Perturbation analysis of job shop scheduling with multiple job priorities

Sensitivity analysis of scheduling parameters in job shops can be an intractable problem because of its exponential complexity and the fact that scheduling-related performance measures often lack an adequate topology. This paper addresses the application of perturbation analysis (PA) for scheduling problems with multi-priority service disciplines. By directly exploiting the special structure inherent in multiple priorities, we devise PA algorithms giving exact sensitivity estimates to a number of performance functions. The developed technique is based on a discrete event dynamic system type representation of the dynamics of schedules