Research on the relationship between anomalous cases in parallel scheduling problems and executor performance

The article examines the problem of optimal task allocation with a partial order, aimedat minimizing the total execution time. It suggests that adding certain conditions, which in-tuitively should improve the result, may actually worsen the objective function.Anomalies arising from varying executor performance are analyzed. It is shown that anuneven task distribution can create "bottlenecks" and reduce efficiency, even if the execu-tors' productivity increases. Conditions for avoiding anomalies are proposed: prioritizedtask assignment, load balancing, and dynamic rescheduling.