Research On The Virtual Organization Establishment For The Random Task In Cloud Manufacturing

With the development of the global economy and the rapid improvement of the Technology, the manufacturing industry is facing more and more pressure. The highly tense competition, the rapid renovation and the prompt innovation demand the enterprise to have stronger ability to exist. In this condition, manufacturing industry is badly in need of a new mode to adapt the new market environment. Along with the concept of the Agile manufacturing(AM) widely recognized by people and the rapid development of the Internet technology, a new manufacturing mode emerged which is called “Cloud Manufacturing(CM)”. It combines the disrupt equipment resources and manufacturing ability by Internet which can produce more benefit than any other manufacturing modes. Nowadays, CM is palying a more and more important role in the manufacturing industry in China. Many manufacturing enterprises, especially the medium and small ones, frequently confront the problem of lacking ability and if they want to seize the chances in the market environment which changes rapidly, they have to take the CM mode which means using the Cloud Service(CS) to build a virtual organization(VO) to become more competitive and win the competition. This paper focuses on the method of how to build a VO in the CM environment and mainly contains the following three aspects.(1) The queue model for the random tasks is buildBefore building the VO, the size of the ability which needs to be extended should be certain. At present, the researches mostly focus on analyzing the tasks to ensure the size and then building the VO to get the ability after the tasks have arrived. The process will waste some time which disables the enterprise to finish the tasks on time and lose chances in the rapidly changing market environment. The paper makes innovations to describe the tasks which have not arrived with probability distribution and build the VO in advance which make the tasks better finished when it really comes. The task is divided based on the N different kinds of ability, and then a model for the N-step tandem queue model is built. We can get the size of the ability which needs to be extended, namely the ability requirement of the VO.(2) The algorithm for evaluating the single CS wiht the fuzzy indexes is designed.When evaluating the fuzzy attribute of the CS, it often can not be described with a precise value. Given this circumstance, in this paper, the triangle fuzzy numbers are used to describe the fuzzy attribute and fuzzy indexes are set to evaluate the CS. Unlike the precise values, it is not possible to compare the fuzzy numbers directly, and we should set a rule which suits the real condition to compare them. By integrating the requirement for the stability into the evaluation, the paper proposes a new algorithm to compare the fuzzy numbers based on the traditional algorithm and the variance. With the method, it can effectively evaluate the CS with fuzzy indexes which is greatly meanful.(3) The CS combination and optimization with QoS restriction is designed.When the enterprise needs many kinds of CS to meet its demand, the CS combination and optimization become a key question. In this paper, the relationship between the CS and the Cloud Enterprise is analyzed and a model with two layers for the CS combination and optimization is built. Based on the model, a method for solving the optimal CS combination is proposed.In the end, a series of experiments are carried out to validate the effectiveness of the proposed method in terms of efficiency and effect.