the  state  of an assembly.
The state of an assembly is an  instance of the conditions  of all its tasks.
SupposeanassemblyhasmtasksTi.i=1.2,  .m.  TaskT1 throughTi
are  done and therestare not  done  yet. Then the  state of  this assembly can
be repremM as
i  m
S,=(  A  =I  DONE(Tj))A(  ,-+I  ,A  -DONE(Tj)).  (1)
That is. S,  represents  the state whose interpretation  of S,  is TRUE. Ndcc
that a formula can also  represent a set  of states. For example, by omitting  4
in S,,
i-1  )I
S',  =  ( A  DONE( Tj ))A  (  j$+l  -DONE(  Tj ))  .  (2)
J -1
can represent a set of two states,  one  with Ti having  been  Qne and  the  other
with Ti  not done yet because bofh  states will give S',  a TRUE  intapetation
tion of  current  subassemblies.  Since at any time instance.,  an  assembly
presents  as  a collection of  subassemblies  and  components.  This  collection
is a partition of the set  of all components  and can also fully describe assem-
bly  states.  Suppose  Ci's. i  = 1,2,  *.  ,  A,  are a  cdlection of  subassem-
blies,  then
The second representation  describa assembly states  by listing a
Sc=(C,.C2,  *'*  .C&),  (3)
k
14
where  uCi=C and  CinCj=O,  ifi+j.  (4)
is a representation  of an assembly state. and C  is the  complete set  of  com-
ponents  of an assembly.
Whiie both representations  described above are  equivalent  in repsent-
ing assembly  states, the first one  is a better  representation for precedence
knowledge while the second  one is more suitable for assembly plan genera-
tion.  Both  of them will be  used  in our assembly planning  system.  For  brev-
ity,  we  shall refer  the  tirst  representation as relational fonn of  state
representation and  denote it by S,.  and  refex  the second representation as
componentfonn of state repmentation and  denote  it by Se. It is not difficult
to transform these two representation  forms  from one to  the other.  Suppose
task Ti relates  two components  cp E  C, and cq E Cy,  then
That is, if Ti is done. then the components  related  by Ti belong to the  same
subassembly; likewise if  two components  belong to the  same  subassembly,
then the task  relating  them is done.
During  an assembly process  a subassembly is usually fixed by a fixture
to  maintain  the  spatial  relationships among component.  Before being
assembled, a component  is kept in a patt feeder which can also  be  viewed
as a type  of  fixture. The  association between  fixture and subassemblies  is
described  by the following  predicate  in the system.
DONE(Ti)=TRUE  ifandonlyif  u=v.  (5)
('""  ifC,isfkedbyafixhmf
C"RRm-Fm"RE(  '1  1 f  )=  FALSE.  XC,  is not fixed by  a fixbuef
where Ci  is a component a  subassanbly. andfis a vrriable
fixture,
if  Ci has  no 6xtllrc with  it
ifCiisfixedbyafixtllrc
if Ci  is a component  in apart  f-.
Tofullydesdbethefixture  . .  lsjtuatiOnh~asslmhlycC~,ach 自动装配规划系统英文文献和中文翻译(4):http://www.youerw.com/fanyi/lunwen_6791.html