Green Cloud Data Center Based On Hybrid Evolutionary Algorithms Research On Resource Scheduling Optimization

After entering the 21 st century,the traditional data center has been difficult to meet the large amount of data computing and storage amount of data,and the data center has gradually extended from the industry of finance and communication operators to other industries.The rapid development of cloud computing and big data technology has brought disruptive changes to traditional data centers,and cloud computing data centers have emerged as a result.Cloud computing is a brand-new computing model with virtualization technology as the core.It combines a large number of computing resources of data centers in different physical locations based on network connections into a virtual pool.And other information services to achieve unified scheduling.The traditional data center is based on the unity of business and service.Its main focus is on application stability,operational reliability,and data security.However,it considers the utilization of computing resources,power consumption,and network congestion.less.With the development of the cloud computing industry,how to reduce energy consumption and network congestion in cloud data centers through resource scheduling optimization has become an important issue that needs to be urgently solved.A two-stage mathematical model for resource scheduling in a green cloud data center was established,and an adaptive and optimized resource scheduling strategy were proposed.In view of the complexity of the problem,in order to obtain a scheduling strategy that can meet the practical application within a reasonable range of time,energy consumption and congestion,a hybrid genetic algorithm based on an adaptive improved operator was designed.Among them: Aiming at the disadvantage that the conventional genetic algorithm is easy to fall into the local search and it is difficult to find the global optimization,an adaptive transformation operator was designed.Cross-over operator and mutation operator were selected to perform adaptive transformation to improve the convergence speed,stability and reliability of the algorithm,so as to reduce the energy consumption,network congestion and two-stage resources of the cloud data center’s onestage resource scheduling the purpose of scheduling the minimum completion time.The main research contents are as follows:(1)Establish a resource scheduling model for cloud data centersAnalyze and describe the resource scheduling problem in the cloud data center,and established a one-stage resource scheduling model with energy consumption and congestion as the target respectively,and refer to the two-stage resource scheduling with the minimum completion time as the maximum time queue for the most completed task Mathematical model and set relevant constraints.(2)Solving resource scheduling models for cloud data centersAn improved operator for adaptive crossover and mutation was designed.An adaptive improved genetic algorithm based on double tangent curve was proposed.First,for a onestage resource scheduling strategy,a conventional genetic algorithm,adaptive improved genetic algorithm and discrete Particle swarm optimization algorithm;then for two-stage resource scheduling,a conventional genetic algorithm based on Johnson rules and an adaptive improved genetic algorithm based on Johnson rules were designed.The orthogonal method was used to adjust the parameters of each algorithm to obtain the optimal parameter combination of each algorithm.(3)Design simulation experiment and result analysisAlgorithm simulation experiments show that the adaptive improved genetic algorithm proposed in the paper meets the expected goal of algorithm improvement.Among them,compared with the conventional genetic algorithm and discrete particle swarm optimization algorithm,the one-stage resource scheduling adaptive improved genetic algorithm can avoid the algorithm being premature,and also improve the stability and reliability of the algorithm;Compared with the conventional genetic algorithm based on Johnson rule,the improved genetic algorithm based on Johnson rule has effectively improved the minimum completion time of the task queue.In summary,the research done can provide theoretical support and service support for related algorithms for the current cloud data center resource scheduling,which can reduce the energy consumption and congestion in the first stage of resource scheduling in the cloud computing environment.The minimum completion time of the task queue in the phased resource scheduling,thereby reducing the overall energy consumption and improving the service quality of the green cloud data center.