학술논문
클린룸에서 조립장비의 공간배치 및 일정계획에 관한 연구 / Assembly scheduling and work space allocation of electronic equipment in the clean room
Document Type
Dissertation/ Thesis
Author
Source
Subject
Language
Korean
Abstract
현재 LCD나 반도체장비와 같은 제조 장비는 주문의 생산일정과 납기일을 고려하여 청정의 작업공간인 클린룸에서 장비를 생산하고 납품하게 된다. 클린룸은 한정된 공간을 할당하여 만든 곳으로 그 범위를 임의의 영역으로 확장할 수 없다. 기존의 클린룸에서의 작업별 주문배치는 스케줄링 전문가의 노하우를 바탕으로 하여 수많은 주문들을 임의로 배치하였다. 그러나 노하우를 바탕으로 영역을 할당하기 때문에 최적의 계획수립을 보장할 수 없었으며 수작업으로 진행되기 때문에 많은 시간이 소요되었다. 이러한 수주기반의 주문배치문제는 “서로 다른 납기를 갖는 작업에 대한 이종병렬기계에서의 일정계획수립”, “반복공정을 가지는 제약적 병렬기계에서의 일정계획수립”, “제품별 배치 하에서 여력공정의 활용을 위한 일정계획연구” 등에서 다양한 알고리즘이 연구되었다. 본 연구는 클린룸을 크기별로 일정한 영역으로 나누고, 각 영역 안에 배치되는 주문을 임의로 생성해 주문의 납기일과 소요일을 고려하여 평균납기일을 최소화시키는 목적을 갖는 알고리즘을 연구하는 것이다. 실험 결과 본 연구에서 제안하는 방법은 생산라인에 작업을 할당하는 이종병렬기계문제에 적용가능하며 평균지연시간을 줄여야 되는 곳인 대량주문처리 시스템에 적용시켜 할당문제를 푸는 곳에 응용되어 활용이 가능하다.
We study a real assemble-to-order system in Korea in which electronic equipments for manufacturing e LCD or semiconductor products are assembled in the pure clean room. Each order has a distinct due date and requires a work space for assembling equipments related to the order. The clean room is already divided by several works spaces with three kinds of sizes, i.e. large, medium and small sizes. Equipments are also classified into three categories: small-, medium- and large-sized equipments according to their sizes. A work space for assembling an equipment should be not less than the size of the equipment. A production manager should determine an assembly schedules for orders and allocate work spaces to orders in the clean room. He is currently doing this job by hand using his experience and know-now. However this manual labor requires huge amount of time and does not guarantee the best work plan which satisfies due dates of orders to the maximum. In this study, we show that heuristic algorithms for the dedicated machine scheduling problem can be applied to our problem after a slight modification. Also, we propose a heuristic algorithm for assembly scheduling and works space allocations to orders. We perform computational experiments to test performances of the algorithms using randomly generated test instances. Test results show that the proposed heuristic algorithm performs much better than the algorithms for the dedicated machine scheduling problem.
We study a real assemble-to-order system in Korea in which electronic equipments for manufacturing e LCD or semiconductor products are assembled in the pure clean room. Each order has a distinct due date and requires a work space for assembling equipments related to the order. The clean room is already divided by several works spaces with three kinds of sizes, i.e. large, medium and small sizes. Equipments are also classified into three categories: small-, medium- and large-sized equipments according to their sizes. A work space for assembling an equipment should be not less than the size of the equipment. A production manager should determine an assembly schedules for orders and allocate work spaces to orders in the clean room. He is currently doing this job by hand using his experience and know-now. However this manual labor requires huge amount of time and does not guarantee the best work plan which satisfies due dates of orders to the maximum. In this study, we show that heuristic algorithms for the dedicated machine scheduling problem can be applied to our problem after a slight modification. Also, we propose a heuristic algorithm for assembly scheduling and works space allocations to orders. We perform computational experiments to test performances of the algorithms using randomly generated test instances. Test results show that the proposed heuristic algorithm performs much better than the algorithms for the dedicated machine scheduling problem.