GSGP-Hardware: FPGA implementation of GSGP

Created by W.Langdon from gp-bibliography.bib Revision:1.8564

@InProceedings{maldonado:2025:GECCOcomp,

• author = "Yazmin Maldonado and Ruben Salas and Joel A. Quevedo and Rogelio Valdez and Leonardo Trujillo",
• title = "{GSGP-Hardware:} {FPGA} implementation of {GSGP}",
• booktitle = "Proceedings of the 2025 Genetic and Evolutionary Computation Conference: Hot off the Press",
• year = "2025",
• editor = "Eric Medvet",
• pages = "45--46",
• address = "Malaga, Spain",
• series = "GECCO '25 Companion",
• month = "14-18 " # jul,
• organisation = "SIGEVO",
• publisher = "Association for Computing Machinery",
• publisher_address = "New York, NY, USA",
• keywords = "genetic algorithms, genetic programming, geometric semantic genetic programming, VHDL, FPGA",
• isbn13 = "979-8-4007-1464-1",
• URL = "https://doi.org/10.1145/3712255.3734246",
• DOI = "doi:10.1145/3712255.3734246",
• size = "2 pages",
• abstract = "Geometric Semantic Genetic Programming (GSGP) enhances traditional GP by using search operators that act on the syntax of solutions with bounded effects on semantics, transforming symbolic regression into a uni-modal search problem. CPU and GPU implementations of GSGP have demonstrated efficiency by eliminating the need to evaluate variable-length syntactic models, but they remain unsuitable for real-time or resource-constrained scenarios. This paper presents GSGP-Hardware, a fully pipelined and parallel FPGA implementation developed using VHDL. Our solution features a custom fixed-point arithmetic division algorithm, concurrent individual evaluation, and optimal mutation step computation. Validated on five real-world benchmarks and synthesized using Vivado AMD-Xilinx, GSGP-Hardware achieves competitive predictive accuracy while delivering unprecedented performance gains; three orders of magnitude in runtime and four orders of magnitude in GPops/s compared to state-of-the-art GPU implementations. This work enables instantaneous symbolic regression with significantly lower power consumption, opening new possibilities for evolutionary algorithms in domains where computational resources and energy efficiency are limited.",
• notes = "GECCO-2025 A Recombination of the 34th International Conference on Genetic Algorithms (ICGA) and the 30th Annual Genetic Programming Conference (GP)",

}

Genetic Programming entries for Yazmin Maldonado Robles Ruben Dario Salas Villegas Joel Antonio Quevedo Felix Rogelio Valdez Leonardo Trujillo

Citations