Research On Task Scheduling Algorithm For High Performance Computing Based On Genetic-Ant Colony

With the advancement of science and technology,high-performance computing has been widely used in fields such as climate simulation,fluid mechanics,molecular dynamics,and biological information.However,in high-performance concurrency,multi-computing system models,data and cloud storage in high-performance computing,the data processing speed and user demand response time cannot be effectively improved.How to allocate resources and load balance the high-performance computing system platform is the core of enhancing performance.In order to improve the utilization of high-performance computing system and reduce the imbalance of system load,this paper proposes a high-performance computing task scheduling algorithm based on genetic-ant colony(GA-LBFACO)by analyzing the characteristics of ant colony algorithm and genetic algorithm.The specific research work is as follows:(1)For the ant colony algorithm in the process of solving high-performance computing task scheduling,only the path pheromone between the servers is considered,and the load status of the servers is ignored.This paper improves the pheromone and heuristic information by redefining the ant colony algorithm.The server load factor is introduced to provide dynamic changes in the server load,and the initial pheromone concentration of the server load factor is added to the transition probability heuristic information,and then the pheromone is updated by adding the pheromone conversion factor.The improved ant colony algorithm can effectively reduce the server load and improve the algorithm's optimization ability.Experiments have proved that the improved ant colony algorithm(LBFACO)is effective in reducing load balancing.However,there are still shortcomings in terms of task execution time.The main reason is that the acquisition speed of pheromone in the early stage of ant colony algorithm is slow,resulting in insufficient pheromone.(2)Aiming at the shortcomings of ant colony algorithm,this paper proposes the idea of combining genetics and improved ant colony algorithm.By studying the characteristics of genetic algorithm,genetic algorithm has better solution space search ability in the early stage,and can quickly search for excellent solutions.The excellent solution searched by genetic algorithm is transformed into the initial pheromone of the improved ant colony algorithm,and the improved ant colony algorithm is adopted in the later stage.The ability to search for precise solutions ultimately results in efficient algorithm convergence speed and global searchability.Experiments have proved that the fusion algorithm GA-LBFACO not only reduces task execution time and server load balancing,but also effectively improves CPU utilization.