Novel Visualization XAI Approach for Assisting Schedulers' Comprehension and Interpretation of ACO Applications in Job Scheduling

Bio-inspired algorithms, such as ant colony optimization (ACO) and others, have been applied to job scheduling. However, ACO applications were viewed as black boxes that were difficult to understand, trust, and accept. To address this issue, this study proposes several novel visualization XAI techniques, including color-encoded disjunctive subgraphs (CEDS), intrinsic modeling decision tree (IMDT), and contrast gradient saliency maps, to help schedulers comprehend and interpret the operation process and results of ACO applications in job scheduling. By focusing only on good feasible solutions only, the CEDS of a job scheduling problem can be constructed (and reconstructed) very quickly, thereby facilitating its manipulation. In addition, it also reveals that most ants end up following similar paths. In addition, the contrastive gradient salient map precisely locates the convergence point of the ACO evolution process and intuitively displays it to the scheduler. Furthermore, IMDT fills the gap in explaining the intrinsic mechanisms of blackbox job scheduling methods such as ACO.