Information Technology Reference
In-Depth Information
(
t
+
1
(
t
)
(
t
+
1
(2)
x
=
x
+
v
i
i
i
The working flow chart of PSO is shown in Figure 7:
Fig. 7.
The working flow chart of PSO
3.3 Algorithm Achieving
Encoding and decoding
The particle's position can't be directly used to represent the solution of scheduling
problem. Reference [6] proposed three-dimensional particles encoding method which
is based on the particle position sequence (PPS) and the particle position round (PPR).
In this paper the method's application field is changed and applied on AGV schedul-
ing of FMS logistics system. It is referred to as particle encoding method based on
PPS-PPR.
Table 1.
Vector representation method about the
i
th partical
1
2
3
…
M
Machine number
Application order
m
i1
i2
i3
…
m
iM
AGV allocation
n
i1
n
i2
n
i3
…
n
iM
In the table1, N is the total number of a scheduling task, the total number of AGV
is J, M is the number of particle swarms, i
∈
[1, M]. The first line means the task
number of AGV application; The second line means the order of the N tasks, when
decoding, sort the whole column of vector particles to the tasks which belong to the
same AGV by the value of m
i (1~N)
; The third line means AGV assignment, when the
upper and lower limit of particles initialization were given, and n
i (1~N)
∈
[1, J + l), by
taking the operation INT (n
ij
) get the natural number between1 to J, after value de-
coded corresponding the AGV number. Therefore, during the algorithm initialization
each dimension vector can be set a range, to some degree it equivalent to the feasible
region range of the algorithm. For example, the first dimension range is 1 ~ N, the
second dimension range is 1.0 ~ 10.0, the third dimension range is 1~J +1 (If the
number of AGV is J). However, the range must be coordination with the algorithm
step selected, otherwise, the phenomenon that particle standing still may appear.
Search WWH ::
Custom Search