The demand-supply imbalance at highly congested (level 3) airports is predominantly managed through the World Airport Scheduling (WASG) guidelines. To provide efficient solutions at congested airports, current research has proposed multi-objective, Mixed Integer Programming (MIP) models, which often result in complex large-scale formulations that cannot be solved in practical computational times, hence requiring the proposition of heuristic algorithms. This paper presents a multi-objective solution framework for supporting the allocation of slots at large coordinated airports. The algorithm allocates requests of higher priority by solving hierarchically a series of MIP models. As the capacity of the airport under consideration becomes saturated, requests of lower priority are divided into subgroups based on an unsupervised learning technique that considers each request’s airport capacity utilisation. A genetic algorithm determines the allocation sequence of the generated subgroups and constructs multiple schedules. Our tests suggest that the algorithm exploits the synergies between Operations Research (OR) and Machine Learning (ML) methodologies and can cope effectively with large airport instances and generate multiple schedules.