Page - 52 - in Algorithms for Scheduling Problems
Image of the Page - 52 -
Text of the Page - 52 -
Algorithms 2018,11, 80
Proof. By Lemma 1 and the construction of the ļ¬rst pass, if at that pass, no IA(b2) arises, the
constructed partial schedule S(B,y) contains all the y-jobswith the lateness nomore than L(Ī“),
whichproves the lemma.
Phase 1, secondpass (schedulingx-jobs): The secondpass is invoked if no IA(b2)during the
execution of the ļ¬rst pass occurs. The ļ¬rst passwill then output a y-complete schedule S(B,y)
(Lemma2),whichis the input for thesecondpass.At thesecondpass, theavailablex-jobsare included
within the idle time intervals remaining inscheduleS(B,y).Now,wedescribeamodiļ¬edversionofa
commonDecreasingNextFitheuristics, adaptedtoourproblemforschedulingbinB (DNFheuristics
forshort). Theheuristicsdealswitha listof thex-jobs forkernelK sorted innon-increasingorderof
theirprocessingtimes,whereasthe jobswiththesameprocessingtimearesortedinthenon-decreasing
orderof theirduedates (i.e.,moreurgent jobsappearearlier in that list). Iteratively, theDNFheuristics
selects thenext jobx fromthelistandtries toschedule itat theearliest idle timemoment tā²atorbehind
its release time. If timemoment tā²+pi isgreater thantheallowableupperendpointof the intervalof
binB, jobxis ignored,andthenextx-job fromthe list is similarlyconsidered(ifnounconsideredx-job
is left, thenthesecondpass forbinBcompletes).Otherwise (tā²+pi fallswithin the intervalofbinB),
the followingstepsarecarriedout.
(A) If jobx canbescheduledat timemoment tā²withoutpushingany-job includedintheļ¬rstpass
(andbecompletedwithin the intervalof thecurrentbin), then it is scheduledat time tā² (being
removedfromthe list), andthenextx-job fromthe list is similarlyconsidered.
(B) Jobx is includedat time tā² if itpushesnoy-job fromscheduleS(B,y). Otherwise, suppose job
x, if includedat time tā²,pushesay-job,andletS(B,+x)bethecorresponding(auxiliary)partial
ED-scheduleobtainedbyrescheduling(right-shifting) respectivelyall thepushedy-jobsbythe
EDheuristics.
(B.1) Ifnoneof theright-shiftedy-jobs inscheduleS(B,+x) result ina latenessgreater than
thecurrent thresholdL(Ī“) (all these jobsbeingcompletedwithin the intervalofbinB),
then job i is included,andthenextx-job fromthe list is similarlyconsidered.
Weneedthe following lemmatocomplete thedescriptionof thesecondpass.
Lemma3. If atStep (B), the right-shiftedy-joby results ina latenessgreater than the thresholdvalue L(Ī“),
then therearises anewkernelK consistingof solelyType (a)y-jobs including joby.
Proof. Since the latenessof joby surpassesL(Ī“), it formspartofanewlyarisenkernel, thedelaying
emerging jobofwhich isx. Furthermore (similarlyas inLemma1),kernelK cannotcontainanyType
(b)y-job,asotherwise, suchay-jobwouldhavebeen includedwithin the idle timeinterval inwhich
jobx is included.Hence, thatType(b)y-jobcouldnotsucceed jobx.
(B.2) If thecondition inLemma3issatisļ¬ed, thenthecorrespondingnewkernel isdeclared,
andthecurrentsetofkernelsandbins isupdated. Similarly,as inStep2, theļ¬rstpass
is called for theļ¬rst newlydeclaredbin (initiatedat timemoment tā²) from thenewly
updatedpartition.
(B.3) If the latenessofsomey-jobsurpasses the thresholdL(Ī“) inscheduleS(B,+x), thenthe
nextx-job fromthe list is similarlyprocessedfromStep(A).
(B.4) If all thex-jobs fromthe listhavebeenprocessed, thenthesecondpass forschedulingbin
Bcompletes.
The following theorem tackles a frontier between the polynomially solvable instances and
intractable ones for ļ¬nding a Pareto-optimal frontier (the two relatedNP-hard problems are the
schedulingproblemwithourprimaryobjectiveandthederivedbinpackingproblem):
52
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