Longitudinal and lateral motion planning poses a significant challenge to achieving full autonomy in automated vehicles. This work focuses on studying the motion planning problem for automated vehicles specifically in a highwaymerging scenario. The problem is modeled as an infinite horizon optimal control problem, taking into account finite control sets for the ego agents and uncontrolled state components of surrounding traffic. For this type of control problem, obtaining a real-time solution that meets both high safety and efficiency requirements can be difficult. In this study, we employ the rollout approach, which involves online optimization following the simulation of a known baseline policy instead of relying solely on extensive offline training. We compare the performance of one and multistep lookahead rollout algorithms against several state-of-the-art benchmark policies in simulation. The simulation results indicate that the rollout algorithm significantly enhances safety while simultaneously maintaining a high average speed within the merging scenario. Furthermore, we conduct simulation studies to assess the rollout methods in adapting to varying behaviors of surrounding vehicles. Additionally, we investigate the impact of different horizon settings and the introduction of terminal cost approximation.