Lead time predictions for a job shop
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
For a job shop, there has been very little research done on the accurate prediction of lead times, despite of the importance of and the advantages to be gained from the ability to predict lead times accurately. In job shop scheduling using integer programming method, etc., to produce fixed schedules, the lead time estimates can be obtained directly from the schedules. But none of these methods is of any appreciable use to industry. Heuristic job shop scheduling, on the other hand, has been implemented and performs satisfactorily. However, the means of predicting the lead times has to be formulated separately. This work investigates existing methods of predicting lead times, for a job shop employing heuristic scheduling. It evaluates their stability, system response rates, and their accuracy under steady state conditions. Before a method can be implemented for a real life job shop, it must be tested under dynamic conditions extensively, and found to be stable. The accuracy of the lead time predicted under such condition, must also be acceptable. Hence, the method with the best accuracy from the evaluation was subjected to such dynamic tests. The results of the tests showed that this method was stable under all the dynamic conditions tested, and predicted lead time with very good accuracy. A new version of this method is formulated. Testing under similar steady state and dynamic conditions, showed that it was superior to the version.