Date of Publication :27th December 2017

Abstract: Task allocation is an important problem in grid/cloud environments in both research and applications. With the rapid development of grid/cloud environments, the features of openness and dynamism of environments put two new challenges to the development of task allocation approaches and strategies in such environments. Firstly, the participants in the environments normally have only local views about the environments due to the administrative independencies between the participants and the limited communication abilities of the participants. Secondly, task allocation methods/ approaches have to handle the dynamism and openness of the environments. In particular, task allocation methods/approaches have to respond to and be resilient from the unpredicted changes in the environments in a quick manner. In the proposed method, both consumers and providers only have local views about the environment. Consumers and providers trade with each other through negotiations in which they make their oer (count-oer) decisions strategically through taking the issues that they are concerned with into account. The experimental results show that the proposed method can achieve desirable performances in terms of the success rate of and prot obtained from the task allocation.

Reference :