A job-shop scheduling method is described utilizing a negotiation-base
d algorithm in autonomous distributed systems. Elements of a productio
n system such as jobs and machines are regard as agents that act like
an independent profit maker. A negotiation procedure is suggested incl
uding price-based bid construction and price revising mechanism. We pr
opose a simulation-based iteration procedure in order to revise the de
cision making parameters. It is shown that the iterative procedure eve
ntually improves the global performance of the entire system. The prop
osed method is expected to have flexibility against various system alt
ernations, addition or deletion of jobs or sudden failure of machines.
We develop a prototype software that utilizes the object oriented con
cept. (C) 1997 Elsevier Science Ltd.