The great majority of data centres are now underutilised, which is a concern. It is estimated that they are only used at 15% of their maximum capacity the great majority of the time. As a result of significant increases in server demand, additional space in these data centres has been added. This study presents the development of a hybrid paradigm for the efficient utilisation of cloud resources. Cloud resource provisioning is made more efficient through the use of machine learning and particle swarm optimization techniques. In order to determine if the model performs as intended, CloudSim is used to simulate the model's performance, and the results are compared against a range of performance measurements, such as the make span, delay, and throughput.