Choose your language

PLRP instances

 

 

You will find here the benchmarks of instances used to test the algorithms dedicated to the periodic capacitated location-routing problem, the results obtained by the researchers on these instances and the best known results.

 

Click here to see information on the files and the data of the instances.

 

To mimic a working week, each instance is made of a 7-day cyclic horizon H with 2 idle days (Saturday and Sunday).

The visit frequency of the customers is between once and three times a week.

The allowed set of combinations of visit days is given in Table 1 and forbids visits on two consecutive days.

The demand of customer j on each day l depends on the chosen service combination r. It varies with the number of days separating two visits, but it can be pre-calculated.

 

Table 1. Allowed set of combinations of visit days

Frequency

Combinations of visit days

1

Monday

Tuesday

Wednesday

Thursday

Friday

2

Monday - Thursday

Monday - Friday

Tuesday - Friday

3

Monday - Wednesday - Friday

 

 

 

Known results

 

The following tables give the best known results for each instance (BKR), obtained either by the algorithm in their published version, or during the parameterization phase.

We can also see a lower bound when one is available (LB), and for each algorithm, the columns have the following meanings:

-     Cost = cost of the solution

-          CPU = running time to obtain the solution (algorithm coded in Visual C++ and tested on a Dell Latitude D420, with a 1.2 GHz Intel Centrino Duo, 1 GB of RAM, and Windows XP).

-          Gap/BKR = gap in percentage between the solution and the best known solution

 

Numbers in bold indicate the best known results.

 

Benchmarks  (capacities on vehicles and depots)

 

 

 

IM*

MAPM**

ELSxPath Relinking***

 

BKS1

Cost2

CPU

gap/BKS

Cost2

CPU

gap/BKS

Cost3

CPU

gap/BKS

20-5-1a

78851

84448

0,61

7,10

85519

0,33

8,46

80546,3

1,1

2,15

20-5-1b

75727

78201

0,67

3,27

76662

0,38

1,23

78564,7

1,3

3,75

20-5-2a

80538

84102

0,61

4,43

84729

0,33

5,20

81827,6

1,1

1,60

20-5-2b

62987

68777

0,63

9,19

64969

0,38

3,15

63492,1

1,1

0,80

50-5-1

157632

189219

5,45

20,04

172479

2,45

9,42

158675,6

5,5

0,66

50-5-1b

142359

159656

7,3

12,15

159472

3,42

12,02

144072,4

9,1

1,20

50-5-2

152376

174001

5,99

14,19

159182

2,64

4,47

153138,6

6,6

0,50

50-5-2b

119925

129385

7,09

7,89

122579

2,66

2,21

122152,8

8,3

1,86

50-5-2bis

177649

213426

6,64

20,14

201040

3,11

13,17

180679,9

7,5

1,71

50-5-2bbis

102824

150134

6,75

46,01

115585

3,72

12,41

103866,5

7,6

1,01

50-5-3

162640

201368

6,25

23,81

184369

3,14

13,36

163986,4

7,4

0,83

50-5-3b

114931

127032

6,97

10,53

130163

2,94

13,25

116288,4

7,7

1,18

100-5-1

364324

391857

48,5

7,56

376857

30,05

3,44

371290,7

40

1,91

100-5-1b

239235

252937

51,97

5,73

247511

23,63

3,46

239804,2

41,2

0,24

100-5-2

287093

287093

68,94

0,00

306784

51,95

6,86

292208,9

49,1

1,78

100-5-2b

174963

226817

52,67

29,64

190201

40,36

8,71

177294,5

41,4

1,33

100-5-3

238031

323875

40,61

36,06

260567

25,72

9,47

240650

32,4

1,10

100-5-3b

181334

209797

52,02

15,70

205891

31,7

13,54

182130,2

41,9

0,44

100-10-1

281896

318582

95,8

13,01

317390

28,33

12,59

284559,8

56

0,94

100-10-1b

224751

237064

115,47

5,48

239659

34

6,63

227180,9

78,9

1,08

100-10-2

282437

288222

119,22

2,05

315764

61,09

11,80

286286,7

70,1

1,36

100-10-2b

179047

196934

121,19

9,99

204853

39,56

14,41

181069,6

84,7

1,13

100-10-3

272463

338819

106,69

24,35

314942

42,11

15,59

277844,9

72,4

1,98

100-10-3b

208890

219830

147,94

5,24

225198

37,8

7,81

211735,2

92,3

1,36

200-10-1

464596

564173

781,48

21,43

558309

497,13

20,17

480716,3

351,1

3,47

200-10-1b

401901

432396

1221,27

7,59

445103

519,2

10,75

406137

547,3

1,05

200-10-2

410514

421012

711,8

2,56

464433

1461,47

13,13

414964,3

298,5

1,08

200-10-2b

324121

324121

1283,7

0,00

354458

915,8

9,36

334163,3

644,3

3,10

200-10-3

574398

599822

834,67

4,43

618414

701,89

7,66

575764

426,7

0,24

200-10-3b

356618

356618

871,02

0,00

382912

433,64

7,37

366447,6

460,1

2,76

MOY

229835

254991

226

12,32

252866

166,7

9,37

233251,3

116,4

1,45

1 BKS is the minimal cost obtained during the 5 runs performed for the ELSxPath Relinking method

2 Cost is the solution cost obtained during the single run performed for the IM et MAPM algorithms

3 Cost is the average solution cost obtained during the 5 runs performed for the ELSxPath Relinking method          

 

 

* C. Prodhon, A Metaheuristic for the Periodic Location-Routing Problem, GOR'07 (Operations Research 2007, Saarbrücken, Germany, 5-7/09/2007. Springer, 2008, 6p.

** C. Prodhon, C. Prins. A Memetic Algorithm with Population Management (MA|PM) for the Periodic Location-Routing Problem, In Hybrid Metaheuristics (Hybrid Metaheuristics 2008 Proceedings, September 2008, Málaga, Spain), M.J. Blesa et al. (éd.), Lecture Notes in Computer Science 5296, pp. 43-57, Springer, 2008. ISBN 978-3-540-88438-5.

*** C. Prodhon. An ELSxPath Relinking Hybrid for the Periodic Location-Routing Problem, In Hybrid Metaheuristics (actes de Hybrid Metaheuristics 2009, October 2009, Udine, Italy), M.J. Blesa et al. (éd.), Lecture Notes in Computer Science 5818, pp. 15-29, Springer, 2009. ISBN 978-3-642-04917-0.


Page last updated 06 January 2009.