Mixed-model assembly lines must adapt to evolving product families across multiple generations while maintaining operational efficiency under uncertain conditions. This work addresses the robust design and reconfiguration of assembly lines where processing times and reconfiguration costs are subject to uncertainty. We extend existing scenario-tree formulations by applying robust optimization to handle operational variability without exponential growth in problem size. Our hybrid approach combines stochastic programming for structural uncertainty (task evolution) with robust optimization for operational parameters (processing times and costs). First computational experiments demonstrate that appropriate robustness settings significantly reduce cycle-time violations while maintaining reasonable robustness cost and computational complexity.