Study on ticket price forecast and task assignment in cloud computing environment

Study on ticket price forecast and task assignment in cloud computing environment

Huang of China University of Science and Technology

Specific work includes: 1 for the price management of aviation revenue, trying to dig out the law of ticket price change, using crawler technology to capture ticket data, based on the ticket price density image, put forward to decide whether to buy ticket cluster_predict_ticket algorithm. 2 the MapReduce version pcluster_predict_ticket algorithm of Cluster_predict_ticket algorithm is proposed for the Cluster_predict_ticket algorithm to deal with the low efficiency of large data flight data. The algorithm is ported to the Hadoop environment, and the algorithm efficiency is improved by using MapReduce programming mode. The experiments show that the pcluster_predict_ticket algorithm is more scalable and efficient without loss of prediction accuracy. 3 aiming at the problem of Hadoop task scheduling under heterogeneous environment, the problem of HTA (Hadoop task Assignment) is proposed, and the HTA problem is modeled using the minimum cost maximum flow. The Λ-flow algorithm is proposed to divide the original task allocation process into multiple rounds, each round is based on the current cluster status, and the execution of the previous task, dynamic task assignment, until all the task allocation is completed, in order to achieve optimal execution efficiency. Compared with other algorithms, the experiment shows that the λ-flow algorithm can better adapt to the dynamic change of the cluster and effectively reduce the execution time of the operation. The results of this paper verify the scalability of the pcluster_predict_ticket algorithm and the effectiveness of the λ-flow algorithm, effectively solve the problem of the efficiency of ticket price prediction under large data environment, and improve the execution efficiency of the Hadoop operation. In this paper, the research on the mapreduce of data mining algorithm is not only applicable to ticket data, but also can be extended to other data mining scenes, meanwhile, the work of this paper can provide some reference for task scheduling problem in cloud computing environment.

Study on ticket price forecast and task assignment in cloud computing environment