Dbr Based Scheduling Method For Re-entrant Manufacturing System

Re-entrant manufacturing system is a class of complex productionsystem abstracted from semiconductor, film and multi-layer printedcircuit boards manufacturing systems. With the development ofinformation technology and microelectronic industry, scheduling problemfor re-entrant manufacturing system is an important research branch inthe area of control science and engineering. A large number of productionscheduling methods, although cannot meet the need of solving practicalproblem, have been presented in literatures thus far. DBR (Drum-buffer-rope) which poses a novel way to scheduling combines material releaseand production control to solve bottleneck problem. The main content ofthis paper is the scheduling method of re-entrant manufacturing systemand the application of DBR method.(1) The Theory of Constraint and the basic DBR based schedulingmethod are introduced. Then, the key techniques of implementing DBRmethod are summarized. Also the problems of the application with DBRand the corresponding solutions are analyzed according to the complexfeatures of re-entrant system in this work. (2) The extended DBR based scheduling algorithm is proposed forthe manufacturing system with the characteristic of re-entrant flow inbottleneck equipment, which maximize the effective capacity ofbottleneck and balance the load of production line. This algorithm makesthe right scheduling decision with the consideration of the correspondingreal-time information of the production line. The effectiveness andsuperiorities are demonstrated through the simulation.(3) The rush order forecasting combined DBR method is proposedfor the problem that the urgent orders in manufacturing system are alwaysinfluenced by producing statement and will impact the regular orders’performance. The ANFIS (Adaptive Neuro-fuzzy Inference System) isfirst applied to build a prediction model for the related information of thecoming rush order. Then the fuzzy inference system is used to decide thecorresponding parameter of the previous DBR method. After that, theproduction line is adjusted before the rush order comes, which will reducethe inference of both the regular order and rush order.