Information Technology Reference
In-Depth Information
Table 3.29.
DDE FSS makespan
m
x
n
Generated
problems
GA
DE
(Solution)
GA/DE
4x4
5
44
39
-
5x10
5
79
79
-
8x15
5
143
138
-
10x25
5
205
202
-
15x25
5
248
253
98.02
20x50
5
468
470
99.57
25x75
5
673
715.4
94.07
30x100
5
861
900.4
95.62
Table 3.30.
DDE FSS total tardiness
m
x
n
Generated
problems
GA
DE
(Solution)
GA/DE
4x4
5
54
52.6
-
5x10
5
285
307
92.83
8x15
5
1072
1146
93.54
10x25
5
2869
2957
97.02
15x25
5
3726
3839.4
97.06
20x50
5
13683
14673.6
93.25
25x75
5
30225
33335.6
90.67
30x100
5
51877
55735.6
93.07
With all the experimentation parameters selected, the FSS problems were evaluated.
Three different objective functions were to be analysed. The first was the makespan.
The makespan is equivalent to the completion time for the last job to leave the system.
The results are presented in Table 3.29.
The second objective is the tardiness. Tardiness relates to the number of tardy jobs;
jobs which will not meet their due dates and which are scheduled last. This reflects the
on-time delivery of jobs and is of paramount importance to production planning and
control [30]. The results are given in Table 3.30.
The final objective is the mean flowtime of the system. It is the sum of the weighted
completion time of the
n
jobs which gives an indication of the total holding or inventory
costs incurred by the schedule. The results are presented in Table 3.31.
Tables 3.29
3.31 show the comparison between Genetic Algorithm (GA) devel-
oped in a previous study for flowshop scheduling [28], compared with DDE. Upon
analysis it is seen that, DE algorithm performs better than GA for small-sized prob-
lems, and competes appreciably with GA for medium to large-sized problems. These
results are not compared to the traditional methods since earlier study of [4] show that
GA based algorithm for flow shop problems outperform the best existing traditional
approaches such as the ones proposed by [16] and [39].
−
Search WWH ::
Custom Search