An improved genetic algorithm for causal discovery

doi:10.23919/JSEE.2025.000015

Journal of Systems Engineering and Electronics ›› 2025, Vol. 36 ›› Issue (3): 768-777.doi: 10.23919/JSEE.2025.000015

• • 上一篇

收稿日期:2023-09-13 出版日期:2025-06-18 发布日期:2025-07-10

An improved genetic algorithm for causal discovery

Tengjiao MAO(), Xianjin BU(), Chunxiao CAI, Yue LU(), Jing DU()

Evaluation Research Department of Center for Strategic Assessment and Consulting, Academy of Military Sciences, Beijing 100097, China

Received:2023-09-13 Online:2025-06-18 Published:2025-07-10
Contact: Xianjin BU E-mail:maotengjiao_ams@126.com;ytbxj@163.com;Olivia9608@hotmail.com;jdstarry@aliyun.com
About author:
MAO Tengjiao was born in 1992. He received his M.S degree in operation research from Army Engineering University of PLA in 2021. He is currently pursuing his Ph.D. degree in military operation research at Academy of Military Science. His research interests include causal inference, military assessment, and operation research. E-mail: maotengjiao_ams@126.com

BU Xianjin was born in 1964. He received his Ph.D. degree from National University of Defense Technology of PLA in 2009. He is currently a researcher in the Center for Strategic Assessment and Consulting, Academy of Military Science. His research interests include causal inference, military assessment, and operation research. E-mail: ytbxj@163.com

CAI Chunxiao was born in 1985. He received his Ph.D. degree from Army Engineering University of PLA in 2012. He is currently a researcher in Center for Strategic Assessment and Consulting, Academy of Military Science. His research interests include operation research and systems engineering. E-mail: caichunxiao1007 @163.com

LU Yue was born in 1996. She received her M.S. degree in computer science from University College Dublin in 2019. She is currently a researcher assistant in the Center for Strategic Assessment and Consulting, Academy of Military Science. Her research interests include military assessment and causal inference. E-mail: Olivia9608@hotmail.com

DU Jing was born in 1979. She received his Ph.D. degree from Army Engineering University of PLA. She is currently a researcher in the Center for Strategic Assessment and Consulting, Academy of Military Science. Her research interests include evaluation theory method and intelligent testing. E-mail: jdstarry@aliyun.com
Supported by:
This work was supported by the National Social Science Fund of China (2022-SKJJ-B-084).

摘要/Abstract

Abstract:

The learning algorithms of causal discovery mainly include score-based methods and genetic algorithms (GA). The score-based algorithms are prone to searching space explosion. Classical GA is slow to converge, and prone to falling into local optima. To address these issues, an improved GA with domain knowledge (IGADK) is proposed. Firstly, domain knowledge is incorporated into the learning process of causality to construct a new fitness function. Secondly, a dynamical mutation operator is introduced in the algorithm to accelerate the convergence rate. Finally, an experiment is conducted on simulation data, which compares the classical GA with IGADK with domain knowledge of varying accuracy. The IGADK can greatly reduce the number of iterations, populations, and samples required for learning, which illustrates the efficiency and effectiveness of the proposed algorithm.

Key words: genetic algorithm (GA), causal discovery, convergence rate, fitness function, mutation operator

. [J]. Journal of Systems Engineering and Electronics, 2025, 36(3): 768-777.

Tengjiao MAO, Xianjin BU, Chunxiao CAI, Yue LU, Jing DU. An improved genetic algorithm for causal discovery[J]. Journal of Systems Engineering and Electronics, 2025, 36(3): 768-777.

图/表 12

参考文献 33

1	PEARL J Causal diagrams for empirical research. Biometrika, 1995, 82 (4): 669- 688. doi: 10.1093/biomet/82.4.669
2	TSIRLIS K, LAGANI V, TRIANTAFILLOU S, et al On scoring maximal ancestral graphs with the max-min hill climbing algorithm. International Journal of Approximate Reasoning, 2018, 102 (9): 74- 85.
3	ASLANI B, MOHEBBI S Ensemble framework for causality learning with heterogeneous directed acyclic graphs through the lens of optimization. Computers & Operations Research, 2023, 152 (8): 106- 148.
4	GUL H, AL-OBEIDAT F, AMIN A, et al Hill climbing-based efficient model for link prediction in undirected graphs. Mathematics, 2022, 10 (22): 4265. doi: 10.3390/math10224265
5	ABREU RS D E, NUNES Y T, GUEDES L A, et al A method for detecting causal relationships between industrial alarm variables using transfer entropy and K2 algorithm. Journal of Process Control, 2021, 106, 142- 154. doi: 10.1016/j.jprocont.2021.09.001
6	XIAO Y, WANG D Y, GAO Y A mobile Bayesian network structure learning method using genetic incremental K2 algorithm and random attribute order technology. Scientific Programming, 2021, (9): 52- 58.
7	HONG Y H, HAO Z F, MAI G Z, et al Causal discovery combining K2 with brain storm optimization algorithm. Molecules, 2018, 23 (7): 1729.
8	GIUDICE E, KUIPERS J, MOFFA G The dual PC algorithm for structure learning. Machine Learning, 2021, (7): 142- 151.
9	BISWAS R, MUKHERJEE S Consistent causal inference from time series with PC algorithm and its time-aware extension. Statistics and Computing, 2024, 34 (1): 217- 237.
10	BISWAS R, SHLIZERMAN E Statistical perspective on functional and causal neural connectomics: the time-aware PC algorithm. PLOS Computational Biology, 2022, 18 (11): e1010653. doi: 10.1371/journal.pcbi.1010653
11	RUSSO F, TONI F Shapley-PC: constraint-based causal structure learning with shapley values. Machine Learning, 2024, 2, 1- 21.
12	POORANI M, NITHYA P, UMAMAHESHWARI B A method for mining infrequent causal associations with swarm intelligence optimization for finding adverse drug reaction. Semantic Scholar, 2014, 3 (1): 25- 32.
13	CAI R C, QIAO J, ZHANG K, et al. Causal discovery from discrete data using hidden compact representation. Proc. of the 32nd Conference on Neural Information Processing Systems, 2018: 2666−2674.
14	CHEN W, CAI R C, ZHANG K, et al Causal discovery in linear non-Gaussian acyclic model with multiple latent confounders. IEEE Trans. on Neural Networks and Learning Systems, 2022, 33 (7): 816- 2827.
15	CAI R C, ZHANG Z J, HAO Z F, et al. Sophisticated merging over random partitions: a scalable and robust causal discovery approach. IEEE Trans. on Neural Networks & Learning Systems, 2017, 29(99): 3623−3635.
16	XIE F, CAI R C, ZENG Y, et al An efficient entropy-based causal discovery method for linear structural equation models with IID noise variables. IEEE Trans. on Neural Networks and Learning Systems, 2020, 31 (5): 1667- 1680. doi: 10.1109/TNNLS.2019.2921613
17	WANG Y T, XUE K, QIAN C. Evolutionary diversity optimization with clustering-based selection for reinforcement learning. Proc. of the International Conference on Learning Representations, 2022: 1−17.
18	BOGGI S, FANO W G Numerical response and causality study of the magnetic permeability-frequency function of NiZn ferrites using genetic algorithm. Journal of Magnetism and Magnetic Materials, 2020, 500, 166- 176.
19	MARICA V G, HOROBET A Conditional granger causality and genetic algorithms in VAR model selection. Symmetry, 2019, 11 (8): 100- 114.
20	MURARI A, ROSSI R, MICHELA G Combining neural computation and genetic programming for observational causality detection and causal modelling. Artificial Intelligence Review, 2022, 56 (7): 6365- 6401.
21	DENG O, NISHIMURA S, OGIHARA A, et al Evolutionary causal discovery with relative impact stratification for interpretable data analysis. Machine Learning, 2024, 4, 163- 181.
22	CHEN Z Y, LIU K, JIAO W P. A genetic algorithm for causal discovery based on structural causal model. Proc. of the Lecture Notes in Computer Science, 2022: 39−54.
23	CESAR B, JUAN V, GRACIELA R, et al Wiener-Granger causality theory supported by a genetic algorithm to characterize natural scenery. Electronics, 2019, 8 (7): 726- 741. doi: 10.3390/electronics8070726
24	VAFAEE F. Learning the structure of large-scale Bayesian networks using genetic algorithm. Proc. of the Association for Computing Machinery, 2014: 855−862.
25	DAI J G, REN J, DU W C, et al An improved evolutionary approach-based hybrid algorithm for bayesian network structure learning in dynamic constrained search space. Neural Computing and Applications, 2018, 32 (5): 1413- 1434.
26	ORPHANOU K, THIERENS D, BOSMAN P. Learning Bayesian network structures with GOMEA. Proc. of the Association for Computing Machinery , 2018: 1007−1015.
27	ZHANG W J, FANG W, SUN J, et al. Learning bayesian networks structures with an effective knowledge-driven GA. Proc. of the IEEE Congress on Evolutionary Computation, 2020. DOI: 10.1109/CEC48606.2020.9185884.
28	CONTALDI C, VAFAEE F, NELSON P Bayesian network hybrid learning using an elite-guided genetic algorithm. Artificial Intelligence Review, 2018, 52 (1): 245- 272.
29	CONTALDI C, VAFAEE F, NELSON P. The role of crossover operator in Bayesian network structure learning performance. Proc. of the Genetic and Evolutionary Computation Conference, 2017: 768−776.
30	AMIRKHANI H, RAHMATI M, LUCAS P, et al Exploiting experts’ knowledge for structure learning of Bayesian networks. IEEE Trans. on Pattern Analysis and Machine Intelligence, 2017, 39 (11): 2154- 2170. doi: 10.1109/TPAMI.2016.2636828
31	ZHU S Y, NG I, CHEN Z T. Causal discovery with reinforcement learning. Proc. of the International Conference on Learning Representations, 2020: 1−17.
32	YAO L Y, CHU Z X, LI S, et al A survey on causal inference. ACM Transactions on Knowledge Discovery from Data, 2021, 15 (5): 785- 796.
33	FELIPPE H, VIOL A, DE ARAUJO D B, et al The Von Neumann entropy for the pearson correlation matrix: a test of the entropic brain hypothesis. Physics, 2021, 20, 74- 81.

Scenario	Sample size
Scenario 1	720
Scenario 2	100
Scenario 3	450
Scenario 4	720

An improved genetic algorithm for causal discovery

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 33

相关文章 0

编辑推荐

Metrics

本文评价

Algorithm	Iteration number	Population size
GA	41	No convergence
IGADK	19	24