Title of article
Event-based MILPmodelsforresource-constrainedprojectscheduling problems
Author/Authors
Oumar Kone´ ، نويسنده , , ChristianArtigues ، نويسنده , , PierreLopez ، نويسنده , , MarcelMongeau، نويسنده ,
Issue Information
ماهنامه با شماره پیاپی سال 2011
Pages
11
From page
3
To page
13
Abstract
In thispaperwemakeacomparativestudyofseveralmixedintegerlinearprogramming(MILP)
formulationsforresource-constrainedprojectschedulingproblems(RCPSPs).
First, wepresentthreediscreteandcontinuoustimeMILPformulationsissuedfromtheliterature.
Second,insteadofrelyingonthetraditionaldiscretizationofthetimehorizon,weproposeMILP
formulationsfortheRCPSPbasedontheconceptof event: the Start/End formulationandthe On/Off
formulation.Theseformulationspresenttheadvantageofinvolvingfewervariablesthanthe
formulationsindexedbytime.Becausethevariablesofthistypeofformulationsarenotfunctionof
the timehorizon,wehaveabettercapacitytodealwithinstancesofverylargeschedulinghorizon.
Finally, weillustrateourcontributionwithaseriesoftestsonvarioustypesofinstancewiththe
MILP formulationsissuedfromtheliterature,togetherwithournewformulations.Wealsocompareour
results witharecentRCPSP-specificexactmethod.Weshowthat,intermsofexactsolving,noMILP
formulationclassdominatestheotheronesandthatastate-of-theartspecialized(non-MILP)method
for theRCPSPisevenoutperformedbyMILPonasetofhardinstances.Furthermore,onanothersetof
hard ‘‘highlycumulative’’RCPSPinstanceswithahighprocessingtimerange,ourOn/Offformulation
outperformsalltheothersMILPformulationsandobtainsresultsclosetotheonesofthespecialized
method.
Keywords
Mixed Integer Linear Programming , Resource-constrained project scheduling , Project Management
Journal title
Computers and Operations Research
Serial Year
2011
Journal title
Computers and Operations Research
Record number
927829
Link To Document