A two-pass scheduling method using discrete event simulation
The scheduling area is well known to generate large combinatorial problems and several modelling techniques have been used to tackle them. One of the most important techniques is Discrete Event Simulation (DES). However, as any other technique DES has its advantages and disadvantages. The main advantage of DES as a scheduling tool is the computational efficiency with which an accurate model of a manufacturing system can be developed. The disadvantage of DES as a scheduling tool is that a schedule is developed through incremental decisions over time, but without appreciation of the effect of each decision on the rest of the schedule. This thesis presents a two-pass finite scheduling method to alleviate the inherent disadvantage of DES as a scheduling tool. The two-pass method is based on two programs which employ a two-phase simulation model. The first pass produces schedules by simple heuristic dispatching rules (first in first out, earliest due date and slack time divided by the number of operations remaining). The second pass of the method tries to improve the schedule based on information from the first pass and a heuristic algorithm designed to improve the production flow. Results from the experiments carried out to test the effectiveness of the method are also presented.