Graphics Programs Reference
In-Depth Information
When the cells are connected, however, we have to take into account some
additional constraint. The transitions of the subnet representing cell i might
all be disabled because busyM
i
is empty, IB
i
is empty, and either BB
i
contains some token but OB
i−1
is empty, or OB
i
contains some token, but
BB
i+1
is empty.
If cell i is blocked because K
i
tokens are in place BB
i
but OB
i−1
is empty, it
is easy to see that OB
i−1
will eventually receive a token so that enterCell
i
will become enabled (and hence also transitions inM
i
and outM
i
will eventu-
ally become enabled). This can be proved by induction, showing that if OB
j
is marked (for some j : 1
≤
j < i
−
1) then OB
j+1
will eventually receive a
token. Indeed, due to the P-invariant above, if OB
j+1
is empty, then K
j+1
tokens must be distributed into places BB
j+1
, IB
j+1
and busyM
j+1
; since
by hypothesis OB
j
is marked, at least one token can eventually flow from
any of the three places listed before, to place OB
j+1
. Moreover, it is always
possible to reach a marking in which OB
1
is marked due to the fact that
transition enterCell
1
can fire as soon as BB
1
contains at least one token
(the reader may easily convince him/herself that this statement is true by
playing the token game on the GSPN model).
Hence, from a marking in which BB
i
is marked,
∀
k : j < k
≤
i
−
1,OB
k
are
empty and OB
j
is not empty, it is eventually possible to reach a marking in
which OB
j+1
is marked; from here it is possible to reach a marking in which
OB
j+2
is marked, and so on until a marking is reached in which OB
i−1
is
marked, so that transition enterCell
i
may fire.
If the subnet of cell i is blocked because K
i
tokens are in OB
i
but BB
i+1
is
empty, then it is possible to show that transition enterCell
i+1
will eventually
become enabled. In fact, it is easy to verify that if BB
j+1
is marked (for
some j : i < j
≤
n) and BB
j
is empty, then BB
j
will eventually receive
a token. Moreover, from a marking in which BB
n
is empty, it is always
possible to reach a marking in which BB
n
is marked due to the fact that
transition exitCell
n
can fire as soon as OB
n
contains at least one token.
Hence from a marking in which OB
i
is marked,
∀
k : i + 1
≤
k < j,BB
k
are empty and BB
j
is not empty, it is possible to reach a marking in which
BB
j−1
is marked. From here it is possible to reach a marking in which
BB
j−2
is marked, and so on, until a marking is reached in which BB
i+1
is
marked, so that transition enterCell
i+1
may fire.
8.2.2
Performance analysis of the Kanban system
In this section we summarize the results of the extensive performance eval-
uation study of a Kanban model presented in [
41]
. The experiments have
the objective of studying the performance and fault tolerance of a Kanban
system with five cells. The studied system is highly homogeneous: all the
cells have the same machining time (the rate of transition outM
i
is 4.0) and
the same number K of cards. A first set of experiments was performed to
Search WWH ::
Custom Search