General Layou-t: of Shop Floor. f o r Experiment. Each specific work station in this layout. w i l l have a !-:nowledge base optimizing the schedule of work through.
APPLICATION OF DISTRIBUTED KNOWLEDGE RASES IN INTELXBIGENT MANUFACTURING D r . Cihan H . D a g l i and G e r a l d E. Hoffman E n g i n e e r i n g Management Department University of Missouri-Rolla R o l l a , MO 65401
of artificial intelliqent methods, a l t h o u g h one method o f t e n tends to determine the system s - t . r u c t u r e whi 1.e o t h e r s provide supp1ement:al f u n c : t i o n s . M a n u f a c t u r i n g p r o c e s s e s a r e v e r y dynamic and s i t u a t i o n s have a t e n d e n c y t o change rapidly. Expert systeins, therefore, r e q u i r e e x t e n s i v e maintenance i n o r d e r .to keep them effective. Also, once d e v e l o p e d and t e s t e d , expert systems c o u l d s t i l l f a i l t o work i n t h e i n t e n d e d operational e n v i ronmerit because s i t u a t i o n s may have changed w i t h i n t h e (21. Knowledge o r i g i n a l environment bases for a complex manufacturing environment have a t e n d e n c y t o b e l.arge and complex t h e m s e l v e s . To r e v i s e o r expand t h e knowledge ba.::e when new it. idould have t o situations occur, r e w r i t t e n t o accommodate t h e new r u l e s o r facts. R e v i s i n g t h e s e coniplex knowledge b a s e s may c r e a t e problem a r e a s i n o t h e r areas of manufacturinq en.vironment, b e c a u s e of t h e i n t e r a c t i o n of the rules used. A method t h a t can b e used t o s i . m p l i f y t h e knowledge b a s e s would be t o cleveI.op smaller specific knowledge b a s e s f o r s p e c i f i c a r e a s w i t h i n t h e manufac:turing environment and l i n k t h e s e knowledge they be b a s e s t o g e t h e r s o n o t on1.y w i l l acting i n parallel t o teach o t h e r b u t i n f o r m a t i o n d e r i v e d from one knowledge b a s e can be passed t o another a s w e l l . T h i s p a p e r examines t h e f e a s i b i l i t y of d e v e l o p i n g s m a l l e r knowl1i:dge b a s e s f o r v e r y s p e c i f i c p u r p o s e s and l i n k i n g them t o a n o t h e r knowledge b a s e t o c o o r d i n a t e t h e flow o f i n f o r m a t i o n and d i r e c t p r o p e r r e s p o n s e s back t o t h e sm;itller knowledge bases. Knowledge bases a're currently d e v e l o p e d by one o r a combination of f o i i r techniques: ( 1 ) a domain e x p e r t i n t e r v i e w e d by R knowledge e n g i n e e r f o r f ac-t.s and r u l e s ; (2) o p e r a t i o n a l d a t a b e i n g reviewed by t h e knowledge e n g i n e e r and coriverteci t o f a c t s and r u l e s ; (3) t h e d0mai.n e x p e r t i r i t e r a c : t w i t h an Intelligent Editing Expert System, o r ;
ABS TRACT _- - . The knowledge b a s e d s y s t e m s d e v e l o p e d a s research o r i n d u s t r i a l prototypes r e c e n t l y a r e g e n e r a l l y ba-tch s y s t e m s and t h e y a r e b u i l t t o be used i n a s e q u e n t i a l environment. Most of t h e m a n u f a c t u r i n g a p p l i c a t i o n s r e q u i r e t h e u s e of systems i n p a r a l l - e l t n v e p r e s e n t t.he dynamics of t h e shop f l o o r . This is possible e i t h e r through independent knowledge bases work stations o p e r a t i n g on s e p a r a t e networked t o g e t h e r o r i m p l e m e n t a t i o n of t h e s e systems on p a r a l l e l computers. T h i s p a p e r c o n s i d e r s t h e former approach w i t h i n t h e domain of kriowledge b a s e d scheduling. The scheduli-ng problem i s a d d r e s s e d t h r o u g h d i s t r i b u t e d knowledge bases t h a t have an a b i l i t y t o p a s s i.nformation back and f o r t h between small knowledge b a s e s f u n c t i o n i n g a t d i f f e r e n t d e c i s i o n making l e v e l s .
__--
-~ INTRODUCTION ~
~
There i s a b i g gal:) among the schedules generated 1.ising various p l a n n i rig systems and schedu 1e s a c t u a 11y used i n t h e shop f l o o r e n v i r o n m e n t . T h i s is basically due to fundamental characteristics of aI.l scheduling difficulties. Effective scheduling i s a knowledge intensive activity requiring a comprehensive model of the f a c t o r y and i t s , environment a t a l l t i m e s i n the [ 41 . schedule generation process Underst.anding t h e d o m a i n ' s c o n s t r a i n t s and t h e i r relative i m p o r t a n c e on t h e f a c t o r y s c h e d u l e and t h e r e l a t i o n s h i p s t o one a n o t h e r c a n p r o v i d e an e f f e c t i v e foundation f o r schedule generation [ S I . Knowledge b a s e d s c h e d u l i n g i s s t i l l v e r y much i n i t s i n f a n c y . Few cornpi-eted systems e x i s t , b u t a number of r e s e a r c h efforts are underway. Steffen I61 a r t i f ic i a1 i n t e 11. i genc e i dent.i f i e s r e s e a r c h i n t o t.wo broad cateqories: knowledge representation and c o n t r o l strategy. Most s y s t e m s use! a c o m b i n a t i o n
239 CH2767-2/89/0000-0239 $1.OO
0 1989
IEEE
A G V ' s movements were:
(4) t h e domain expert. c a n i n t e r f a c e "I earns" w i t h an i n d u c t i o n program which from examples and s c e n a r i c ) s p r e s e n t e d by t h e e x p e r t [ 11.
t y p e and amount ot: raw m a t e r i a l s needed a t work c e l l s -- t y p e and l o c a t i o n of p r o d u c t s n e e d i n g t o b e moved - b u f f e r s t o c k n e e d i n q s t o r a g e and re t r i eva 1 -- a n t i c i p a t i o n of j o b i::ompletion - movement r a t e s and t i m e s between t h e v a r i o u s work c e l l s limitations of si.ze and weight a l l o w a n c e s p e r AGV
-
CONCEPT A small manufacturing plant was conceptualized i n o r d e r t o experiment with t h i s process. The g e n e r a l o u t l i n e and a r e a s of t h e manufal.turing shop i s shown i n F i g u r e 1. The domain s p e c i f i c a r e a f o r t h e knowledge b a s e s was t o o p t i m i z e t h e s c h e d u l i n g o f work a t each work s t a t i o n i n o r d e r t o meet a weekly quota.
Raw bterial Inventory
&.lling Operations
Lathe OPns
C i r c u i t Board Assembly W.Ork--C-elJ
~
For t h e C i r c u i t Board Assembly, a different s e t o f o b j e c t i \ r e s w e r e needed Some t o o p t i m i z e t h e assembly s c h e d u l e . of t h e c o n s i d e r a t i o n s r e q u i r e d were:
Heat Treatment
-
Board
Stock Quality Control
-
F i g u r e 1. General Layou-t: of Shop F l o o r f o r Experiment Each s p e c i f i c work station in this will have a !-:nowledge b a s e layout o p t i m i z i n g t h e s c h e d u l e of work t h r o u g h that p a r t i c u l a r work ctiill.. In t h i s cell was experiment, each work r e p r e s e n t e d by an IBM 615% work s t a t i o n . The knowledge b a s e s t h a t was used t o of i n f o r m a t j on coordinate the flow r e s i d e d i n a n o t h e r IBM 6152 work s t a t i o n . The work s t a t i o n s were l i n k e d t o e a c h o t h e r and t h e c o n t r o l l i n g knowledge b a s e u s i n g a t o k e n r i n g network f o r t h e Local (LAN) . Each particular Area Network knowledge b a s e f o r e a c h work s t a t i o n was d e s i g n e d t o accompli s h t a s k s p a r t i cu 1a r t o t h a t work c e l l i t r e p r e s e n t e d . Some examples of t h e s e t a s k s c l r e shown f o r s e l e c t e d work c e l l s i n t h e following sections.
actual s e q u e n c i n g of t a s k s t o b e performed +normal j o b v e r s u s e x p e d i t e d j o b +board f a b r i c a t i o n +circuit track + s o l d e r mask +component s p a c i n g +component o r i e n t a t i o n +composite s e l e c t i o n +general layout movement of the product o r raw m a t e r i a l s t o and from t h e A G V ' s o r d e r i n g new m a t e r i a l s when s h o r t i n f o r m i n g t h e c o n t r o l l i n g knowledge b a s e when + j o b i s completed + t o t a l j o b s complet.ed or ].ate +maintenance probl.ems +completion of r e p a i r s
M i -
l l and Lathe Work C e l l s __
The m i l l i n g and l a t h e work c e l l s N machines o f e a c h t y p e c o n s i s t i n g of would r e q u i r e s c h e d u l i n g of: j o b s f o r b o t h normal o p e r a t i o n s and expedited orders - r o b o t i c arm movement t o l o a d and unload t h e A G V ' s -- t o o l c o n t r o l f o r +replacement +changeovers - i n f o r m i n g t h e c o n t r o l . l i n g knowledge b a s e of j o b s completed and any l a t e j o b s - i n f o r m i n g t h e c o n t r o l l i n g knowledge b a s e of maintenance problems o r r e p a i r s needed. -
AGV Work C e ll ~
Similarily. t h e o t h e r work c e l l s . such a s t h e raw m a t e r i a l c e l l . , h e a t t r e a t m e n t c e l l , assembly c e l l and q u a l i t y will r e q u i r e knowledge control cell., bases developed to fulfill their requirements i n optimizing t h e scheduling of j o b s t h r o u g h them. The c o m p l e t i o n of jobs, need f o r movement o f raw m a t e r i a l s
The Automated Guided V e h i c l e (AGV) knowledge b a s e needed t o s c h e d u l e t h r e e A G V ' s t o minimize i d l e t i m i ? of t h e A G V ' s and ensure timely mo.,rement of raw m a t e r i a l s , w o r k - i n - p r o c e s s , and f i n i s h e d goods. Some of t h e c o n s i d e r a t i o n s needed by t h e knowledge b a s e i n :;cheduling t h e
240
or products and the need for raw materials would be passed from one knowledge base to another for its use in schedu1ing and implementi nq . The controlling knowledge base, in turn, must coordinate all t.he information being generated and direct the proper knowledge to the proper locations. Additionally, it must be able to handle routine challenges to everyday operations and must have a way of' resolving new problems that arise. This may be a set of rules to use if a particular sequence of occurrences happen or it may mean contacting a human user to direct the solution of the problem. Some of the areas this controlling knowledge base must coordinate or infotmation passed along are:
various knowledge bases. 'I'he information can then be updated as chai-igesoccur and these changes passed on to the knowledge bases. Some initial assumptioi~swere made to simplify the process and get the experiment started. These assumptions were : (1) to use a batch p:t'ocess in which all jobs were scheduled at the start of a particular time period rather than jobs continually arriving. For our purposes, all jobs were scheduled at. the beginning of each day. (2) that the system updates itself as to job status and revi.sing schedules every two hours rather than real time. This was done in order to speed up the tests and because there were no actual machines to provide actual real-time feedback. Two hours was chosen arbitrarily to allow rapid progress over the daily interval but still allowed enough detail for analysis of ongoing operations.
display to the user the current status of: +raw material inventory +buffer stockage +finished goods compLeted +job status +quality inspection results +AGV status and locations +late job data +scheduling rules currently being used at work cells - pass information to other knowledge bases needed to schedule jobs - tie in with simulation/animation to display current shop status - develop rules to handle changes in floor conditions and direct work cells of the implementations -
__--
