Page - 32 - in Algorithms for Scheduling Problems
Image of the Page - 32 -
Text of the Page - 32 -
Algorithms 2018,11, 66
non-fixed jobs fromthe setJ [Bk],whicharedistributed to theblockBk in thevertex j∈Vt of the
solutiontreesuchthat thearc (j,u)belongs to thesetEt.
At the first iteration of Procedure 5, the setP(J [B11]) = {J [B11], . . . ,∅} of all subsets of the
setJ [B11] is constructed and 2|J [B 1
1]| partial permutations are generated for all subsetsP(J [B11])
of the non-fixed jobsJ [B11]. The constructed solution tree T1 = (V1,E1) consists of the vertexes
0,1, . . . ,2|J [B11]| and2|J [B11]| arcs connectingvertex0with theothervertexes in the treeT1. Foreach
generated permutation πu(Bk;J[u]), whereJ[u] ∈ P(J [Blk]), the penalty φu determined in (6) is
calculated, which is equal to the difference between the lengths of themaximal possible relative
perimeterPermax of theoptimalityboxOB(πk,T) and the relativeperimeter of theoptimalitybox
OB(πu(Bk;J[u]),T),whichmaybeconstructedfor thepermutationπu(Bk;J[u]):
PerOB(πu(Bk;J[u]),T)=∑
i≤k Perumax−φu, (6)
wherePerumaxdenotes themaximal lengthof therelativeperimeterof thepermutationπu(Bk;J[u]),T).
Thepenaltyφu iscalculatedusingO(n)-Procedure3describedintheproofofTheorem5.Thecomplete
permutation π0→u(Bk;J[u])with the end πu(Bk;J[u]) is determined based on the permutations
πs(Bc;J[s]), where each vertex s belongs to the chain (0 → u) between vertexes 0 and u in the
solution treeTt=(Vt,Et)andeachblockBcbelongs to thesetBu={B1,B2, . . . ,Bk}. Thepermutation
π0→u(Bk;J[u]) includesall jobs,whicharefixed in theblocksBc∈Buordistributed to theirblocks
Bc∈Bu in theoptimalorder for thepenaltyφu.
TheaimofProcedure5 is toconstructacomplete jobpermutationπ f(Bm;J[f])=πk∈S such
that thepenaltyφf isminimal forall jobpermutations fromthesetS.At thenext iteration,apartial
permutationπs(B2;J[s]) ischosenfromtheconstructedsolutiontreesuchthatthepenaltyφs isminimal
amongthepermutationscorrespondingto the leafsof theconstructedsolutiontree.
At the second iteration, the setP(J [B22]) = {J [B22], . . . ,∅} of all subsets of the setJ [B22]
is generated and 2|J [B22]| permutations for all subsetsP(J [B22]) of the jobs from the block B2 are
constructed. For each generated permutation πv(B22;J[v]), whereJ[v] ∈ P(J [B22]), the penalty
φv is calculated using the equality φv = φl+Δφk, where the edge [l,v] belongs to the solution
treeT2=(V2,E2)andΔφk denotes thepenalty reached for theoptimalpermutationπ0→v(B2;J[v])
constructedfromthepermutationπv(B1;J[v])usingO(n)-Procedure3. For theconsiderationat the
next iteration,onechooses thepartialpermutationπd(Bc;J[d])with theminimalvalueof thepenalty
for thepartialpermutations in the leavesof theconstructedsolutiontree.
Thewholesolution tree is constructedsimilarlyuntil there isapartialpermutationwithasmaller
valueof thepenaltyφt in theconstructedsolutiontree.
4.2. TheApplicationofProcedure5 to theSmallExample
Table 1 presents input data for the example of the problem1|pLi ≤ pi ≤ pUi |∑Ci described
in Section 3.1. The jobs J4, J5 and J7 are non-fixed: J [B1] = {J4, J5}, J [B2] = {J4, J5, J7},
J [B3]={J4, J6, J7},J [B4]={J7}. The job J4 must be distributed either to the block B1, B2, or to
theblockB3. The job J5mustbedistributedeither to theblockB1,or to theblockB2. The job J7must
bedistributedeither to theblockB2,B3,or to theblockB4. Therelativeperimeterof theoptimalitybox
OB(πk,T) forany jobpermutationπk∈Scannotbegreater thanPermax=4×2=8dueto theupper
boundontherelativeperimeterPerOB(πk,T)given in (5).
At the first iterationofProcedure5, thesetP(J [B11])={∅,{J4},{J5},{J4, J5}} is constructedand
permutationsπ1(B11;∅),π 2(B11; J4),π 3(B11; J5)andπ 4(B11; J4, J5)aregenerated. Foreachelementof the
setP(J [B11]),we construct apermutationwith themaximal lengthof the relativeperimeter of the
optimalityboxandcalculate thepenalty.Weobtainthefollowingpermutationswiththeirpenalties:
π0→1(B11;∅)=(J1, J2, J3),φ1=1 19
30≈1.633333;π0→2(B11; J4)=(J4, J2, J1, J3),φ2=1,5;π0→3(B11; J5)=
(J1, J5, J2, J3),φ3=11330≈1.433333;π0→4(B11; J4, J5)=(J4, J5, J1, J2, J3),φ4=11330≈1.433333.
32
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