Game-theoretic, Market and Meta-Heuristics Approaches for Modelling Scheduling and Resource Allocation in Grid Systems

Task scheduling and resource allocation are the crucial issues in any large scale distributed system, such as Computational Grids (CGs). However, traditional computational models and resolution methods cannot effectively tackle the complex nature of Grid, where the resources and users belong to many administrative domains with their own access policies, users' privileges, etc. Recently, researchers are investigating the use of game theoretic approaches for modelling task and resource allocation problems in CGs. In this paper, we present a compact survey of the most relevant research proposals in the literature to use game-based models for the resource allocation problems and their resolution using metaheuristic methods. We emphasize the need of the translation of the traditional economical models into the game scenarios and the use of metaheuristic schedulers for solving such games in order to address the new complex scheduling and allocation criterions. We study the case of asymmetric Stackelberg game used for modelling the Grid users' behavior, where the security and reliability criterions are aggregated and defined as the users' costs functions. The obtained results show the efficiency of the hybridization of heuristic-based approaches with game models, which enables to include additional requirements and features into the computational models and tackle more effectively the resolution of the applied schedulers.