Greedy and metaheuristics for the offline scheduling problem in grid computing

In grid computing a number of geographically distributed resources connected through a wide area network, are utilized as one computations unit. The NP-hard offline scheduling problem in grid computing consists of assigning jobs to resources in advance. In this paper, five greedy heuristics and two metaheuristics for solving the offline scheduling problem are introduced. Computationally evaluating the heuristics shows that all heuristics find useful solutions with a gap of 20\% between upper and lower bounds. The metaheuristics give better results than the greedy heuristics, but also have larger time usage. All heuristics solve instances with up to 2000 jobs and 1000 resources, thus the results are useful both with respect to running times and to solution values.