Evaluation of multi-objective decentralized scheduling for applications in Grid environment

In grid environments applications require dynamic scheduling for optimized assignment of tasks on available resources, so the optimization represents a key solution for scheduling. This paper presents an evaluation of multi-objective decentralized scheduling models for the problem of task allocation. It also presents a survey of existing optimization solutions for grid scheduling. The surveyed scheduling solutions are: random and best of n random, exhaustive search, simulated annealing, game theory, ad-hoc greedy scheduler, and genetic algorithm for decentralized scheduling. We carry out our experiments with various scheduling scenarios and with heterogeneous input tasks and computation resources. We also present the methods to evaluate and validate the described scheduling methods. We present several experimental results that offer a support for near-optimal algorithm selection.