=Paper=
{{Paper
|id=Vol-1623/paperls6
|storemode=property
|title=The Band Collocation Problem: a Library of Problems and a Metaheuristic Approach
|pdfUrl=https://ceur-ws.org/Vol-1623/paperls6.pdf
|volume=Vol-1623
|authors=Hakan Kutucu,Arif Gursoy,Mehmet Kurt,Urfat Nuriyev
|dblpUrl=https://dblp.org/rec/conf/door/KutucuGKN16
}}
==The Band Collocation Problem: a Library of Problems and a Metaheuristic Approach==
https://ceur-ws.org/Vol-1623/paperls6.pdf
The Band Collocation Problem: a Library
of Problems and a Metaheuristic Approach
Hakan Kutucu1 , Arif Gursoy2 , Mehmet Kurt3 , and Urfat Nuriyev2
1
Karabuk University, Department of Computer Engineering, Karabuk, Turkey
hakankutucu@karabuk.edu.tr,
2
Ege University, Department of Mathematics, Izmir, Turkey
arif.gursoy@ege.edu.tr, urfat.nuriyev@ege.edu.tr
3
Izmir University, Department of Mathematics and Computer Science, Izmir, Turkey
mehmet.kurt@izmir.edu.tr
Abstract. In this paper, we consider the Band Collocation Problem (BCP)
which may find an application in telecommunication networks, to design an
optimal packing of information flows on different wavelengths into groups for
obtaining the highest available cost reduction using wavelength division mul-
tiplexing (WDM) technology. We give a review of its mathematical models.
The linear and nonlinear models have been implemented in GAMS (the Gen-
eral Algebraic Modeling System) and solved using the CPLEX and KNITRO
solvers, respectively. Then, we introduce the BCP Library (BCPLib) includ-
ing 1296 problem instances with different properties that can be accessed at
http://www.izmir.edu.tr/bps. Finally, we improve a simulated annealing (SA)
meta-heuristic to solve the BCP. The proposed algorithm is performed using
two local search methods for several test instances of the BCPLib and com-
pared with the solutions obtained by a genetic algorithm. Experimental results
showed that the proposed algorithm improves the quality of solutions.
Keywords: Bandpass problem, Combinatorial optimization, Mathematical mod-
eling, Telecommunication, Simulated annealing
1 Introduction
The Bandpass Problem (BP) whose first mathematical model was presented in 2009
is a combinatorial optimization problem which may be used in telecommunication
systems [1]. Due to the development in the technology, some problems become invalid
or useless and they need to be updated. In this sense, Nuriyev et. al. announced the
Band Collocation Problem (BCP) by extending the BP due to incompatibility with
real life implementations at the present time [7].
The BP is related to transmitting data over fiber optic networks using the Dense
Wavelength Division Multiplexing (DWDM) technology [1]. The data is transmitted
from a source to other stations on different wavelengths in a single fiber optic cable.
Copyright c by the paper’s authors. Copying permitted for private and academic purposes.
In: A. Kononov et al. (eds.): DOOR 2016, Vladivostok, Russia, published at http://ceur-ws.org
The Band Collocation Problem 465
Stations add/drop data onto/from the cable via an optical device called Add/Drop Mul-
tiplexers (ADM). Special cards in ADMs control each wavelength. They can add/drop
(extract) data at some wavelengths to/from a network path [5]. Stations do not have
to receive all data on the cable.
Each special card is normally responsible for one wavelength. However, according
to the BP, a practical programmable ADM can add/drop multiple wavelengths if they
are neighboring to each other, that is, if they are consecutive. In the BP, a group of
consecutive wavelengths is called bandpass and the length of a bandpass is represented
by a positive integer B called bandpass number.
Companies want to reduce the costs of constructing the network. Actually, this is
the goal of the BP by maximizing the number of bandpasses. The key idea of the BP
is to gather up requested data’s wavelengths at any station.
In order to model this problem, we use a binary matrix corresponding to the network
traffic. Consider a communication network. The communication is conducted on m
different wavelengths to carry data to be sent to n different stations. This situation
is described by a binary matrix A = aij : if data carried on wavelength i = 1, ..., m is
requested by station j = 1, ..., n then aij = 1 otherwise aij = 0.
Consider as an example matrix A shown in Fig. 1 which represents a network traffic.
For this example, if each special card in ADMs controls one wavelength, then we need
21 cards. If some special cards are programmed to handle two consecutive wavelengths
(this means that the bandpass number B is 2), then we would need 13 cards (8 for
bandpasses, see Fig. 1.(a), 5 for single). If some special cards are programmed for three
consecutive wavelengths (this means B is 3), then would we need 21 cards since there
is no any bandpass when B = 3 as it can be seen in Fig. 1.(b).
(a) (b)
Fig. 1. (a) The Bandpasses when B = 2 (b) There is no any bandpass when B = 3.
Special cards in ADMs are expensive and IT managers try to reduce the number
of these cards used in ADM. How is it possible? The answer is to reorder the wave-
lengths. For this purpose, the BP asks reordering of the rows of a given matrix so
that the number of B non-overlapping consecutive 1’s is maximized. Note that in the
communication network, reordering of the rows of the matrix simply corresponds to a
reassignment of the DWDM wavelengths.
Fig. 2 shows a reordering of the rows of the matrix given in Fig. 1. For this re-
ordering, if some special cards are programmed for B = 2 consecutive wavelengths,
then we need 13 cards again (8 for bandpasses, 5 for single). However, if some special
cards are programmed for B = 3 consecutive wavelengths, then we need 13 cards (4
466 Hakan Kutucu et al.
for bandpasses, 9 for single). Thus, this reordering of rows allows the use of fewer cards
when B = 3.
(a) (b)
Fig. 2. (a) A reordering of the rows of the matrix in Fig. 1 and bandpasses when (a) B = 2
and (b) B = 3.
It is clear that to find an optimal solution to the BP, we must perform an exhaustive
search over all row permutations. There are total of m! different permutations of m
wavelengths. This number grows faster than exponentially with m. Therefore, this is
not reasonable. Babayev et al. also have proved that the BP is NP-hard [1].
Recent changes in ADM technology made the BP ineffective. We can explain the
deficiencies of the BP as follows:
Technology allows an ADM to drop a wavelength even if it does not carry any
information. Therefore, (a) a bandpass may contain zero elements. (b) The bandpass
number B may be not fixed. (c) the BP ignores costs of the programmable cards. Thus,
the BCP is proposed.
In the next section, we give a review of the mathematical models of the BCP. In
Section 3, we introduce a library of problem instances for researchers to develop efficient
computational solution methods, open for public use. In Section 4, we improve a sim-
ulated annealing (SA) metaheuristic to solve the BCP and also present computational
results.
2 Mathematical Models of the BCP
In this section, we review all models of the BCP. We first introduce some nota-
tion that will be used throughout the models. Let A = (aij ) be a binary matrix
of dimension m × n, Bk = 2k be a length of bands and ck be a cost of the Bk -
Band, where (k = 0, 1, . . . , t = blog2 mc). π is a permutation of the rows such that
π = (π(1), π(2), ..., π(m)). The goal of the BCP is to find an optimal permutation of
rows of the matrix that minimizes the total cost of Bk -Bands in all columns subject to
the constraints. The decision variables in the models are as follows:
1 if row i is relocated to position r,
xir =
0 otherwise,
k 1 if row i is the first row of a Bk -Band in column j,
yij =
0 otherwise
The Band Collocation Problem 467
and
k 1 if aij is an element of a Bk -Band in column j,
zij =
0 otherwise.
2.1 The Combinatorial Model
The combinatorial formulation of the BCP introduced first by Nuriyev et al. is the
following [7].
t m−B
X Xk +1 X
n
k
Minimize ck yπ(i)j (1)
k=0 i=1 j=1
subject to
m
X
k k
Bk · yπ(l)j ≤ zπ(i)j , k = 0, ..., t, j = 1, ..., n, l = 1, ..., m − Bk + 1, (2)
i=l
Xt m−B
Xk +1 m
X
k
Bk · yπ(i)j ≥ aij , j = 1, ..., n, (3)
k=0 i=1 i=1
l+B
X k −1 X
t
k p
yπ(l)j + yπ(i)j ≤ 1, k = 0, ..., t, j = 1, ..., n, l = 1, ..., m − Bk + 1, (4)
i=l+1 p=0
t
X
k
yπ(i)j ≤ 1, i = 1, ..., m, j = 1, ..., n, (5)
k=0
Xt
k
zπ(i)j ≥ aij , i = 1, ..., m, j = 1, ..., n, (6)
k=0
Xt
k
zπ(i)j ≤ 1, i = 1, ..., m, j = 1, ..., n, (7)
k=0
k k
yij , zij ∈ {0, 1}, i = 1, ..., m, j = 1, ..., n, k = 0, ..., t, (8)
The explanations of each constraint will be given at the end of this section.
2.2 The Linear Model
A linear programming model of the BCP introduced first by Gursoy et al. is the fol-
lowing [4].
t m−B
X Xk +1 X
n
k
Minimize ck yij (9)
k=0 i=1 j=1
subject to
m
X
xir = 1, i = 1, . . . , m (10)
r=1
468 Hakan Kutucu et al.
m
X
xir = 1, r = 1, . . . , m (11)
i=1
l+B
Xk −1 X
m
k
Bk · ylj ≤ aij · xir , k = 0, . . . , t, j = 1, . . . , n, l = 1, . . . , m − Bk + 1 (12)
r=l i=1 t m−Bk +1 m
X X X
k
Bk · yij ≥ aij , j = 1, . . . n (13)
k=0 i=1 i=1
m−BXk +1 Xm
k k
Bk · yij = zij , k = 0, . . . , t, j = 1, . . . , n (14)
i=1 i=1
l+B
X k −1
k
yij ≤ 1, k = 0, . . . , t, j = 1, . . . , n, l = 1, . . . , m − Bk + 1 (15)
i=l t
X
k
yij ≤ 1, i = 1, . . . , m, j = 1, . . . , n (16)
t k=0 m
X X
k
zij ≥ arj · xri , i = 1, . . . , m, j = 1, . . . , n (17)
k=0 r=1
t
X
k
zij ≤ 1, i = 1, . . . , m, j = 1, . . . , n (18)
k=0
k k
xir , yij , zij ∈ {0, 1}, i, r = 1, . . . , m, j = 1, . . . , n, k = 0, . . . , t. (19)
2.3 The Nonlinear Model
A nonlinear programming model of the BCP introduced first by Nuriyev et al. is the
following [8].
t m−B
X Xk +1 X
n
k
Minimize ck yij (20)
k=0 i=1 j=1
subject to
l+B
X k −1 X
m
k k
Bk · ylj ≤ zij · xir , k = 0, . . . , t, j = 1, . . . , n, l = 1, . . . , m − Bk + 1 (21)
r=l i=1
and (10),(11),(13),(15)-(19).
The constraints (10) express the fact that row i must be relocated into one new
position r only, (11) express that only one row i must be relocated to each new position
r, (2), (12) and (21) guarantee to find the coordinates of Bk -Bands, (13) say that the
total length of bands in column j can not be less than the number of 1’s in the same
column, (14) identify the elements of Bk -Band in column j, (4), (5) and (15) guarantee
that no two bands may have a common element, (3) and (16) guarantee that any non-
zero entry of the permuted matrix belongs to a unique band Bk , (6), (7), (17) and (18)
say that each non-zero entry of the permuted matrix has to be an element of a band
Bk . In the models, all decision variables are binary which are indicated by (8) and (19).
The linear and nonlinear models have been implemented in GAMS (the General
Algebraic Modeling System) and solved using the CPLEX and KNITRO solvers, re-
spectively.
The Band Collocation Problem 469
3 Online Library
As an extension of the Bandpass problem, the BCP is NP-hard. Therefore, it is needed
to improve heuristic or metaheuristic algorithms. Potential researchers working on the
BCP may need problem instances to check and compare their solutions.
First of all, we focused on what the comparison criteria of the solution algorithms
would be. Main question was what the changing of the results was depending on? So,
we considered all components of the BCP such as number of rows, number of columns,
cost of a Bk − Band, density of the binary matrix.
Solution algorithms can be compared according to different number of rows, columns
and different density of the binary matrix. Here, the density of a matrix is the ratio of
the number of its nonzero entries to the total number of its entries.
It is clear that the upper bound of the length of a Bk −Band depends on the number
of rows. For example, if the number of rows equals 12, then there are B0 = 20 = 1,
B1 = 21 = 2, B2 = 22 = 4 and B3 = 23 = 8 bands. The costs c1 , c2 , c3 and c4 are
corresponding to each of these bands. Now, we face with a new question: How should we
determine the costs? In particular, how should the price be increased from ci to ci+1 ?
It cannot be random, since it is not suitable to compare algorithms. So, we identified
six growth rates (ρ = 0.05, 0.1, 0.2, 0.3, 0.4 and 0.5) for this purpose, and used the
following formula for all problem instances:
ck+1 = (2 − ρ) · ck , where k = 0, 1, ..., t = blog2 mc
In the library, there are three different matrices of the same type. (TX-M1, TX-M2
and TX-M3), there are six different cost alternatives for the same matrix.
There are currently available 72 matrix types, 216 different matrices and 1296 prob-
lem instances. Potential contributors to this library can provide their instances. Infor-
mation on new instances or optimum solutions for library problems is also appreciated
(to be included into the library). At this point we should note that a solution must
contain the permutation (ordering) of rows of the matrix, the starting position of each
a Bk − Band in all columns and the total cost.
The Band Collocation Problem Library (BCPLib) is available online at
http://www.izmir.edu.tr/bps. There are details about instances on “Remarks for In-
stances and Solutions” page. Fig. 3 shows the page of the Instances of Unknown Optimal
Value which includes both the properties of 1296 instances and their excel file.
There will be properties of instances of known optimal value in “Instance of Known
Optimal Value” and will be an excel file which will be included them. When we find
some solutions to the instances, we will share results in “Solution for Instances of Un-
known Optimal Value” and “Solutions for Instances of Known Optimal Value” pages.
Fig. 4 shows an excel file which includes instances. Each work sheet in the file has
a TXMY matrix and six cost ratios.
4 A Solution Approach to the BCP Using the Simulated
Annealing
Simulated annealing (SA) is a popular metaheuristic approach for solving combinatorial
optimization problems. The key point of simulated annealing is that it provides to find
470 Hakan Kutucu et al.
Fig. 3. Properties of the Instances.
Fig. 4. The excel file of all instances.
a global optimum by escaping from local optima. SA introduced by Kirkpatrick et
al. was inspired by the annealing process of physical systems [6]. At each iteration
of a simulated annealing algorithm applied to an optimization problem, the objective
function generates values for two solutions (the current solution and a newly produced
solution) are compared. Better solutions are always accepted, while a fraction of bad
The Band Collocation Problem 471
solutions are accepted in the hope of escaping local optima in search of global optima.
The probability of accepting bad solutions depends on a temperature parameter, which
is typically geometrically decreasing with each iteration of the algorithm.
The pseudocode of the proposed SA algorithm is given in Algorithm 1. The proposed
SA utilizes two reproduction operators, separately: one is to reverse a subset of the
rows, the other is to swap two random rows. The first reproduction operator is actually
2−opt local search method first proposed by Croes [2] for solving the traveling salesman
problem. The second one is a well-known λ − interchange mechanism introduced by
Osman and Christofides [10]. In each reproduction operator, the initial temperature
was set to 5000. The cooling parameter α = 0.9998. The SA terminates when the
temperature falls down 0.1 as indicated line 4 of Algorithm 1. The other possible
termination condition is independent of the temperature and says that the simulated
annealing can stop when there is no change in the cost of the matrix after a fixed number
of iterations. Another termination condition is that the current cost value equals the
optimum value. However, the optimum values of the test instances in BCPLib are not
known, yet. Therefore, the SA algorithm lasts until the temperature falls down 0.1
which requires more time. The cost of the current solution (in line 5 of Algorithm 1)
is computed by a dynamic programming algorithm proposed by Nuriyeva in [9].
Algorithm 1: Pseudocode of the SA.
1 Set the initial temperature T ;
2 Set the cooling parameter α;
3 Generate an initial solution π and calculate its cost f (π);
4 while T > 0.1 do
5 Generate a new solution π 0 by a reproduction operator and calculate its cost f (π 0 );
6 ∆f = f (π 0 ) − f (π);
7 if ∆f < 0 then
8 π = π0 ;
9 else if random[0, 1] < exp(− ∆f
T
) then
10 π = π0 ;
11 else
12 Discard π 0 ;
13 end
15 /* Update temperature */
16 T = T × α;
17 end
4.1 Computational experiments
Our algorithm have been implemented in C and tested on i7-5600U machine with
a 2.60 GHz processor and 8GB RAM. The test problems are taken from the BCPLib
described above [11]. We perform 50 independent runs to get reliable statistical results.
The entries of Table 1 and 2 presented below are:
472 Hakan Kutucu et al.
Initial Cost : the cost value of the input matrix,
m : the number of rows in the matrix,
d : the density of 1’s in the matrix,
n : the number of columns in the matrix,
ratio : the cost ratio ρ,
Best Genetic : the best value obtained by the genetic algorithm among
12 crossover and mutation techniques,
Best : the best value obtained by the SA,
Worst : the worst value obtained by the SA,
Average : the average value among 50 runs,
Improvement : the relative improvement between the best genetic value and our
best value: (Best Genetic−Best)
(Best Genetic)
× 100,
Time : average CPU time in seconds.
The computational results presented in Table 1 and 2 show that our proposed SA
using both local search methods gives better results than the results obtained by the
genetic algorithm reported in [11]. The genetic algorithm and our proposed SA found
the same best values for T1-M1-R10, T1-M1-R30, T3-M1-R10. This is probably that
these values are the optimum values for the instances. As it can be seen λ−interchange
local search method compared to 2 − opt is better in terms of the solution quality.
The Band Collocation Problem 473
5 Conclusion
In this paper, we considered the Band Collocation Problem. We gave a review of its
mathematical models. The linear and nonlinear models have been implemented in
GAMS (the General Algebraic Modeling System) and solved using the CPLEX and
KNITRO solvers, respectively. We also introduced the Band Collocation Problem Li-
brary (BCPLib) which is meant to provide researchers with a set of test problems
having various properties. Generating problem instances whose optimal results known
is the subject of future work. Finally, we improved a simulated annealing (SA) meta-
heuristic to solve the BCP. The proposed algorithm was performed using two local
search methods for several test instances of the BCPLib and compared with the solu-
tions obtained by a genetic algorithm. Experimental results showed that the proposed
algorithm improves the quality of solutions. Moreover, λ − interchange local search
method is better than 2 − opt local search method for the BCP in the SA algorithm.
Acknowledgement
The authors would like to thank the anonymous referees for their valuable comments
that considerably improved the presentation of the paper. This work is supported by
the Scientific and Technological Re-search Council of Turkey-TUBITAK 3001 Project
(Project No.:114F073)
References
1. Babayev, D. A., Bell, G. I. and Nuriyev, U. G.: The Bandpass Problem ”Combinatorial
Optimization and Library of Problems. J. Comb. Optim. 18, 151–172 (2009)
2. Croes, G. A.: A method for solving traveling salesman problems. Operations Res. 6, 791–
812 (1958)
3. GAMS development Corporation (GDC), General Algebraic Modeling Systems (GAMS),
Washington, DC. www.gams.com (2008)
4. Gursoy, A., Kutucu, H., Kurt, M., Nuriyev, U.: A Binary Integer Linear Programming
Model for the Band Collocation Problem. Theoretical and Applied Aspects of Cybernetics,
TAAC’2015, 131–134 (2015)
5. Kaminov, I.P. et al.: A Wideband All Optical WDM Network. IEEE Journal on Selected
Areas in Communications 14 (5), 780–799 (1996)
6. Kirkpatrick, S., Gelatt, C.D. and Vecchi, M.P.: Optimization by Simulated Annealing.
Science 220, 671 – 680 (1983)
7. Nuriyev, U., Kurt, M., Kutucu, H., Gursoy, A.: The Band Collocation Problem and Its
Combinatorial Model. The International Conference Mathematical and Computational
Modelling in Science and Technology, August 2-7, 140–141 (2015)
8. Nuriyev, U., Kutucu, H., Kurt, M. and Gursoy, A.:The band collocation problem in telecom-
munication networks. Proceedings of the 5th International Conference “Control and Opti-
mization with Industrial Applications” (COIA-2015), 362–365 (2015)
9. Nuriyeva F.:On a generalized sequential partially covering problem. Appl. Comput. Math.,
15(2) 240–243 (2016)
474 Hakan Kutucu et al.
10. Osman, I.H. and Christofides, N.: Capacitated clustering problems by hybrid simulated
annealing and tabu search. International Transactions in Operational Res., 1(3) 317 – 336
(1994)
11. The Band Collocation Problem Online Library (BCPLib), http://www.izmir.edu.tr/bps
(Lad: 25.04.2016)
The Band Collocation Problem 475
Table 1. Computational Results using 2 − opt local search method.
Instance Initial Cost m n d ratio Best Best Worst Average improvement Time
(%) Genetic (%) (sec.)
T1-M1-R10 24500 12 6 35 0.1 23140 23140 23140 23140 0.00 1.04
T1-M1-R30 23500 12 6 35 0.3 19660 19660 19660 19660 0.00 1.04
T3-M1-R10 50560 12 6 75 0.1 47680 47680 48330 47910 0.00 1.03
T3-M1-R30 41690 12 6 75 0.3 37340 36640 37640 37154 1.87 1.03
T4-M1-R10 43710 16 8 35 0.1 41960 41770 42060 41873 0.45 2.02
T4-M1-R30 41190 16 8 35 0.3 35710 35270 35920 35391 1.23 2.03
T6-M1-R10 89430 16 8 75 0.1 84820 84530 85730 84982 0.34 1.76
T6-M1-R30 67760 16 8 75 0.3 63780 63270 64440 63685 0.80 1.67
T7-M1-R10 78150 24 10 35 0.1 76700 75890 76800 76350 1.06 3.73
T7-M1-R30 69310 24 10 35 0.3 62950 62760 64760 63729 0.30 3.74
T9-M1-R10 161860 24 10 75 0.1 155870 154440 157650 155910 0.92 3.74
T9-M1-R30 125700 24 10 75 0.3 112050 110750 114640 113036 1.16 3.78
T10-M1-R10 123840 32 12 35 0.1 122780 120720 122630 121640 1.68 6.48
T10-M1-R30 105480 32 12 35 0.3 101370 96830 101400 99021 4.48 6.49
T12-M1-R10 258130 32 12 75 0.1 248400 247710 251640 249082 0.28 6.53
T12-M1-R30 172800 32 12 75 0.3 165360 163670 165670 164507 1.02 6.58
T13-M1-R10 185240 40 14 35 0.1 183170 179590 182350 181075 1.95 10.46
T13-M1-R30 162240 40 14 35 0.3 152800 147970 152100 149791 3.16 10.45
T15-M1-R10 377800 40 14 75 0.1 362300 360060 364960 362707 0.62 10.50
T15-M1-R30 262160 40 14 75 0.3 237600 236350 243210 239335 0.53 10.57
T16-M1-R10 254600 48 16 35 0.1 251980 247600 250990 249478 1.74 15.23
T16-M1-R30 222500 48 16 35 0.3 205570 199890 208520 204040 2.76 15.24
T18-M1-R10 510780 48 16 75 0.1 493430 489480 496280 493317 0.80 15.27
T18-M1-R30 350030 48 16 75 0.3 326330 324820 332620 328721 0.46 15.40
T19-M1-R10 340640 56 18 35 0.1 328690 324820 327460 325948 1.18 20.74
T19-M1-R30 314490 56 18 35 0.3 276330 268590 278200 273353 2.80 20.76
T21-M1-R10 686690 56 18 75 0.1 646570 645950 656520 650793 0.10 20.84
T21-M1-R30 472820 56 18 75 0.3 424070 420690 437540 431766 0.80 20.98
T22-M1-R10 425490 64 20 35 0.1 411140 405340 411210 407808 1.41 27.18
T22-M1-R30 375220 64 20 35 0.3 327420 315620 329530 323386 3.60 27.18
T24-M1-R10 859860 64 20 75 0.1 825270 824920 832590 827887 0.04 27.31
T24-M1-R30 489600 64 20 75 0.3 484440 481980 489600 486262 0.51 27.49
T25-M1-R10 527560 72 22 35 0.1 517920 509940 514550 511843 1.54 35.12
T25-M1-R30 462480 72 22 35 0.3 426750 411170 429560 420558 3.65 35.31
T27-M1-R10 1065400 72 22 75 0.1 1017110 1011590 1025480 1018535 0.54 35.43
T27-M1-R30 624480 72 22 75 0.3 599250 592350 604780 598669 1.15 35.58
T28-M1-R10 641050 80 24 35 0.1 622670 607340 616550 612774 2.46 44.12
T28-M1-R30 568120 80 24 35 0.3 493910 480370 502370 489237 2.74 44.25
T30-M1-R10 1314500 80 24 75 0.1 1206320 1197910 1217230 1206502 0.70 44.54
T30-M1-R30 777680 80 24 75 0.3 703520 695570 714690 704054 1.13 45.36
T31-M1-R10 767110 88 26 35 0.1 737840 719090 728970 723282 2.54 60.07
T31-M1-R30 680630 88 26 35 0.3 565920 539520 561860 549678 4.66 54.63
T33-M1-R10 1514390 88 26 75 0.1 1446220 1434670 1460990 1448456 0.80 54.69
T33-M1-R30 921580 88 26 75 0.3 861830 858160 879990 869746 0.43 56.27
T34-M1-R10 892920 96 28 35 0.1 881670 868240 875100 871299 1.52 65.91
T34-M1-R30 778400 96 28 35 0.3 716030 704370 726510 714474 1.63 65.69
T36-M1-R10 1792900 96 28 75 0.1 1730450 1723880 1750670 1738069 0.38 66.30
T36-M1-R30 1069280 96 28 75 0.3 1040420 1029500 1053770 1042701 1.05 66.61
T37-M1-R10 225600 56 12 35 0.1 216180 211830 214820 213401 2.01 14.00
T37-M1-R30 207400 56 12 35 0.3 177850 168990 176960 173282 4.98 14.03
T39-M1-R10 457780 56 12 75 0.1 426140 422620 432070 426282 0.83 14.04
T39-M1-R30 315410 56 12 75 0.3 278760 276150 286270 280926 0.94 14.17
T40-M1-R10 253890 64 12 35 0.1 241010 236840 240670 239164 1.73 16.51
T40-M1-R30 220980 64 12 35 0.3 190230 181710 189890 185774 4.48 16.50
T42-M1-R10 517620 64 12 75 0.1 491910 486020 494450 490491 1.20 16.55
T42-M1-R30 293760 64 12 75 0.3 292150 290300 293760 293402 0.63 16.68
T43-M1-R10 287590 72 12 35 0.1 278070 269870 274070 271595 2.95 19.82
T43-M1-R30 252870 72 12 35 0.3 220250 207960 219100 212597 5.58 19.65
T45-M1-R10 576240 72 12 75 0.1 547710 539530 548340 543795 1.49 19.56
T45-M1-R30 341250 72 12 75 0.3 321260 316430 325900 320419 1.50 20.53
T46-M1-R10 318760 80 12 35 0.1 305810 296210 301170 298161 3.14 24.15
T46-M1-R30 276330 80 12 35 0.3 232970 219240 231080 224844 5.89 27.14
T48-M1-R10 656320 80 12 75 0.1 600090 591300 605790 595983 1.46 23.99
T48-M1-R30 392230 80 12 75 0.3 349090 345570 359660 352293 1.01 22.38
T49-M1-R10 354240 88 12 35 0.1 336450 321850 329660 325968 4.34 24.95
T49-M1-R30 316900 88 12 35 0.3 248310 237010 249990 242347 4.55 25.03
T51-M1-R10 703220 88 12 75 0.1 659660 654950 670830 661264 0.71 25.16
T51-M1-R30 420250 88 12 75 0.3 386880 385770 399620 393017 0.29 25.35
T52-M1-R10 382010 96 12 35 0.1 371540 359100 364510 361886 3.35 27.84
T52-M1-R30 334090 96 12 35 0.3 287280 274980 288230 280537 4.28 27.87
T54-M1-R10 763250 96 12 75 0.1 718540 714850 731960 721593 0.51 28.05
T54-M1-R30 454580 96 12 75 0.3 424010 422410 437350 428571 0.38 28.22
476 Hakan Kutucu et al.
Table 2. Computational Results using λ − interchange local search method.
Instance Initial Cost m n d ratio Best Best Worst Average improvement Time
(%) Genetic (%) (sec.)
T1-M1-R10 24500 12 6 35 0.1 23140 23140 23330 23144 0.00 0.85
T1-M1-R30 23500 12 6 35 0.3 19660 19660 20170 19670 0.00 0.85
T3-M1-R10 50560 12 6 75 0.1 47680 47680 48420 47936 0.00 0.86
T3-M1-R30 41690 12 6 75 0.3 37340 36640 38310 37136 1.87 0.85
T4-M1-R10 43710 16 8 35 0.1 41960 41860 42340 42041 0.24 1.68
T4-M1-R30 41190 16 8 35 0.3 35710 35570 37050 36109 0.39 1.69
T6-M1-R10 89430 16 8 75 0.1 84820 84530 85440 84939 0.34 1.72
T6-M1-R30 67760 16 8 75 0.3 63780 63270 64290 63827 0.80 1.69
T7-M1-R10 78150 24 10 35 0.1 76700 76070 77440 76710 0.82 3.83
T7-M1-R30 69310 24 10 35 0.3 62950 62790 65590 63983 0.25 3.84
T9-M1-R10 161860 24 10 75 0.1 155870 154410 156880 155409 0.94 3.97
T9-M1-R30 125700 24 10 75 0.3 112050 110750 113640 112506 1.16 3.93
T10-M1-R10 123840 32 12 35 0.1 122780 121620 123610 122594 0.94 6.76
T10-M1-R30 105480 32 12 35 0.3 101370 96330 102580 99527 4.97 6.69
T12-M1-R10 258130 32 12 75 0.1 248400 246040 248960 247800 0.95 6.97
T12-M1-R30 172800 32 12 75 0.3 165360 163170 165230 164364 1.32 6.68
T13-M1-R10 185240 40 14 35 0.1 183170 180790 183800 182052 1.30 10.61
T13-M1-R30 162240 40 14 35 0.3 152800 147690 152330 150137 3.34 10.62
T15-M1-R10 377800 40 14 75 0.1 362300 358140 362650 360460 1.15 10.68
T15-M1-R30 262160 40 14 75 0.3 237600 235260 239530 237175 0.98 10.73
T16-M1-R10 254600 48 16 35 0.1 251980 248540 251650 250212 1.37 15.46
T16-M1-R30 222500 48 16 35 0.3 205570 198160 206850 202997 3.60 15.48
T18-M1-R10 510780 48 16 75 0.1 493430 486900 493000 489605 1.32 15.94
T18-M1-R30 350030 48 16 75 0.3 326330 319100 327000 323881 2.22 15.97
T19-M1-R10 340640 56 18 35 0.1 328690 325660 330010 327338 0.92 21.53
T19-M1-R30 314490 56 18 35 0.3 276330 267780 277690 271147 3.09 21.51
T21-M1-R10 686690 56 18 75 0.1 646570 641830 648540 644890 0.73 21.64
T21-M1-R30 472820 56 18 75 0.3 424070 416190 425330 420458 1.86 21.76
T22-M1-R10 425490 64 20 35 0.1 411140 406160 412610 409402 1.21 28.17
T22-M1-R30 375220 64 20 35 0.3 327420 312560 328440 320969 4.54 28.24
T24-M1-R10 859860 64 20 75 0.1 825270 814710 824910 819750 1.28 28.96
T24-M1-R30 489600 64 20 75 0.3 484440 480280 489600 486612 0.86 28.55
T25-M1-R10 527560 72 22 35 0.1 517920 510110 515430 513040 1.51 35.38
T25-M1-R30 462480 72 22 35 0.3 426750 408930 420230 414461 4.18 35.74
T27-M1-R10 1065400 72 22 75 0.1 1017110 1005880 1014550 1009427 1.10 39.02
T27-M1-R30 624480 72 22 75 0.3 599250 590470 598910 595666 1.47 44.02
T28-M1-R10 641050 80 24 35 0.1 622670 610900 616330 613781 1.89 54.54
T28-M1-R30 568120 80 24 35 0.3 493910 470530 488900 479258 4.73 50.56
T30-M1-R10 1314500 80 24 75 0.1 1206320 1189580 1201650 1194789 1.39 44.56
T30-M1-R30 777680 80 24 75 0.3 703520 684100 697060 691123 2.76 44.84
T31-M1-R10 767110 88 26 35 0.1 737840 719550 733890 726716 2.48 54.23
T31-M1-R30 680630 88 26 35 0.3 565920 529220 549270 537201 6.49 54.57
T33-M1-R10 1514390 88 26 75 0.1 1446220 1422480 1432650 1428080 1.64 55.37
T33-M1-R30 921580 88 26 75 0.3 861830 840680 858170 848478 2.45 56.28
T34-M1-R10 892920 96 28 35 0.1 881670 868940 875870 872717 1.44 66.67
T34-M1-R30 778400 96 28 35 0.3 716030 688390 708550 696284 3.86 66.68
T36-M1-R10 1792900 96 28 75 0.1 1730450 1703600 1723180 1712658 1.55 67.04
T36-M1-R30 1069280 96 28 75 0.3 1040420 1014750 1025700 1020757 2.47 67.45
T37-M1-R10 225600 56 12 35 0.1 216180 213450 216320 214827 1.26 14.14
T37-M1-R30 207400 56 12 35 0.3 177850 168770 175100 171903 5.11 14.15
T39-M1-R10 457780 56 12 75 0.1 426140 418630 426090 422009 1.76 14.23
T39-M1-R30 315410 56 12 75 0.3 278760 268530 276790 272117 3.67 14.31
T40-M1-R10 253890 64 12 35 0.1 241010 238630 242690 240730 0.99 16.68
T40-M1-R30 220980 64 12 35 0.3 190230 179560 187270 183624 5.61 16.70
T42-M1-R10 517620 64 12 75 0.1 491910 481710 488650 484858 2.07 16.78
T42-M1-R30 293760 64 12 75 0.3 292150 288690 293760 293520 1.18 16.89
T43-M1-R10 287590 72 12 35 0.1 278070 271020 275350 273560 2.54 19.66
T43-M1-R30 252870 72 12 35 0.3 220250 204450 213230 209032 7.17 19.68
T45-M1-R10 576240 72 12 75 0.1 547710 534920 541480 537892 2.34 19.79
T45-M1-R30 341250 72 12 75 0.3 321260 315680 320820 318202 1.74 19.90
T46-M1-R10 318760 80 12 35 0.1 305810 298190 302830 300189 2.49 22.60
T46-M1-R30 276330 80 12 35 0.3 232970 213650 223490 219515 8.29 22.65
T48-M1-R10 656320 80 12 75 0.1 600090 584150 592600 588024 2.66 22.77
T48-M1-R30 392230 80 12 75 0.3 349090 339870 347420 344442 2.64 22.94
T49-M1-R10 354240 88 12 35 0.1 336450 325350 333560 329375 3.30 25.57
T49-M1-R30 316900 88 12 35 0.3 248310 230280 242380 236380 7.26 26.69
T51-M1-R10 703220 88 12 75 0.1 659660 644100 652570 648823 2.36 31.10
T51-M1-R30 420250 88 12 75 0.3 386880 375280 385330 379620 3.00 31.27
T52-M1-R10 382010 96 12 35 0.1 371540 361130 367440 364329 2.80 34.43
T52-M1-R30 334090 96 12 35 0.3 287280 265990 276520 271182 7.41 34.49
T54-M1-R10 763250 96 12 75 0.1 718540 702880 711630 707132 2.18 34.69
T54-M1-R30 454580 96 12 75 0.3 424010 408520 420580 415028 3.65 34.90