Visual interactive simulation models for dynamic job shop rescheduling

In this research a visual, interactive simulation model is used to compare the performance of scheduling heuristics in a dynamically rescheduled job shop. A dynamic scheduling process means that the priority of jobs waiting to be processed in a work station is re-evaluated whenever the work station becomes available. The prioritization can be a function of job features or a function of the state of the job shop, or both. The complex behavior of dynamic rescheduling processes led to the use of animations and visualizations to augment the results of a formal design of experiments research methodology.; A 10 work station job shop with balanced, random routing of jobs was simulated for a variety of loads, scheduling heuristics, and maintenance policies. Classic heuristics such as AWINQ and MST were reformulated to use information on the state of the entire job shop. A new heuristic was uncovered called Preventive Maintenance-Break Down (PMBD) that performed quite well. The PMBD heuristic uses only information about a shop's work station failure history and maintenance program to sequence jobs. This completely state-dependent approach performed as well as SPT in flow time comparisons.; The results indicated that even moderate preventative maintenance can improve job shop performance. The load of jobs arriving to the shop and the particular performance feature of interest influence the selection of a scheduling heuristic. In general, the heuristics that used shop state information performed well.; The use of visual, interactive simulation proved extremely valuable as an analysis tool and as a communications aid. The visualization successfully transformed a process model of the job shop into an action space rendering of motion, color, and statistical details that were valuable contributions to this research. In particular, the animation of the mathematical algorithm for the scheduling heuristics improved the validation of the simulation model and offers a challenging opportunity for further research.