Modern transportation systems are increasingly challenged by rising diversity and density in urban traffic, driven by technological advancements and increased travel demands. Dynamic lane allocation provides a pivotal solution by adapting lane usage based on real-time traffic conditions to enhance urban mobility and mitigate congestion. This research introduces a novel dynamic lane allocation strategy that allows ride-sharing vehicles access to dedicated bus lanes during select periods and locations to alleviate urban congestion. We use the MnMS (Multimodal Network Modeling and Simulation) framework, which employs a trip-based Macroscopic Fundamental Diagram (MFD) to support diverse travel options and accelerate computation. Additionally, this study integrates a Local Correction Factor (LCF) that refines speed estimations on individual links for more accurate and realistic route choices. In this study, a Large Neighborhood Search (LNS) optimization solver is employed to efficiently solve complex problems by exploring extensive solution neighborhoods iteratively. Experiments on a 6x6 Manhattan grid network show that optimized dynamic lane allocation significantly outperforms static lane usage, particularly under high demand, reducing total travel time by approximately 15.3%. The proposed approach is identified as crucial for urban traffic management, offering substantial potential to reduce congestion and enhance mobility in densely populated cities.