonance suppression, while semi-active control extends adaptability to changing operating conditions. NSS/QZS concepts further reduce effective dynamic stiffness and can substantially improve seat isolation, but their practical performance depends on payload, stroke, nonlinear stability, and the width of the low-stiffness region. Intelligent control and machine-learning methods offer a route toward condition-dependent tuning, although sensing requirements, model uncertainty, robustness, and vehicle-level validation remain unresolved. The main research gaps are coordinated drum-cab-seat optimization, multi-condition robust objectives, simultaneous vertical-pitch-roll whole-body-vibration criteria, and experimental validation under deformable terrain. A staged framework is proposed that links excitation identification, model validation, sensitivity screening, multi-objective optimization, robust verification, and vehicle testing.