CODeM-Toolkit

A Generator for Uncertain Scalable Stochastic Multiobjective Test Problems

The code in this repository is a C++ implementation of the CODeM toolkit that modifies test problems for multiobjective optimization to stochastic problems.

Some examples for stochastic problems that use different features of the toolkit can be found in misc/examples/CODeMProblems.h.

Once compiled, the application allows the user to evaluate an example problem with two sets of decision vectors: random solutions and Pareto optimal solutions for the underlying deterministic problem. For every decision vector, both deterministic objective vector and samples from the stochastic objective vector are produced.

Compile

The file CODeM.pro is configured for Qt framework, and can be compiled using qmake.

Additionally, Makefile can be used to directs make on how to compile and link the program when qmake is not available on the user's machine. This Makefile is intended to be used on Linux systems or Windows systems via mingw.

How to Use Makefile

1. Open a terminal and cd to the root of the project folder.
2. Type make in the terminal. This will compile the project and generate the executable.
3. To run a test, simply type make test after step 2.
4. Type make clean to remove all compiled objects and the executable.

Makefile Customisation

1. Change the value of TARGET to use a user defined name for the executable.
2. Change the value of CXX to use a different compiler. By default the compiler is g++. To compile the project using the Makefile type make
3. Change the command line arguments to make test. The test is configured to run the executable with the --default option. Specialised options can be used instead.

Run

The application accepts command line arguments such as the problem instance, dimensions, number of solutions etc. For a full list of options type CODeM, CODeM -h or CODeM --help.

The output from the program is formatted as Matlab syntax for convenient analysis of the results. To write the results into a file instead of the console, use the option -f FILENAME or --file FILENAME. The program also allows to generate an output in JSON format. Use -j FILENAME or --jsonFile FILENAME for this.

Citation

Please use the following citation when referring to this work in a sceintific publication:

S. Salomon, R.C. Purshouse, I. Giagkiozis, and P.J. Fleming. A Toolkit for Generating Scalable Stochastic Multiobjective Test Problems. In Proceedings of the 2016 Annual Conference on Genetic and Evolutionary Computation, Denver, CO, USA, 2016

@inproceedings{Salomon2016Toolkit,
    address = {Denver, CO, USA},
    author = {Salomon, S. and Purshouse, R.C. and Giagkiozis, I. and Fleming, P.J.},
    booktitle = {Proceedings of the 2016 Annual Conference on Genetic and Evolutionary Computation},
    isbn = {9781450342063},
    title = {{A Toolkit for Generating Scalable Stochastic Multiobjective Test Problems}},
    year = {2016}
}