YevheniiSemendiak/tud_master_benchmarks

Master thesis benchmarks analysis and results. The text is here:

👋

The experiments results and analysis

To reproduce the analysis, provided in the master thesis, please perform following steps:

1. Clone this repository.
2. Install the dependencies: pip install seaborn pandas numpy matplotlib.
3. Execute the following notebooks:
   • analyse_parameter_tuning to check the results of parameter tuning for each meta-heuristic.
   • analyze_first_bench to analyze the results of main experiment set, created to verify the proposed concept applicability.
   • analyse_second_bench to analyze the influence of modified version BRISEv2 (will be published soon) configuration influence on the performance of created online selection hyper-heuristic with parameter control in low-level heuristics.

Code

Together with the experiment results, this repository, yet partially, contains a source code of the developed system. Mostly it is represented by the created search space representation approach.

The examples of code usage may be found in the corresponding folder.