John D Lamb

Links

University of Aberdeen ·
Contact details ·
Research interests ·
Publications ·
Diary ·
Personal page ·

Computer code

You can use the computer code here under the terms of the GNU General Public License. We ask also that if you use it in any published work, you acknowledge the authors of the code appropriately. Please contact J.D.Lamb@abdn.ac.uk if you have any queries about using this code.

C++ code for portfolio selection by interior-point optimisation

This code was developed by John D Lamb and uses the GNU Scientific Library, ccgsl and ipo. It has optional components that use the commercial ILOG CPLEX library.

C++ code for Data Envelopment Analysis of investment funds

This code was developed by John D Lamb and uses the GNU Scientific Library. It has optional components that use the commercial ILOG CPLEX library.

The code is designed to run on a GNU Linux/unix operating system and has not been tested on any variety of Microsoft Windows.

C++ code for central location problems

This code was developed by John D Lamb and uses the Boost Graph Library and the GNU Scientific Library. It has optional components that use the commercial ILOG CPLEX library.

The code is designed to run on a GNU Linux/unix operating system and has not been tested on any variety of Microsoft Windows.

Connected graphs based on TSPLIB

TSPLIB is a well-known library of test cases for TSP algorithms. The TSPLIB problems are defined on complete graphs. For many location problems, it is useful to have test graphs that are 2-connected but not complete.

The graphs in the collection below are created from the TSPLIB graphs. Each graph is created from the corresponding TSPLIB graph by adding edges randomly until a 2-connected graph incident with all the original nodes is found. Where there is an edge, its length is identical to that of the corresponding TSPLIB graph.

The collection contains a graph for each TSPLIB graph on 101 or fewer nodes. For the format of the files, see the TSPLIB documentation.

C++ code for supervisor–searcher algorithms

This code was developed by John D Lamb and Steve Liu and uses the GNU Scientific Library. It has optional components that use MPI for parallel computation.

The code is designed to run on a GNU Linux/unix operating system. It can be compiled on newer versions of Microsoft windows using the gcc compiler. We have not tested it extensively on this platform.