Skip to content

Leative/MDEIntelligence22-MDO-crossover-evaluation

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Towards a Configurable Crossover Operator for Model-Driven Optimization - Evaluation Data

Evaluation artifacts for the paper Towards a Configurable Crossover Operator for Model-Driven Optimization submitted to MDE Intelligence 2022.

MDEOptimiser

A precompiled version of the optimzation tool MDEOptimiser with integrated crossover operator is contained as a command line version in the folder MDEOptimiser.

Results

The detailed results of the experiments of the evaluation of the paper can be found in folder results. More information about the results can be found in the README contained in that folder.

Note that the raw data generated during the experiments is attached as an archive to the tag 1.0.0 on Github and can be downloaded there.

Usecase

All artifacts of the use case (CRA case) are located in the folder usecases/CRA. The subfolder models contains the EMF meta-model, the instance models and the Henshin model transformations used as mutation and repair operators. The subfolder src contains the java implementation of the meta-model and, in the package guidancefunctions, the implementation of the feasibility constraints and objectives. The package crossoverdefinition contains a class used to specify the problem part of the meta-model. Both packages need a part of the adapted MDEOptimiser as a dependency which is delivered as a local p2 repository in usecases/p2repo. The configuration files used by MDEOptimiser to specify the different variants of evolutionary algorithms are located in the folder configs.

Compiling the use case

The use case is provided as a maven project. To build it, execute the command mvn clean compile in the folder usecases. The build process may take some time as dependencies need to be resolved and downloaded. The successful build was tested with Apache Maven 3.6.3 and Java 11.0.6. As the build configuration uses Tycho 2.0.0 newer Java versions (e.g. Java 16) might not allow for a successful build without adaptions.

Running experiments in batches

To simplify the execution of a batch of experiments a bash script is contained in the main folder. To run 10 evolutionary computations for all configurations of the CRA case, for example, a call to the script would look like:

./experiment-runner.sh "usecases/CRA/target/classes/" "usecases/CRA/configs/" "usecases/CRA/"

Note that the MDEOptimiser argument specifying the path to the project folder (e.g. usecases/CRA) should not contain backwards navigation (e.g., usescases/../usecases/CRA) as an Exception is thrown by MDEOptimiser in this case.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Packages

No packages published