Dynamic workloads in cloud computing create challenges for efficient task scheduling due to unpredictable resource demands. Traditional static scheduling methods are often inadequate in such environments. To overcome these challenges, dynamic task scheduling provides a flexible method that enhances both system performance and resource efficiency. Therefore, a novel Deep Mobile Network (DeepMobileNet) is devised for dynamic task scheduling in cloud environments. Initially, multi-objective functions are utilized for task scheduling within cloud environments that are evaluated based on reliability, energy, actual task running time, cost, resource utilization, makespan, earliest start time, memory capacity, task priority, earliest finish time, and throughput. Finally, task scheduling is done by the proposed hybrid DeepMobileNet architecture that integrates MobileNet and Deep Belief Network (DBN). Moreover, scheduled tasks are assigned to Virtual Machines (VMs) and servers to complete all tasks as efficiently as possible. The performance of the proposed hybrid DeepMobileNet architecture is assessed based on factors such as makespan, energy consumption, and resource utilization. The proposed hybrid DeepMobileNet architecture attained a makespan of 10.999 s, energy of 0.259 kwh, resource utilization of 95.347
