Calibrating Agentic AI Autonomy in Semiconductor Wafer Fabrication: A Spherical Fuzzy Multi-Criteria Governance Framework

Authors

1 College of Computing, Georgia Institute of Technology, Atlanta, USA 30332
Corresponding author: ychen841@gatech.edu

DOI:

https://doi.org/10.63646/SPWY7566

Keywords:

Agentic AI; Semiconductor Manufacturing; Spherical Fuzzy Sets; Multi-Criteria Decision Making; Human-AI Collaboration; Autonomy Governance

Abstract

Agentic artificial intelligence is beginning to reach beyond dashboards and recommendations into direct execution on the semiconductor fabrication floor, yet which shop-floor decisions warrant that authority, and how tightly each should stay bounded, remains unsettled. This study develops a spherical fuzzy multi-criteria governance framework for calibrating Agentic AI autonomy across seven wafer-fabrication decisions: predictive maintenance triggering, lot dispatch reprioritization, yield-excursion containment, in-spec recipe parameter adjustment, cross-tool capacity reallocation, quality deviation disposition, and alternate material substitution. Fifteen criteria spanning Automation Readiness, Risk Criticality, and Control Infrastructure Maturity are evaluated by three practitioner panels using spherical fuzzy numbers; criterion weights are derived with the spherical fuzzy Indifference Threshold-based Attribute Ratio Analysis method, and delegation suitability is ranked with spherical fuzzy TOPSIS. A separate non-compensatory dimension-averaging procedure then assigns each decision one of five autonomy tiers, from diagnostic-only to bounded autonomous execution, using the weakest governing dimension rather than an overall score. Lot dispatch reprioritization and predictive maintenance triggering rank highest for delegation suitability, while quality deviation disposition and alternate material substitution rank lowest. Autonomy assignment diverges from this ranking: the top-ranked decision is capped below full autonomy by its risk and control profile in the same way as the second-ranked decision, but for a different governing reason, illustrating that a high delegation-suitability score does not by itself justify a high autonomy ceiling. A six-scenario sensitivity analysis shows the bottom two decisions remain fixed in every scenario while the middle of the ranking reorders modestly, with Spearman correlations to the baseline between 0.929 and 1.000. The findings indicate that regulatory and customer-specification exposure, not raw automation readiness, is the strongest single discriminator among these decisions, and that governance capability rather than technical readiness is the binding constraint on how far autonomy can be extended.

References

[1] C.-Y. Lin, Generative AI for intelligent manufacturing virtual assistants in the semiconductor industry, IEEE Robotics and Automation Letters 10 (4) (2025) 4132–4139. https://doi.org/10.1109/LRA.2025.3544506.

[2] J. Jeon, Y. Sim, H. Lee, C. Han, D. Yun, E. Kim, S.L. Nagendra, M.B.G. Jun, Y. Kim, S.W. Lee, J. Lee, ChatCNC: Conversational machine monitoring via large language model and real-time data retrieval augmented generation, Journal of Manufacturing Systems 79 (2025) 504–514. https://doi.org/10.1016/j.jmsy.2025.01.018.

[3] T. Harries, M. Hartnoll, M. Hafezianrazavi, H. Meek, A. Nassehi, Digital twins for predictive maintenance, Procedia CIRP 118 (2023) 306–311. https://doi.org/10.1016/j.procir.2023.06.053.

[4] M.A. Hatefi, Indifference threshold-based attribute ratio analysis: a method for assigning the weights to the attributes in multiple attribute decision making, Applied Soft Computing 74 (2019) 643–651. https://doi.org/10.1016/j.asoc.2018.10.050.

[5] F. Kutlu Gündoğdu, C. Kahraman, Spherical fuzzy sets and spherical fuzzy TOPSIS method, Journal of Intelligent & Fuzzy Systems 36(1) (2019) 337–352. https://doi.org/10.3233/JIFS-181401.

[6] L. Xu, S. Almahri, S. Mak, A. Brintrup, Multi-agent systems and foundation models enable autonomous supply chains: opportunities and challenges, IFAC-PapersOnLine 58(19) (2024) 795–800. https://doi.org/10.1016/j.ifacol.2024.09.200.

[7] L. Xu, S. Mak, M. Minaricova, A. Brintrup, On implementing autonomous supply chains: a multi-agent system approach, Computers in Industry 161 (2024) 104120. https://doi.org/10.1016/j.compind.2024.104120.

[8] S.M.R. Naqvi, et al., Survey on ontology-based explainable AI in manufacturing, Journal of Intelligent Manufacturing 35 (2024) 3605–3627. https://doi.org/10.1007/s10845-023-02304-z.

[9] A. Kusiak, Federated explainable artificial intelligence (fXAI): a digital manufacturing perspective, International Journal of Production Research 62 (1–2) (2024) 171–182. https://doi.org/10.1080/00207543.2023.2238083.

[10] I. Drori, D. Te'eni, Human-in-the-loop AI reviewing: feasibility, opportunities, and risks, Journal of the Association for Information Systems 25(1) (2024) 98–109. https://doi.org/10.17705/1jais.00867.

[11] B. Zhang, Y. Leng, B. Huo, N. Yu, X. Jin, Editorial: human-in-the-loop system design and control adaptation for behavior-assistant robots, Frontiers in Neuroscience 18 (2024) 1418331. https://doi.org/10.3389/fnins.2024.1418331.

[12] R. Busch, P. Czerner, M. Wahl, B. Choubey, Yield prediction in semiconductor manufacturing using two-step machine learning, International Journal of Production Research (2025). https://doi.org/10.1080/00207543.2025.2601804.

[13] J. Senoner, T. Netland, S. Feuerriegel, Using explainable artificial intelligence to improve process quality: evidence from semiconductor manufacturing, Management Science 68(8) (2022) 5704–5723. https://doi.org/10.1287/mnsc.2021.4190.

[14] F. Kutlu Gündoğdu, A spherical fuzzy extension of MULTIMOORA method, Journal of Intelligent & Fuzzy Systems 38 (1) (2020) 963–978. https://doi.org/10.3233/JIFS-179462.

[15] I.M. Sharaf, A new approach for spherical fuzzy TOPSIS and spherical fuzzy VIKOR applied to the evaluation of hydrogen storage systems, Soft Computing 27(8) (2023) 4403–4423. https://doi.org/10.1007/s00500-022-07749-7.

[16] G.E. Modoni, M. Sacco, A human digital-twin-based framework driving human centricity towards Industry 5.0, Sensors 23(13) (2023) 6054. https://doi.org/10.3390/s23136054.

[17] M. Krupas, E. Kajati, C. Liu, I. Zolotova, Towards a human-centric digital twin for human–machine collaboration: a review on enabling technologies and methods, Sensors 24(7) (2024) 2232. https://doi.org/10.3390/s24072232.

[18] Y. Xia, Z. Xiao, N. Jazdi, M. Weyrich, Generation of asset administration shell with large language model agents: toward semantic interoperability in digital twins in the context of Industry 4.0, IEEE Access 12 (2024) 84863–84877. https://doi.org/10.1109/ACCESS.2024.3415470.

[19] Y. Xia, M. Shenoy, N. Jazdi, M. Weyrich, Towards autonomous system: flexible modular production system enhanced with large language model agents, in: 2023 IEEE 28th International Conference on Emerging Technologies and Factory Automation (ETFA), 2023, pp. 1–8. https://doi.org/10.1109/ETFA54631.2023.10275362.

[20] L. Wang, et al., A survey on large language model based autonomous agents, Frontiers of Computer Science 18 (6) (2024) 186345. https://doi.org/10.1007/s11704-024-40231-1.

[21] M. Müller, T. Müller, B. Ashtari Talkhestani, P. Marks, N. Jazdi, M. Weyrich, Industrial autonomous systems: a survey on definitions, characteristics and abilities, at-Automatisierungstechnik 69 (1) (2021) 3–13. https://doi.org/10.1515/auto-2020-0131.

[22] M.A. Khalifa, A.T. Azar, W. El-Shafai, A comprehensive systematic review of cyber-physical systems security: threats, challenges, and defense in ICS/OT environments, Engineering Reports 8(7) (2026). https://doi.org/10.1002/eng2.70866.

[23] A. Gomez-Cabrera, P.J. Escamilla-Ambrosio, M.A. Al Faruque, A review of resilience enhancement techniques for cyber-physical systems, IEEE Access 13 (2025) 138061–138082. https://doi.org/10.1109/ACCESS.2025.3593257.

[24] J.D. Lee, K.A. See, Trust in automation: designing for appropriate reliance, Human Factors 46(1) (2004) 50–80. https://doi.org/10.1518/hfes.46.1.50_30392.

[25] K.A. Hoff, M. Bashir, Trust in automation: integrating empirical evidence on factors that influence trust, Human Factors 57(3) (2015) 407–434. https://doi.org/10.1177/0018720814547570.

[26] M. Wang, J. Zhang, P. Zhang, M. Jin, Cooperative multi-agent reinforcement learning for multi-area integrated scheduling in wafer fabs, International Journal of Production Research 63(8) (2025) 2871–2888. https://doi.org/10.1080/00207543.2024.2411615.

[27] N.H. Abd Wahab, K. Hasikin, K.W. Lai, K. Xia, L. Bei, K. Huang, X. Wu, Systematic review of predictive maintenance and digital twin technologies: challenges, opportunities, and best practices, PeerJ Computer Science 10 (2024) e1943. https://doi.org/10.7717/peerj-cs.1943.

[28] G.A. Susto, A. Schirru, S. Pampuri, S. McLoone, A. Beghi, Machine learning for predictive maintenance: a multiple classifiers approach, IEEE Transactions on Industrial Informatics 11(3) (2015) 812–820. https://doi.org/10.1109/TII.2014.2349359.

[29] L.A. Zadeh, Fuzzy sets, Information and Control 8(3) (1965) 338–353. https://doi.org/10.1016/S0019-9958(65)90241-X.

[30] K.T. Atanassov, Intuitionistic fuzzy sets, Fuzzy Sets and Systems 20(1) (1986) 87–96. https://doi.org/10.1016/S0165-0114(86)80034-3.

[31] C.L. Hwang, K. Yoon, Multiple Attribute Decision Making: Methods and Applications, Springer-Verlag, Berlin, 1981. https://doi.org/10.1007/978-3-642-48318-9.

[32] S. Jafarzadeh Ghoushchi, S. Shaffiee Haghshenas, A. Memarpour Ghiaci, G. Guido, A. Vitale, Road safety assessment and risks prioritization using an integrated SWARA and MARCOS approach under spherical fuzzy environment, Neural Computing and Applications 35(6) (2023) 4549–4567. https://doi.org/10.1007/s00521-022-07929-4.

[33] M. Helal, Y. Dwivedi, E. Ismagilova, et al., Reinterpreting agency theory in the era of artificial intelligence: conceptualizing delegated agency, AI & Society 41 (2026) 7063–7074. https://doi.org/10.1007/s00146-026-03034-5.

[34] P. Wang, G. Su, W. Yang, P. Jing, Research on safety interlock system design and control experiment of combined support and anchor equipment, Sensors 22(16) (2022) 6058. https://doi.org/10.3390/s22166058.

[35] R.A. Freeman, Reliability of interlocking systems, Process Safety Progress 13(3) (1994) 146–152. https://doi.org/10.1002/prs.680130307.

[36] R. Beregi, G. Pedone, B. Háy, J. Váncza, Manufacturing execution system integration through the standardization of a common service model for cyber-physical production systems, Applied Sciences 11(16) (2021) 7581. https://doi.org/10.3390/app11167581.

[37] T. Song, Y. Qiao, Y. He, J. Li, N. Wu, B. Liu, A new framework of the EAP system in semiconductor manufacturing internet of things, Electronics 12(18) (2023) 3910. https://doi.org/10.3390/electronics12183910.

[38] W. Xiong, D.D. Wu, J.H.Y. Yeung, Semiconductor supply chain resilience and disruption: insights, mitigation, and future directions, International Journal of Production Research 63(9) (2025) 3442–3465. https://doi.org/10.1080/00207543.2024.2387074.

[39] C. Rudin, Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead, Nature Machine Intelligence 1(5) (2019) 206–215. https://doi.org/10.1038/s42256-019-0048-x.

[40] D.K. Das, M.M.H. Mostafa, From assistance to autonomy: nonlinear human factors and system-level impacts on road transportation across SAE levels 0–5, Sustainability 18(12) (2026) 6033. https://doi.org/10.3390/su18126033.

[41] E. Billing, F. Fraboni, L. Gualtieri, P.H. Rosen, P. Thorvald, Editorial: human factors and cognitive ergonomics in advanced industrial human-robot interaction, Frontiers in Robotics and AI 12 (2025) 1564948. https://doi.org/10.3389/frobt.2025.1564948.

[42] L. Wright, S. Davidson, Digital twins for metrology; metrology for digital twins, Measurement Science and Technology 35 (2024) 051001. https://doi.org/10.1088/1361-6501/ad2050.

[43] D. Kim, P. Kang, S. Cho, H.J. Lee, S. Doh, Machine learning-based novelty detection for faulty wafer detection in semiconductor manufacturing, Expert Systems with Applications 39(4) (2012) 4075–4083. https://doi.org/10.1016/j.eswa.2011.09.088.

[44] K. Liu, Y.Q. Chen, T. Zhang, S. Tian, X. Zhang, A survey of run-to-run control for batch processes, ISA Transactions 83 (2018) 107–125. https://doi.org/10.1016/j.isatra.2018.09.005.

[45] M. Valori, G. Prange-Lasonder, J. Saenz, R. Behrens, C. Bidard, E. Lucet, I. Fassi, Editorial: safety in close human-robot interaction, Frontiers in Robotics and AI 10 (2023) 1288713. https://doi.org/10.3389/frobt.2023.1288713.

[46] V. Kulothungan, Using blockchain ledgers to record AI decisions in IoT, IoT 6(3) (2025) 37. https://doi.org/10.3390/iot6030037.

[47] Q. Nie, D. Tang, J. Ding, W. Qian, C. Liu, A multi-agent and vision language model framework of digital twin for shop-floor self-decision-making control towards Industry 5.0 high-performance manufacturing, Journal of Industrial Information Integration 53 (2026) 101161. https://doi.org/10.1016/j.jii.2026.101161.

[48] M.R. Douissa, K. Jabeur, A non-compensatory classification approach for multi-criteria ABC analysis, Soft Computing 24 (2020) 9525–9556. https://doi.org/10.1007/s00500-019-04462-w.

[49] A. Verbruggen, Revocability and reversibility in societal decision-making, Ecological Economics 85 (2013) 20–27. https://doi.org/10.1016/j.ecolecon.2012.10.011.