Page 292 - Mechatronics for Safety, Security and Dependability in a New Era

Ch55-I044963.fm Page 276 Tuesday, August 1, 2006 9:39 PM
Ch55-I044963.fm
276
276 Page 276 Tuesday, August 1, 2006 9:39 PM
1. Re-allocation of the demand hardly took place unless due time of final inspection changed or
some WIPs pass other preceding WIPs.
2. In many cases, as there are less than tens of WIPs concurrently being processed in the whole
system, re-allocation of the demand takes only a few seconds. In the case of building materials
consisting multiple parts, as the multiple re-allocation processes occur, it may takes more than
ten seconds.
We used multiple agent support systems in the implementation, in order to ease future extension
of functionality. From the view point of the system extend-ability, we found followings.
1. All information on the actual process achievements are kept within a database. This makes any
other agents being able to utilize these data for processing and user interface purpose within a
few seconds, such as status display systems and e-mail sender programs.
2. It seems to be appropriate that the conversion from the identifier of RFID to building material
identifier should be done by specialized subsystem in the management system. There may be
many class of RFIDs other than expressing WIPs.
In this implementation, we introduced some limitations to simplify the system, as follows.
1. There is no problem solving engine to minimize cost.
To find best solution of re-allocation, it is required to minimize cost to re-distribute building
materials, or in some cases it is required to determine which WIP to be scraped. Solving these
problems may need massive computational power, because of the combinational explosion.
2. There is no clear decision rules to be embedded within the system.
Some incidents can be processed automatically without human interventions, but some require
human approvals. There is no clear border and the best way depends on its environment.

6 CONCLUSION

Through this implementation described above, we found that the integrated process management
system including both part-manufacturing and building construction is feasible enough. In this
financial year, we are now going to carry out a field test, applying this system to the actual
manufacturing and construction sites.

ACKNOWLEDGMENTS

This research activity has been carried out as a part of the Intelligent Manufacturing Systems
(IMS) international research program: "Innovative and Intelligent Parts-oriented Construction
(IF7-II)." We appreciate the kind guidance of each members of this project.

REFERENCES
[1] Takata M and Arai E. (2001). Implementation of a Layer Structured Control System on the
"Glue Logic". Global Engineering, Manufacturing and Enterprise Networks pp.488-496, Kluwer
Academic Publishers.
[2] Takata M and Arai E. (2005). Implementation of a Data Gathering System with Scalable
Intelligent Control Architecture. KNOWLEDGE AND SKILL CHAINS IN ENGENEERING
AND MANUFACTURING: Information Infrastructure in the Era of Global Communications,
pp.261-268, Springer.