Seite - 8 - in Algorithms for Scheduling Problems
Bild der Seite - 8 -
Text der Seite - 8 -
Algorithms 2018,11, 18
TheonlydifferencebetweenConditions 3 and4 iswhether dk+2 = 0ornot. This implies that
periodk is the lastoff-peakperiodandperiodk+2doesnotexist.Hence,periodkmustbe in the last
cycle (i.e.,on the lastday)andthereare twoscenarios thatshouldbeconsideredas follows(thezero
pointsof the twoscenariosaredifferent).
Scenario1: In this scenario,neitherperiodk+1norperiodk+2existsandfourpossiblepositions
markedinFigure6aaregiven. Similar toCondition3,periodk is criticaldueto its lowestpriceand
the longest remaining idle time. To insert job i intoperiodk, theother jobsprocessedwithinperiodk
shouldbemovedrightor left.Hence, twopositionsareanalyzedfirst.
1. Tooccupyoff-peakperiodsasmuchaspossible,asetofalreadyinserted jobs inperiodk should
bemovedto therightmostside,andthen job icanbeprocessedacrossperiodskandk−1.
2. All inserted jobs inperiodk shouldbemovedleft so that job icanbeprocessedwithinperiodk.
Notethatthemovementof jobsmayleadtoachangeintheirelectricitycosts,definedasmovement
cost. The insertioncostof job i isequal to theelectricitycost forprocessing job iplus thecorresponding
movementcost if somealreadyinserted jobsmustbemovedtoenlarge the idle timeofperiod(s). If the
movementcost iszero, thenthe insertioncost isequal to theelectricitycost forprocessingthe job.
Let cost1and cost2, respectively,denote the insertioncostsof theabovecorrespondingpositions.
Since thepower consumption rate of job i is not higher thananyoneof the already inserted jobs,
it followsthat cost1≤ cost2. That is, thePosition1 is theonlyone that shouldbeconsideredwhen job i
is inserted intoperiodk.Next, letksmdenote theoff-peakperiodwithsubmaximalremaining idle time.
Positions3and4aregivenas follows.
3. Suppose that job i isprocessedacrossperiodsksm,ksm+1,andksm+2. If Iksm is slightlysmaller
than Ik, cost3maybe less than cost1asperiod ksm+1 is amid-peakperiod. Hence,Position3
needs tobeconsidered.
4. Similar toCondition3, job icanbeprocessedwithinamid-peakperiod.
(a) Illustration of Scenario 1
(b) Illustration of Scenario 2
Figure6. IllustrationofCondition4.
8
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