1998 Fiscal Year Final Research Report Summary
Study on Optimization of Integrated Scheduling Problems for FMS
| Project/Area Number |
08650463
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
System engineering
|
|---|
| Research Institution | Kyoto Institute of Technology |
|---|
Principal Investigator |
KISE Hiroshi Kyoto Institute of Technology, Facultyof Engineering and Design, , Professor, 工芸学部, 教授 (10027807)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
OKURA Hiroyuki Kyoto Institute of Technology, Faculty of Engineering and Design, Associate Prof, 工芸学部, 助教授 (80135649)
|
|---|
| Project Period (FY) |
1996 – 1998
|
| Keywords | FMS / FMC / Robotic cell / AS / RS / Optimal scheduling / Branch-and bound algorithm / Heuristics / Markov process |
|---|
| Research Abstract |
The purpose of this study is to construct integrated models mathematically for scheduling problems that arise in FMS (flexible manufacturing systems) and to develop their optimal solution methods.
Atypical FMS consists of machining centers with many kinds of cutting tools, AGVs (automated guided vehicles), robots and anAS/RS (automated store and retrieval system). Especially, a small-sized FMS with a few machining centers is called FMC (flexible manufacturing cell), and a FMS with one robot a robotic cell.
The outline of the results obtained in this study is as follows.
(1) An efficient exact algorithm was developed for a FMC scheduling problem that has two machining centers, and an intermediate station between them for washing and cooling parts processed.
(2) Approximate algorithms that give more accurate solutions than previous ones were developed for robotic cell scheduling problems that have two or three machining centers, each having a finite buffer.
(3) An exact solution algorithm based on a branch-and-bound procedure was developed for a FMS scheduling problem with arbitrary number of machining centers. It was demonstrated that this algorithm has the capability of solving problem instances 10 times as size as ones the previous algorithms can solve.
(4) A new planing problem that arises in an AS/RS was formulated, and approximate algorithms were proposed based on the graph theory.
(5) A few mathematical properties were found for a Markov process that is a basic stochastic one of scheduling for probabilistic scheduling problems in FMS.
|
