In the dispatch strategies of ride-sharing platforms, reinforcement learning can simultaneously consider multiple objectives such as passenger waiting time, driver income, and road load, rather than optimizing only a single dimension.