The Research Of Data Warehouse Etl Distribution And Scheduling Model

In recent years, with the development of society, competition between enterprises is becoming increasingly fierce. So enterprises have to improve the competitiveness of the company for its own survival and development. It is the key to business success depends on the correct business decision or not. Therefore, it is growing that enterprise dependent on the decision support system. Enterprise have called for the establishment of data warehouse to store, analyze, process data, and ultimately provide the basis for the corporate decision-making. Data warehouse system in the process of handling data, often dealing with the relationship between the data precedence constraints. This process is a data warehouse in one of the most challenging. This article is a precedence constraint to deal with this multi-task scheduling problem.This paper analyzes the research situation, and introduces related knowledge, and then constraints on task scheduling problem in making the required number. With a directed acyclic graph that the relationship between tasks, a mathematical model. In solving the model, the presented popular method, and we compare the advantages and disadvantages of each method. Finally, we select genetic algorithm to solve this model.In the process of genetic algorithm, task, the processor that are encoded. It has precedence constraints between the tasks. The individual by encoded may not satisfy the constraint, so this paper adjusts the result by a reverse adjacency list. Depending on the requirements of the shortest, and giving the corresponding fitness function. When making individual choices, using roulette wheel method and the best combining individual preservation. According to the individual on the task of encoding processors, respectively, cross-individual variation also used different strategies. At last, we confirm the validity of the algorithm by the simulation experiment...