We introduce SEvoBench, a C++ framework for the benchmarking of evolutionary single-objective optimization. The SEvoBench framework brings the following highlights:
- Modular DE and PSO algorithms implemented at the software level using the Strategy and Builder design patterns, with additional evolutionary algorithms integrated.
- The CEC2014 and CEC2017 benchmark suites are redesigned using the Curiously Recurring Template Pattern (CRTP).
- The CEC2010 large-scale benchmark suite is accelerated using SIMD technology.
- Parallel testing is supported.
- Implemented using C++20, following Modern C++ style.
SEvoBench is still under development, and some features and functionalities of it need improvement.
Note
The full documentation is under development.
If you use this software in your work, please cite it as below.
@software{Yongkang_SEvoBench_A_C_Framework_2025,
author = {Yongkang, Yang},
month = feb,
title = {{SEvoBench:A C++ Framework for Evolutionary Single-Objective Optimization Benchmarking}},
doi= {10.5281/zenodo.10707581} ,
url = {https://github.com/yangyongkang2000/SEvoBench},
version = {2.0.0},
year = {2025}
}