MODELLING FLEXIBLE MANUFACTURING SYSTEMS - Modelling with Generalised Stochastic Petri Nets - page 195

Table 8.5: Specification of the transitions of the GSPN of Fig. 8.14

transition rate semantics

LU 0.25 single-server

mv LU−1 mv LU−23 0.5 single-server

outM 1

0.1 single-server

mv 1−23

0.5 single-server

outM 2 0.05 single-server

outM 2 0.0667 single-server

outM 3 0.0333 single-server

outM 3

0.1 single-server

mv 23−LU 0.5 single-server

transition weight priority ECS

waitAGV 1 1 1 1

waitAGV 2 1 1 1

waitAGV 3 1 1 1

part a 20 1 2

part b 80 1 2

inM 2

1 1 2

inM 2

1 1 3

inM 3 1 1 4

inM 3 1 1 5

transition (observe that transitions inM 2 , inM 2 , inM 3 , and inM 3 are not
in the same ECS because they are mutually exclusive with each other).
The ECS identified by number 1 comprises all the immediate transitions
representing the acquisition of the AGV for a transport. They have equal
weight, meaning that in case of conflict for the use of the AGV, the choice is
performed using a uniform probability distribution. ECS number 2 is some-
how unexpected: while it is natural to associate transitions part a and part b
with a common ECS, and assign them a weight according to the required
production mix, it is annoying to have transition inM 2 in the same ECS.
However, this cannot be avoided since transition part b is in SC relation with
transition inM 2 due to the inhibitor arc from place inM 23 (which is in the
output set of part b ) to transition inM 2 . ECS number 2 may cause confusion,
indeed it is a non free-choice ECS and some immediate transitions exist that
are in causal connection with the transitions in the ECS and have the same
priority. For instance let us consider a marking with two tokens in place
inM 23 , one token in place choice, one token in place idleM 3 , and one token
in place idleM 2 ; the firing sequences part b , inM 2 , inM 3 and inM 3 , part b ,