Hierarchical two-level game model for tasks scheduling in a desktop grid

Abstract

Desktop grids utilize idle computational resources within a research organization. Task scheduling is a very important issue for a desktop grid due to its significant impact on computational performance. In the paper we propose a game-theoretical model of task scheduling in a desktop grid. The model allows to consider the trade-off between the server load and the total time of computations. The numerical results based on a virtual drug screening project are presented, and the Nash equilibrium is found.