Skip to main content

Table 5 Gurobi solver metrics on noisy artificial growing data

From: Global exact optimisations for chloroplast structural haplotype scaffolding

\(\mathsf {|{}V|}\)

\(\mathsf {|{}E|}\)

\(\mathsf {|{}SC1|}\)

\(\mathsf {|{}SC2|}\)

\(\mathsf {|{}IR|}\)

\(|{\mathcal {L}}|\)

Time

Opt.

% Gap

Nodes

Iter.

186

372

20

20

20

152

1.01

39.00

0.00

1

2862

26

24

43

30

5.16

39.00

6888

280

688

20

20

40

212

6.89

79.00

0.00

1

6727

28

23

89

50

12.26

79.00

19930

366

946

20

20

60

262

42.52

123.50

3.64

1

16821

25

26

132

60

79.28

119.00

50246

452

1208

20

20

80

320

90.06

161.50

1.54

1

27822

22

23

181

78

295.22

159.00

129174

556

1366

20

20

100

322

196.04

199.00

0.00

1

42292

23

24

231

40

244.41

199.00

92009

662

1804

20

20

120

412

1007.59

242.50

1.44

1

84639

29

26

276

90

1434.16

239.00

210798

736

1946

20

20

140

454

1108.09

283.00

1.41

1

228198

24

26

318

92

3540.79

279.00

691619

822

2212

20

20

160

514

1118.76

323.00

1.24

1

86592

26

27

358

112

2449.55

319.00

287146

902

2362

20

20

180

542

1591.18

363.00

1.10

1

91936

26

27

398

100

2576.6

359.00

269958

996

2656

20

20

200

602

2294.85

404.00

1.24

1

116315

26

26

446

120

3747.02

399.00

351501

  1. The column descriptions are the same as the one in Table 4. Except that because of the noise on multiplicities, the sum of contig multiplicity for each region can change: this is the value below the number of contig in columns \(\mathsf {|{SC1}|}\), \(\mathsf {|{SC2}|}\) and \(\mathsf {|{IR}|}\). Similarly, because of the noise on the number of links, the below value for \(\mathcal {L}\) corresponds to the number of noisy links