Genetic Algorithm based Task Scheduling in High-Performance Cluster Computing

In recent years, the demand for increased computational power has driven advancements in both High-Performance Computing (HPC) systems and multi-core technology. As these technologies play an important role in enhancing the power computation capabilities of different resources, traditional approaches for task scheduling in HPC had faced challenges like inefficient resource allocation and high task waiting time. Therefore, this research proposes Genetic Algorithm (GA) which refines the core scheduler for each task and enhances the configuration of task scheduling strategy. In core task scheduling, GA is utilized for minimizing average scheduling time by optimally allocating tasks across computing nodes. The GA operates through evolutionary processes by including population initialization, selection, crossover, and mutation which iteratively refined task allocation solutions. The proposed GA effectively reduced task waiting time, improved system throughput and achieved better results in terms of resource utilization (90.27%) when compared with existing Particle Swarm Optimization (PSO) respectively.