Seite - 16 - in Algorithms for Scheduling Problems
Bild der Seite - 16 -
Text der Seite - 16 -
Algorithms 2018,11, 18
4.2. RandomlyGenerated InstancesStudies
In these tests,weutilize twosetsof randomlygenerated instances toevaluate theperformanceof
theproposedalgorithm. For thefirst setof small-size instances, therangeof jobs is [20,100]. For the
secondsetof large-size instances, thenumberof jobs is set as500, 1000, 2000, 3000, 4000, and5000.
Theprocessing time ti is randomlygeneratedfromauniformdistribution(30,210)minandthepower
consumptionperhourpi is randomlygenerated in (30,100)kW.Tomeasure theeffectof theproposed
algorithm,parameter e is setas e=1.2,1.5,2,3.
Foreachgroupofn (n≤2000)and e, 10 randominstancesaregenerated, thentheaveragevalues
of10testsarecalculatedandrecorded.Whenthenumberof jobs issetas3000,4000,and5000,GIHhas
torunformore than4h(the longest isnearly twodays) tofindafeasiblesolution. Thus, considering
thefeasibilityof theexperiment,only3randominstancesaregeneratedinsuchagroupoftests.All the
averagevaluesarerecordedinTables5and6.Meanwhile, forthelarge-sizeinstances,weaddtworules
toGIHtoreduce thecomputationtimewithoutchangingthecomputationalaccuracy. The improved
algorithmisnamedGIH2.
Table5.Computational results for thesmall-size instances.
Instance GIH GIH-F
n e m TECH CTH (s) TECF CTF (s) G (%) R
20 1.2 12.0 1634.1 0.034 1632.5 0.002 −0.10% 17.0
1.5 15.0 1370.1 0.037 1370.1 0.002 0.00% 18.5
2.0 19.0 1295.7 0.041 1295.1 0.002 −0.05% 20.5
3.0 28.5 1168.0 0.056 1168.0 0.001 0.00% 56.0
30 1.2 18.0 2414.6 0.064 2415.7 0.002 0.05% 32.0
1.5 20.0 2274.6 0.065 2274.1 0.002 −0.02% 32.5
2.0 28.5 2005.0 0.083 2005.0 0.002 0.00% 41.5
3.0 39.5 1741.0 0.119 1741.0 0.002 0.00% 59.5
40 1.2 23.0 3342.1 0.096 3342.0 0.004 0.00% 24.0
1.5 28.0 2900.1 0.109 2899.3 0.003 −0.03% 36.3
2.0 36.0 2775.6 0.143 2775.0 0.003 −0.02% 47.7
3.0 52.0 2380.3 0.194 2380.3 0.002 0.00% 97.0
50 1.2 27.5 4242.5 0.137 4242.4 0.005 0.00% 27.4
1.5 34.0 3733.0 0.164 3732.6 0.003 −0.01% 54.7
2.0 43.0 3243.8 0.212 3243.2 0.004 −0.02% 53.0
3.0 64.5 2940.6 0.315 2940.6 0.003 0.00% 105.0
60 1.2 34.0 4820.8 0.204 4819.7 0.006 −0.02% 34.0
1.5 40.0 4536.5 0.224 4536.3 0.004 0.00% 56.0
2.0 52.0 4029.0 0.293 4028.9 0.004 0.00% 73.3
3.0 78.0 3544.1 0.464 3544.1 0.004 0.00% 116.0
70 1.2 37.5 6133.5 0.249 6132.2 0.007 −0.02% 35.6
1.5 46.0 5416.3 0.303 5416.2 0.007 0.00% 43.3
2.0 61.0 4676.0 0.413 4675.8 0.004 0.00% 103.3
3.0 90.0 4024.9 0.643 4024.9 0.005 0.00% 128.6
80 1.2 43.0 7073.1 0.321 7072.9 0.009 0.00% 35.7
1.5 53.0 6049.6 0.401 6049.6 0.006 0.00% 66.8
2.0 68.5 5348.1 0.554 5348.1 0.007 0.00% 79.1
3.0 101.5 4514.4 0.868 4514.3 0.005 0.00% 173.6
90 1.2 48.0 8128.5 0.399 8128.4 0.009 0.00% 44.3
1.5 58.0 6772.5 0.501 6772.4 0.011 0.00% 45.5
2.0 77.5 6172.7 0.697 6172.6 0.008 0.00% 87.1
3.0 104.1 5228.2 1.196 5228.2 0.009 0.00% 132.9
100 1.2 53.5 8623.5 0.509 8622.9 0.017 −0.01% 29.9
1.5 64.0 7607.1 0.614 7607.0 0.011 0.00% 55.8
2.0 86.5 6896.8 0.927 6896.8 0.014 0.00% 66.2
3.0 128.0 5815.2 1.482 5815.1 0.009 0.00% 164.7
16
zurück zum
Buch Algorithms for Scheduling Problems"
Algorithms for Scheduling Problems
- Titel
- Algorithms for Scheduling Problems
- Autoren
- Frank Werner
- Larysa Burtseva
- Yuri Sotskov
- Herausgeber
- MDPI
- Ort
- Basel
- Datum
- 2018
- Sprache
- englisch
- Lizenz
- CC BY 4.0
- ISBN
- 978-3-03897-120-7
- Abmessungen
- 17.0 x 24.4 cm
- Seiten
- 212
- Schlagwörter
- Scheduling Problems in Logistics, Transport, Timetabling, Sports, Healthcare, Engineering, Energy Management
- Kategorien
- Informatik
- Technik