Research And Implementation Of Cloud Computing Task Scheduling Algorithm Based On Improved Genetic Algorithm

In cloud computing,task scheduling algorithm is one of the key factors for service providers to reduce costs and improve service quality.How to use scheduling algorithm to efficiently allocate computing resources to minimize the scheduling time and cost are an important issue in cloud computing research.At present,most of the improved genetic algorithm for cloud computing task scheduling is effective,but there is still a lot of room for improvement,which can do better in the cloud computing task scheduling.Therefore,this thesis studies how to improve the efficiency of cloud computing task scheduling,shorten the task completion time and reduce the costs by improving the genetic algorithm.The specific improvements of genetic algorithm are as follows:1.There are mapping errors in the binary coding of genetic algorithm when the continuous function is discretized.Especially when the length of individual chromosome coding is short,it may not meet the corresponding accuracy requirements.When the length of individual chromosome coding is too long,the search space of the algorithm will increase sharply,and the performance of the whole genetic algorithm will be reduced.In order to overcome the shortcoming of binary coding,in this thesis use a method of combining two coding methods,which use binary coding in the optimal length range of binary coding chromosomes and floating-point coding in the beyond range.2.In the early stage of evolution,the crossover and mutation probability of adaptive genetic algorithm is too low,which easily leads to the shortage of local optimum.To overcome this shortcoming,a parameter k~n related to evolutionary generation is added to the crossover probability to reduce the probability of the algorithm falling into the local optimal solution.The improved genetic algorithm is used to optimize the task scheduling of cloud computing,and a comparative experiment is carried out.The experimental results show that the improved algorithm is better than adaptive genetic algorithm(AGA)and simple genetic algorithm(SGA)in total task completion time and load balancing.Compared with double fitness genetic algorithm(DFGA),although the total task completion time is slightly insufficient,but it is better in load balancing.