Timed Petri Net Based Scheduling for Mechanical Assembly Integration of Planning and Scheduling

Akio INABA  Fumiharu FUJIWARA  Tatsuya SUZUKI  Shigeru OKUMA  

Publication
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences   Vol.E81-A   No.4   pp.615-625
Publication Date: 1998/04/25
Online ISSN: 
DOI: 
Print ISSN: 0916-8508
Type of Manuscript: Special Section PAPER (Special Section on Concurrent Systems Technology)
Category: 
Keyword: 
scheduling,  timed Petri net,  assembly,  

Full Text: PDF>>
Buy this Article




Summary: 
In scheduling problem for automatic assembly, planning of task sequence is closely related with resource allocation. However, they have been separately carried out with little interaction in previous work. In assembly planning problem, there are many feasible sequences for one mechanical product. In order to find the best assembly sequence, we have to decide the cost function for each task a priori and make decision based on summation of costs in sequence. But the cost of each task depends on the machine which executes the allocated task and it becomes difficult to estimate an exact cost of each task at planning stage. Moreover, no concurrent operation is taken into account at planning stage. Therefore, we must consider the sequence planning and the machine allocation simultaneously. In this paper, we propose a new scheduling method in which sequence planning and machine allocation are considered simultaneously. First of all, we propose a modeling method for an assembly sequence including a manufacturing environment. Secondly, we show a guideline in order to determine the estimate function in A* algorithm for assembly scheduling. Thirdly, a new search method based on combination of A* algorithm and supervisor is proposed. Fourthly, we propose a new technique which can take into consider the repetitive process in manufacturing system so as to improve the calculation time. Finally, numerical experiments of proposed scheduling algorithm are shown and effectiveness of proposed algorithm is verified.