郑州大学计算智能实验室

Computational Intelligence Laboratory

Sparse Opt


Sparse Optimization

Sparse Optimization test functioncode

According to compressed sensing theory, an unknown sparse or compressive signal can be recovered from a few measured values, which are much less than those used in previous theories such as Nyquist sampling theorem. Many single-objective sparse optimization algorithms have been proposed to solve the sparse optimization problems. However, the regulation parameter in single-objective sparse optimization problem is difficult to be set. Several multi-objective sparse optimization algorithms have been proposed to eliminate the regulation parameter, but the reconstruction accuracy is not satisfactory. These test functions aim to promote the design of novel algorithms to solving spare optimization problems.


CEC2018 Technical Report 'Problem Definitions and Evaluation Criteria for the CEC Special Session on Evolutionary Algorithms for Sparse Optimization'paper

In the sparse optimization test suite of CEC’2018, a set of sparse optimization test problems of various complexities are designed, such as problems with different signal length, various measured values and different sparsities. Some of the designed problems are involved with noise to make it difficult to be solved. In addition, a fair and appropriate evaluation criterion is given to assess the performance of different sparse reconstruction algorithms.







If you use these sparse optimization test problems, please cite our Technical Report:


"J.J. Liang, M. Gong, H. Li, C.T. Yue, and B.Y. Qu, “Problem definitions and evaluation criteria for the CEC special session on evolutionary algorithms for sparse optimization”, Technical Report, Computational Intelligence Laboratory, Zhengzhou University, Zhengzhou, China, Report #2018001, Feb. 2018