malbarbo / dcmstp-instances Goto Github PK

A collections of scripts to download instances and solvers for the degree constrained minimum spanning tree problem. It also provides an uniform interface to run the solvers.

Clone this repository using the command

git clone https://github.com/malbarbo/dcmstp-instances

Before usage, it is necessary to download the solvers and instances. Some solvers are available as binaries, while others need to be compiled from source. To download (and build) all solvers these dependencies are necessary: make, curl, gcc, maven and jdk. (In a Debian based system these dependencies can be installed with the command: apt-get install make curl gcc maven openjdk-9-jdk-headless)

After dependencies installation, execute the command make in project directory. Solvers are download to target/solvers directory and instances are download to target/instances directory.

To delete all solvers and instances, delete the target directory (Alert: to avoid removing important files, never put any file in the target directory)

The solve script provides an uniform interface to execute the solvers. For example, to solve the target/instances/variable/ANDINST/tb1ct100_1.txt instance with the solver cgbc, execute

./solve cgbc file target/instances/variable/ANDINST/tb1ct100_1.txt

To check if the produced solution is valid, redirect the output to a file and execute the check script

./solve concorde 2 target/instances/variable/ANDINST/tb1ct100_1.txt > result
./check result

Note that some solvers does not print the solution, in this case, the solution cannot be checked.

Running ./solve in the project directory will show detailed usage information.

• CG-BC, BCP and BCP-BC from [3] (Download links are broken)

• Choco from [6]

• Concorde from [1] (can only be used when dc = 2)

• LKH from [7] (can only be used when dc = 2)

Write a script and put it solvers directory. The script should accept as a command line argument the instance file path and output to stdout tree lines, the first contains the time in seconds used by the solver, the second contains the value of the objective function and the third line the solution (like 1-2 3-4, where 1-2 and 3-4 are edges of the tree). The script can write logging information to stderr.

The script is responsible for converting the input file if necessary, calling the real solver with the appropriated parameters, parsing the output of the solver and printing the expected output. See the scripts in solvers directory to see examples of how this can be done.

The input file is a text file that contains integers numbers. The first two numbers are the number of vertices n and the number of edges m. The following 3 * m values are the edges. Each edge is represented by a pair of numbers (vertices) in the range [1, n] followed by the weight. The last 2 * n values represents the degree constraint of each vertex (pairs of vertex number followed by the degree constrain).

Ok, this format has some redundancies, like the vertex number in the degree constraint section. Also if the graph is complete (all downloaded instances are!) the end vertices of the edges are redundant. Anyway, this format was chosen because it is already used by some solvers and it is easy to parse.

Example of input file with 4 vertices and 6 edges:

4 5
1 3 10
1 4 20
2 3 30
2 4 40
3 4 50
1 1
2 1
3 2
4 2

Vertices 1 and 2 has degree constraint of 1 and vertices 3 and 4 degree constraint of 2.

• CRD, SYM, STR and SHRD instances from [9]

• Misleading and R instances from [8] and [10]

• ANDINST, DE, DR, LH-E and LH-R instances from [2], [5] and [3]

1. D. L. Applegate, R. E. Bixby, V. Chvatál, and W. J. Cook, The Traveling Salesman Problem: A Computational Study. Princeton University Press, 2006. http://www.jstor.org/stable/j.ctt7s8xg

2. R. Andrade, A. Lucena, and N. Maculan, “Using Lagrangian dual information to generate degree constrained spanning trees,” Discrete Applied Mathematics, vol. 154, no. 5, pp. 703–717, Apr. 2006. https://doi.org/10.1016/j.dam.2005.06.011

3. L. H. Bicalho, A. S. da Cunha, and A. Lucena, “Branch-and-cut-and-price algorithms for the Degree Constrained Minimum Spanning Tree Problem,” Computational Optimization and Applications, vol. 63, no. 3, pp. 755–792, Apr. 2016. https://doi.org/10.1007/s10589-015-9788-7

4. T. N. Bui, X. Deng, and C. M. Zrncic, “An Improved Ant-Based Algorithm for the Degree-Constrained Minimum Spanning Tree Problem,” IEEE Transactions on Evolutionary Computation, vol. 16, no. 2, pp. 266–278, Apr. 2012. https://doi.org/10.1109/TEVC.2011.2125971

5. A. S. da Cunha and A. Lucena, “Lower and upper bounds for the degree-constrained minimum spanning tree problem,” Networks, vol. 50, no. 1, pp. 55–66, Aug. 2007. https://doi.org/10.1002/net.20166

6. J. G. Fages, X. Lorca, and L. M. Rousseau, “The salesman and the tree: the importance of search in CP,” Constraints, vol. 21, no. 2, pp. 145–162, Apr.

7. https://doi.org/10.1007/s10601-014-9178-2

8. K. Helsgaun, “General k-opt submoves for the Lin–Kernighan TSP heuristic,” Mathematical Programming Computation, vol. 1, no. 2–3, pp. 119–163, Oct. 2009. https://doi.org/10.1007/s12532-009-0004-6

9. J. Knowles and D. Corne, “A new evolutionary approach to the degree-constrained minimum spanning tree problem,” IEEE Transactions on Evolutionary Computation, vol. 4, no. 2, pp. 125–134, Jul. 2000. https://doi.org/10.1109/4235.850653

10. M. Krishnamoorthy, A. T. Ernst, and Y. M. Sharaiha, “Comparison of Algorithms for the Degree Constrained Minimum Spanning Tree,” Journal of Heuristics, vol. 7, no. 6, pp. 587–611, Nov. 2001. https://doi.org/10.1023/A:1011977126230

11. G. R. Raidl and B. A. Julstrom, “Edge sets: an effective evolutionary coding of spanning trees,” IEEE Transactions on Evolutionary Computation, vol. 7, no. 3, pp. 225–239, Jun. 2003. https://doi.org/10.1109/TEVC.2002.807275

12. P. Martins and M. C. de Souza, “VNS and second order heuristics for the min-degree constrained minimum spanning tree problem,” Computers & Operations Research, vol. 36, no. 11, pp. 2969–2982, Nov. 2009. https://doi.org/10.1016/j.cor.2009.01.013