Evaluating the Robustness of the Alienated Ant Algorithm in Grids

Grid systems are developed in very dynamic and heterogeneous environments, characterized by an high level of unpredictability. Some system performance, in fact, may vary due to sudden failures or to the occurrence of circumstances unforeseeable deterministically (e.g. failure of the machines, load higher than the expected and inaccuracy in the estimation of parameters). The presence of a robust scheduling algorithm allows to face such perturbations, avoiding degradation of system performance. The paper aims to evaluate the robustness of the Alienated Ant Algorithm, a recently proposed metaheuristic and successfully applied to scheduling in distributed and highly dynamic environments such as Grid, subjecting the algorithm to stress conditions to assess its performance.