Economic-Based Modelling for Resource Scheduling in Grid Computing

In decentralized computational environments applications or services belong to different organizations with different objectives. Grid computing model is appropriate for such environments, in particular for large scale computations. In this paper, decentralization is modeled in terms self-interested, economically inspired agents that attempt to achieve their goals. The objective of this work is to model the Grid scheduling problem based on the nature of the Grid environment and the characteristics of the Grid entities focusing on the resource utilization and completion time objectives. We create a mapping between the Grid scheduling problem and the combinatorial allocation problem and propose an adequate economic-based optimization model and a bidding language for the Grid scheduling problem. The proposed approach is being validated through a prototype implementation using the Coordinated Intelligent Rational Agent (CIR-Agent) model.