Empirical study of self-configuring genetic programming algorithm performance and behaviour

17th International Scientific Conference "RESHETNEV READINGS"

Semenkin, E.^1 ; Semenkina, M.^1

Siberian State Aerospace University Named after Academician M.F. Reshetnev, 31 KrasnoyarskiyRabochiy prospect, Krasnoyarsk

660014, Russia^1

关键词: Algorithm performance;    Comparative analysis;    Performance improvements;    Real-world problem;    Selective pressure;    Self-configuring genetic programming;    Symbolic programming;    Uniform crossover operator;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/70/1/012004/pdf
DOI  :  10.1088/1757-899X/70/1/012004

来源: IOP

【 摘 要 】

The behaviour of the self-configuring genetic programming algorithm with a modified uniform crossover operator that implements a selective pressure on the recombination stage, is studied over symbolic programming problems. The operator's probabilistic rates interplay is studied and the role of operator variants on algorithm performance is investigated. Algorithm modifications based on the results of investigations are suggested. The performance improvement of the algorithm is demonstrated by the comparative analysis of suggested algorithms on the benchmark and real world problems.