The present work addresses the scheduling optimization problem in the context of online food delivery systems, focusing on strategic order aggregation and courier batch assignment under typical operational constraints. By representing the process as a variant of the Online Meal Delivery Scheduling Problem (MDSP), we analyze order characteristics, courier resources, and system objectives using classical scheduling theory terminology. Clustering orders via K-means and DBSCAN is formalized as batch scheduling for minimization of the makespan and total tardiness. Competitive analysis is applied to benchmark heuristic algorithms, with theoretical and empirical evidence supporting 21–42% improvement in total delivery time and 13–28% reduction in courier workload. Machine learning clustering integrates with online assignment to produce schedules approaching the lower bound of competitive ratio, ensuring resource-efficient, real-time order fulfillment.