Page - 35 - in Algorithms for Scheduling Problems
Image of the Page - 35 -
Text of the Page - 35 -
Algorithms 2018,11, 66
optimalitybox(Section3),wepropose the followingAlgorithm3forconstructing the jobpermutation
ĻkāS for theproblem1|pLi ⤠pi⤠pUi |āCi,whoseoptimalityboxOB(Ļk,T)provides theminimal
valueof theerror functionF(Ļk,T)amongallpermutationsS.
Algorithm3
Input: Segments [pLi ,p U
i ] for the jobs JiāJ .
Output: ThepermutationĻkāSandoptimalityboxOB(Ļk,T),whichprovide
theminimalvalueof theerror functionF(Ļk,T).
Step1: IF theconditionofTheorem3holds
THENOB(Ļk,T)=ā
foranypermutationĻkāS
andtheequalityF(Ļk,T)= n(n+1)
2 holdsSTOP.
Step2: IF theconditionofTheorem4holds
THENusingProcedure2+F(Ļk,T)construct
thepermutationĻkāS suchthatbothequalities
PerOB(Ļk,T)=nandF(Ļk,T)=0holdSTOP.
Step3:ELSEdetermine thesetBofallblocksusingthe
O(n logn)-Procedure1described in theproofofLemma1
Step4: IndextheblocksB={B1.B2, . . . ,Bm}accordingto increasing
leftboundsof theircores (Lemma3)
Step5: IFJ =B1THENproblem1|pLi ⤠pi⤠pUi |āCi is calledproblemP1
(Theorem5)set i=0GOTOstep8ELSEset i=1
Step6: IF thereexist twoadjacentblocksBrā āBandBrā+1āB such
thatBrā ā©Brā+1=ā
; let rdenote theminimumof theabove index rā
in theset{1,2, . . . ,m}THENdecompose theproblemP into
subproblemP1with thesetof jobsJ1=āŖrk=1Bk andsubproblemP2
with thesetof jobsJ2=āŖmk=r+1BkusingLemma4;
setP=P1,J =J1,B={B1,B2, . . . ,Br}GOTOstep7ELSE
Step7: IFB ={B1}THENGOTOstep9ELSE
Step8:Construct thepermutationĻs(i)withtheminimalvalueof
theerror functionF(Ļk,T)usingProcedure3+F(Ļk,t)
IF i=0orJ2=BmGOTOstep12ELSE
Step9: IF thereexistsablock in thesetBcontainingmore thanone
non-ļ¬xed jobsTHENconstruct thepermutationĻs(i)with
theminimalvalueof theerror functionF(Ļk,T) for theproblemP1
usingProcedure5+F(Ļk,t)GOTOstep11
Step10:ELSEconstruct thepermutationĻs(i)withtheminimalvalueof the
error functionF(Ļk,T) for theproblemP1 usingProcedure3+F(Ļk,t)
Step11:Construct theoptimalityboxOB(Ļs(i),T) for thepermutationĻs(i)
usingAlgorithm1 IFJ2 =Bm and iā„1THEN
set i := i+1,P=P2,J =J2,B={Br+1,Br+2, . . . ,Bm}
GOTOstep6ELSEIFJ2=BmTHENGOTOstep8
Step12: IF i>0THENsetv= i,determine thepermutation
Ļk=(Ļ s(1),Ļs(2), . . . ,Ļs(v))andtheoptimalitybox:
OB(Ļk,T)=Ćiā{1,2,...,v}OB(Ļs(i),T)GOTOstep13
ELSEOB(Ļk,T)=OB(Ļs(0),T)
Step13:TheoptimalityboxOB(Ļk,T)has theminimalvalue
of theerror functionF(Ļk,T)STOP.
InSection6,wedescribe theresultsof thecomputationalexperimentsonapplyingAlgorithm3
to therandomlygeneratedproblems1|pLi ⤠pi⤠pUi |āCi .
6.ComputationalResults
For thebenchmark instances1|pLi ⤠pi⤠pUi |āCi,where therearenopropertiesof therandomly
generated jobsJ , whichmake theproblemharder, themid-point permutationĻmidāp = Ļe ā S,
whereall jobs JiāJ areorderedaccordingtothe increasingof themid-points p U
i +p L
i
2 of theirsegments
[pLi ,p U
i ], isoftenclose to theoptimalpermutation. Inourcomputationalexperiments,wetestedseven
35
back to the
book Algorithms for Scheduling Problems"
Algorithms for Scheduling Problems
- Title
- Algorithms for Scheduling Problems
- Authors
- Frank Werner
- Larysa Burtseva
- Yuri Sotskov
- Editor
- MDPI
- Location
- Basel
- Date
- 2018
- Language
- English
- License
- CC BY 4.0
- ISBN
- 978-3-03897-120-7
- Size
- 17.0 x 24.4 cm
- Pages
- 212
- Keywords
- Scheduling Problems in Logistics, Transport, Timetabling, Sports, Healthcare, Engineering, Energy Management
- Categories
- Informatik
- Technik