Task Scheduling Based On Improved Differential Evolution Algorithm In Cloud Computing

The rate of computer resource utilization and distribution of calculation ability arealways imbalance on the internet. Some applications require massive memory resourceand powerful computing, meanwhile there are numerous computing equipment andmemory resource sit idle. To realize the distributed share between resource andcomputational capabilities, and the quick-rise amount of network data, hence theemergence of cloud computing. Cloud computing developed from grid computing,parallel computing and distributed computing, it’s a rising business computing mode,the core concept of which is to distribute the calculation task onto the resource poolconstitute by many computer, enable the user to obtain computing power, memoryspace and information service on demand, resources and tasks in cloud are massive, so,the focal point and difficulty in cloud computing is to make full use of cloud resourcesto dispatch tasks efficiently.The article fully discussed the research background and current situation of cloudcomputing, based on systematic research of cloud computing technology, differentialevolution and classic scheduling algorithm, put forward a new multi-strategy mutationdifferential evolution of differential evolution based on modified differential evolutionalgorithm, the algorithm added category of normal distribution based on standarddifferential evolution algorithm, choose different mutation strategy according to thecharacteristics of each classification, difference vector individuals in mutation strategyuse roulette to choose based on individual similarity, improved DE algorithm’sshortcomings of slow convergence rate and local optimum tendency, could effectivelysolve combinatorial optimization problem. Finally, expand simulation platformCloudSim, and realized the algorithm in the article, after plentiful simulationexperiment and contrastive analysis, the result shows that this algorithm is able to achieve relatively short total tasks completion time and improve resource utilization.