Abstract: Job scheduling is a key challenge in cloud computing due to the large number of resources and services involved. The choice of scheduling strategy strongly affects system performance, and poor decisions can even cause failures. Since testing directly on real cloud systems is impractical, researchers rely on simulations to develop and evaluate scheduling algorithms. However, generating simulation input data manually is time-consuming and can lead to inconsistencies, making fair comparison between algorithms difficult. This study makes three key contributions. First, we introduce a tool that automatically generates simulation data for CloudSim Plus. Second, we use this tool to generate a dataset that is employed to evaluate our proposed algorithm; this dataset is also published publicly to enable other researchers to use it as a benchmark for testing and comparing their algorithms with ours. Finally, we propose a job scheduling algorithm based on a hybrid approach that combines a Genetic Algorithm with Particle Swarm Optimization (PSO). The proposed algorithm was evaluated on both synthetic and real datasets where it demonstrated competitive performance compared to existing approaches in the field.

Key words: benchmark dataset, cloud computing, CloudSim plus, genetic algorithm, PSO, job scheduling, resource allocation, simulation data generation