Introduction of novel dispatching rules for grid scheduling algorithms

Grid scheduling problems are dynamic as the jobs and resources in the system vary overtime. The Rule-based scheduling algorithms are a new trend in grid scheduling which are applicable in dynamic grid environments; The arriving jobs waiting for execution is ordered according to a certain rule and they are dispatched for processing according to that order. Grid scheduling algorithms usually involve more than one objective function. In particular, it is not enough to minimize the Maximum completion time (Makespan) objective function without taking into consideration the lateness of the jobs (Tardiness). In this paper two novel dispatching rules are proposed that aims to minimize the maximum completion time of jobs (Makespan) and at the same time reduce the average lateness of these jobs (Tardiness). Furthermore, the proposed dispatching rules are combined with a local Tabu search algorithm that can be programmed to further optimize various grid scheduling objective functions. A complete set of comparison between the various combinations of techniques is presented based on simulation results obtained using Grid Sim.