Seite - 120 - in Algorithms for Scheduling Problems
Bild der Seite - 120 -
Text der Seite - 120 -
Algorithms 2018,11, 50
Figure3.Permutationsofcompetingoperations inbranch-and-boundalgorithm.
2.3.2.Gradient-AlikeAlgorithm
Theoppositeof thementionedaboveapproachwouldbesearchingfora localminimumamong
all combinationswithagradientalgorithm[11].Here,wewill computeananalogueof thederivative
that isusedtodefinethesearchdirectionateachpoint [12]. ‘Derivative’ foreachcompetingoperation
is calculatedby trying to shift it byoneposition to the left. In case theobjective function reduces
(or increases formaximization problems)we shift the operation to the leftwith the defined step
(thestepratio ismeasured inanumberofpositionoperationskipsmoving to the left). The formal
presentationof thedesignedgradient-alikealgorithmisdescribedbyProcedure2.
Procedure2: gradientalgorithm
1.FindtheinitialsolutionoftheLPproblem (1)–(3)forthecombinationck∈{0,1},k=1,. . . ,K
thatcorresponds to thecasewhenallordersare inarow(firstoperationof the followingorderstarts
onlyafter the lastoperationofprecedingorder iscompleted).
2.Remember thesolutionandkeepthevalueofobjective functionΦas temporarilybest result
Opt=Φ.
3.Select the lastorder f= |F|.
4.Select the resource r= |R| that isusedbyfirstoperation i= 1of thesequencegraphof the
manufacturingprocess.
5. The current optimization variable for gradient optimization is selected—position of
operation iof theorder f ontheresource r< i,r, f>.
6.Condition:Does theselectedresource rhavemorethanoneoperation inthemanufacturing
procedure thatallocates it?
6.1. Ifyes then
Beginoptimizingposition< i,r, f>
6.1.1.Set thestepofshifting theoperation tomaximum(shifting to the leftmostposition).
D(i,r, f)=max
6.1.2.Findthe ‘derivative’ofshifting theoperation i to the left
Shift theoperation ionepositionto the left
Find thesolutionof theLPproblem (1)–(3) forcurrentcombinationandcalculate the
objective functionΦ
ConditionA: is thesolutionfeasible?
If feasible then
ConditionB: Is theobjectivefunctionvalueΦbetter thantemporarilybest resultOpt?
Ifyes then
Wefoundtheoptimizationdirectionforposition< i,r, f>,proceedtop. 6.1.3
Ifnot then
120
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