- author = "Yi-Chao Huang and Xin-Xin Xu and Jian-Yu Li and Sam Kwong and Zhi-Hui Zhan and Jun Zhang",
- title = "Generative Evolutionary Computation: An Automatic Gene Targeting Differential Evolution Via Genetic Programming",
- booktitle = "Proceedings of the 2025 Genetic and Evolutionary Computation Conference Companion",
- year = "2025",
- editor = "Aniko Ekart and Nelishia Pillay",
- pages = "623--626",
- 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, evolutionary computation, DE, generative artificial intelligence, generative evolutionary computation: Poster",
- isbn13 = "979-8-4007-1464-1",
-
URL = "
https://doi.org/10.1145/3712255.3726588",
-
DOI = "
doi:10.1145/3712255.3726588",
- size = "4 pages",
- abstract = "Evolutionary computation (EC) is a kind of artificial intelligence (AI) for optimization. However, traditional EC algorithms require the careful design of parameters and/or operators from experts. Designing an operator with generalization ability is a tough task, and the manually design process is often limited by the structures of existing operators, which lacks diversity. In order to explore operators with diverse structures and good generalization ability so as to provide inspiration for the manually design of operators, this paper proposes a generative EC approach for the automatic design of operators by using genetic programming (GP). We apply the generative approach to a typical EC variant named gene targeting differential evolution (GTDE), so as to propose a new automatic GTDE (AGTDE) algorithm. The AGTDE uses the advantage of GP in optimizing structural features to automatically generate and refine the targeting vector generation operator within GTDE. This way, the operator of AGTDE is generated automatically rather than manually designed, which is more robust and optimal. The experimental results demonstrate that AGTDE is capable of identifying appropriate and even better operator for solving optimization problems, when compared with GTDE with manually design operator. Moreover, the results show that the operator obtained for one problem can be applied to other problems and obtain promising results, which reflect the robustness and generalization ability of the operator generated by GP. Therefore, this generative approach may provide some novel avenues of thought for the automatically design of EC algorithm operators.",
-
notes = "CEC2017 Test Suite D100
School of Computer Science and Engineering, South China University of Technology, Guangzhou 510006, China
A Recombination of the 34th International Conference on Genetic Algorithms (ICGA) and the 30th Annual Genetic Programming Conference (GP)",
