流程型工廠之群組排程研究

The Investigation of Group Scheduling Problem on Flowshop

陳宗明(Tsung-Ming Chen)

指導教授 ： 張百棧

元智大學/工程學院/工業工程與管理學系/碩士(1992年)

摘要

在應用群組技術(Group Technology,GT.)而形成的單元製造系統 (
Cellular Manufacturing System )中，其生產環境及製造形態皆與傳統
工廠有所不同，在生產排程方面；由於工件必須以群組的形態來加以處理
，所以必須應用適合此特殊生產形態的排程技術，而此種與GT配合的生產
排程便稱為群組排程(Group Scheduling)，由於群組排程牽涉到二層面(
Two Phase)的問題，不僅要決定各群內工件的加工順序且同時也必須決定
群與群之間處理的先後順序，由此可知群組排程比一般傳統排程的問題更
為複雜，本論文主要在探討以最小總流程時間為目標之流程型群組排程問
題，其內容如下所示： 1.將現有之各種群組排程下限值演算法加以改進
，提出三種較有效率之新下限值演算法，並利用實驗分析來探討這些下限
值的優劣。 2.在啟發解上不但將一般傳統排程的演算法應用到群組排程
的問題來相互比較；同時也嘗試著根據群間排程的特殊處發展另一種與現
存數種觀念不同的演算法，並利用實驗分析來與其它現存之各種演算法做
比較及研究。

關鍵字

群組技術 ； 彈性製造單元 ； 群組排程

並列摘要

Recently, Group Technology (GT) has been widely applied in
industries and thus a new type of manufacturing system called
cellular manufacturing system (CMS) is formed. The
manufacturing environment in CMS is different from conventional
factories, especially in production scheduling, since jobs have
been preprocessecd and grouped within different part family,
therefore, the scheduling problem within such a system is
totally different from the conventional scheduling problem and
further development of applicable scheduling algorithm is
needed. This new type of scheduling problem combined with the
GT is called group scheduling (GS) problem and it involved two
different phases : one is to decide the processing sequence of
each group and the other is to sequence each job within each
group. Therefore, the computational complexing of the GS
problem is even complicated than the conventional scheduling
problem. This research investigated the GS problem on the flow
shop environment with the objective of minimizing the makespan.
We improved the lower bound proposed by earlier researchers and
provided three different algorithms for the calculation of
lower bound. Experimental design was also applied to further
analyze the differences among these bounds. In the development
of heuristic algorithms, we are not only applied the algorithm
used in conventional factories to the group scheduling problem,
among groups. At last, we also made some conclusions based on
the experiment applied to the algorithms mentioned above.

並列關鍵字

Group Technology ； Cellular Manufacturing System ； Group Scheduling