Sustainable energy planning in industrial supply chains requires a digitally integrated decision architecture capable of modeling uncertainty, aligning stakeholder priorities, and optimizing infrastructure deployment. In this study, we introduced a hybrid AI-driven (Artificial Intelligence) framework that combines expert-based weighting, neutrosophic uncertainty modeling, and adaptive learning to support strategic planning across energy, logistics, and infrastructure domains. The framework began with the Best-Worst Method (BWM) to derive consistent weights for four meta-criteria: Information strength, balance, data reliability, and lever readiness. These weights were applied to five strategic criteria clusters: Energy performance, environmental impact, logistics service, production stability, and risk and resilience, which were evaluated using Interval-Valued Neutrosophic IVN-TOPSIS (Technique for Order Preference by Similarity to Ideal Solution). Each criterion was expressed as an interval-valued neutrosophic number $ {A}_{i} = ({T}_{i}, {I}_{i}, {F}_{i}) $, where $ {T}_{i} $, $ {I}_{i} $, and $ {F}_{i} $ represented degrees of truth, indeterminacy, and falsity, respectively. The closeness coefficient $ {CC}_{i} $ was computed to rank alternatives under uncertainty. These outputs were embedded in a Deep Reinforcement Learning (DRL) agent, where the reward function $ R = (s, a) $ was shaped by the normalized IVN-TOPSIS scores, enabling real-time policy refinement while preserving expert-defined priorities. Applied to OCP's phosphate supply chain, the model revealed that energy and environment jointly account for 55% of the total strategic weight, confirming their dominant role in decarbonization and cost control. This integrated architecture enhances decision robustness, transparency, and operational relevance. While we focused on strategic criteria modeling, in future work, we will incorporate chemical interaction modeling, particularly the stable complexation mechanisms between phosphate components and energy vectors, to further support infrastructure deployment and sustainable logistics optimization.
Citation: Fadoua Tamtam, Mustapha Amzil, Wissam Jenkal, Larbi Yacoubi, Amina Tourabi. Hybrid AI-driven decision architecture for sustainable industrial planning: Integrating BWM, IVN-TOPSIS, and DRL in OCP's digital supply chain[J]. AIMS Energy, 2025, 13(6): 1538-1559. doi: 10.3934/energy.2025057
Sustainable energy planning in industrial supply chains requires a digitally integrated decision architecture capable of modeling uncertainty, aligning stakeholder priorities, and optimizing infrastructure deployment. In this study, we introduced a hybrid AI-driven (Artificial Intelligence) framework that combines expert-based weighting, neutrosophic uncertainty modeling, and adaptive learning to support strategic planning across energy, logistics, and infrastructure domains. The framework began with the Best-Worst Method (BWM) to derive consistent weights for four meta-criteria: Information strength, balance, data reliability, and lever readiness. These weights were applied to five strategic criteria clusters: Energy performance, environmental impact, logistics service, production stability, and risk and resilience, which were evaluated using Interval-Valued Neutrosophic IVN-TOPSIS (Technique for Order Preference by Similarity to Ideal Solution). Each criterion was expressed as an interval-valued neutrosophic number $ {A}_{i} = ({T}_{i}, {I}_{i}, {F}_{i}) $, where $ {T}_{i} $, $ {I}_{i} $, and $ {F}_{i} $ represented degrees of truth, indeterminacy, and falsity, respectively. The closeness coefficient $ {CC}_{i} $ was computed to rank alternatives under uncertainty. These outputs were embedded in a Deep Reinforcement Learning (DRL) agent, where the reward function $ R = (s, a) $ was shaped by the normalized IVN-TOPSIS scores, enabling real-time policy refinement while preserving expert-defined priorities. Applied to OCP's phosphate supply chain, the model revealed that energy and environment jointly account for 55% of the total strategic weight, confirming their dominant role in decarbonization and cost control. This integrated architecture enhances decision robustness, transparency, and operational relevance. While we focused on strategic criteria modeling, in future work, we will incorporate chemical interaction modeling, particularly the stable complexation mechanisms between phosphate components and energy vectors, to further support infrastructure deployment and sustainable logistics optimization.
| [1] |
Setyadi A, Pawirosumarto S, Damaris A (2025) Toward a resilient and sustainable supply chain: Operational responses to global disruptions in the Post-COVID-19 Era. Sustainability 17: 6167. https://doi.org/10.3390/su17136167 doi: 10.3390/su17136167
|
| [2] |
Echeverri-Martínez R, Alfonso-Morales W, Caicedo-Bravo EF (2020) A methodological Decision-Making support for the planning, design and operation of smart grid projects. AIMS Energy 8: 627–651. https://doi.org/10.3934/energy.2020.4.627 doi: 10.3934/energy.2020.4.627
|
| [3] |
Li T, Sun J, Fei L (2025) Application of multiple-criteria decision-making technology in emergency decision-making: Uncertainty, heterogeneity, dynamicity, and interaction. Mathematics 13: 731. https://doi.org/10.3390/math13050731 doi: 10.3390/math13050731
|
| [4] |
Abdullah AG, Utami HP, Gunawan B, et al. (2025) Multi-criteria decision-making for wind power project feasibility: Trends, techniques, and future directions. Clean Eng Technol 27: 100987. https://doi.org/10.1016/j.clet.2025.100987 doi: 10.1016/j.clet.2025.100987
|
| [5] |
Pamučar D, Ecer F, Cirovic G, et al. (2020) Application of improved Best Worst Method (BWM) in real-world problems. Mathematics 8: 1342. https://doi.org/10.3390/math8081342 doi: 10.3390/math8081342
|
| [6] |
Gulum P, Ayyildiz E, Taskin Gumus A (2021) A two level interval valued neutrosophic AHP integrated TOPSIS methodology for post-earthquake fire risk assessment: An application for Istanbul. Int J Disaster Risk Reduct 61: 102330. https://doi.org/10.1016/j.ijdrr.2021.102330 doi: 10.1016/j.ijdrr.2021.102330
|
| [7] |
Mohammadpour M, Mohammadpour A (2025) Cross-Material forecasting-optimization of CO2 adsorption using deep transfer learning and particle swarm optimization: From porous carbons to Metal-Organic frameworks. Sep Purif Technol 378: 134655. https://doi.org/10.1016/j.seppur.2025.134655 doi: 10.1016/j.seppur.2025.134655
|
| [8] |
Imandoust M, Kermani Alghorayshi ST, Zahedi R, et al. (2025) Multi-objective optimization of hybrid solar-wind MSF-RO cogeneration desalination systems using NSGA-Ⅱ to enhancing energy efficiency, cost reduction and sustainability. Sep Purif Technol 378: 134740. https://doi.org/10.1016/j.seppur.2025.134740 doi: 10.1016/j.seppur.2025.134740
|
| [9] |
Li J, Cui H (2025) Knowledge prior deep meta-reinforcement learning-based load frequency control of isolated sustainable energy system considering electricity prosumers. J Clean Prod 504: 145410. https://doi.org/10.1016/j.jclepro.2025.145410 doi: 10.1016/j.jclepro.2025.145410
|
| [10] |
Guo S, Xiao C, Zheng Y, et al. (2021) Removal and potential mechanisms of Cr(Ⅵ) contamination in phosphate mining wasteland by isolated Bacillus megatherium PMW-03. J Clean Prod 322: 129062. https://doi.org/10.1016/j.jclepro.2021.129062 doi: 10.1016/j.jclepro.2021.129062
|
| [11] |
Raza MA, Aman MM, Abro AG, et al. (2023) Modelling and development of sustainable energy systems. AIMS Energy 11: 256–270. https://doi.org/10.3934/energy.2023014 doi: 10.3934/energy.2023014
|
| [12] |
Peura P, Sjöholm P (2015) Sustainable or distributed energy—or both? clarifying the basic concepts of reforming the energy sector. AIMS Energy 3: 241–254. https://doi.org/10.3934/energy.2015.2.241 doi: 10.3934/energy.2015.2.241
|
| [13] |
Tramarico C, Petrillo A, Andrade H, et al. (2025) Advancing circular supplier selection: Multi-criteria perspectives on risk and sustainability. Sustainability 17: 6814. https://doi.org/10.3390/su17156814 doi: 10.3390/su17156814
|
| [14] |
Nguyen PH, Tran TH, Thi Nguyen LA, et al. (2023) Streamlining apartment provider evaluation: A spherical fuzzy multi-criteria decision-making model. Heliyon 9: e22353. https://doi.org/10.1016/j.heliyon.2023.e22353 doi: 10.1016/j.heliyon.2023.e22353
|
| [15] |
Hassoun Nedjar N, Djebbar Y, Djemili L (2023) Application of the analytical hierarchy process for planning the rehabilitation of water distribution networks. Arab Gulf J Sci Res 41: 518–538. https://doi.org/10.1108/AGJSR-07-2022-0110 doi: 10.1108/AGJSR-07-2022-0110
|
| [16] |
Akcan S, Güldeş M (2019) Integrated multicriteria decision-making methods to solve supplier selection problem: A case study in a hospital. J Healthc Eng 2019: 5614892. https://doi.org/10.1155/2019/5614892 doi: 10.1155/2019/5614892
|
| [17] |
Puška A, Štilić A, Pamučar D, et al. (2024) Introducing a novel multi-criteria ranking of alternatives with weights of criterion (RAWEC) model. MethodsX 12: 102628. https://doi.org/10.1016/j.mex.2024.102628 doi: 10.1016/j.mex.2024.102628
|
| [18] |
Melnik-Leroy GA, Dzemyda G (2021) How to influence the results of MCDM?—Evidence of the impact of cognitive biases. Mathematics 9: 121. https://doi.org/10.3390/math9020121 doi: 10.3390/math9020121
|
| [19] |
Cinelli M, Kadziński M, Gonzalez M, et al. (2020) How to support the application of multiple criteria decision analysis? Let us start with a comprehensive taxonomy. Omega 96: 102261. https://doi.org/10.1016/j.omega.2020.102261 doi: 10.1016/j.omega.2020.102261
|
| [20] |
Jia Y, Wang Z (2024) Application of artificial intelligence based on the fuzzy control algorithm in enterprise innovation. Heliyon 10: e28116. https://doi.org/10.1016/j.heliyon.2024.e28116 doi: 10.1016/j.heliyon.2024.e28116
|
| [21] |
Hosseini S, Ivanov D (2020) Bayesian networks for supply chain risk, resilience and ripple effect analysis: A literature review. Expert Syst Appl 161: 113649. https://doi.org/10.1016/j.eswa.2020.113649 doi: 10.1016/j.eswa.2020.113649
|
| [22] |
Kuo HA, Hong TY, Chien CF (2025) A deep reinforcement learning based digital twin framework for resilient production planning under demand uncertainty and an empirical study in semiconductor wafer fabrication. Comput Ind Eng 208: 111389. https://doi.org/10.1016/j.cie.2025.111389 doi: 10.1016/j.cie.2025.111389
|
| [23] |
Sebastian L, Rita S, Al-Absy MSM, et al. (2025) Neutrosophic approach for patient health monitoring using R programming. Discov Appl Sci 7: 261. https://doi.org/10.1007/s42452-025-06662-6 doi: 10.1007/s42452-025-06662-6
|
| [24] |
Wang J, An N (2024) Research on evaluation method of green suppliers under pythagorean fuzzy environment. Sustainability 16: 9124. https://doi.org/10.3390/su16209124 doi: 10.3390/su16209124
|
| [25] |
Balezentis T, Siksnelyte-Butkiene I, Streimikiene D (2021) Stakeholder involvement for sustainable energy development based on uncertain group decision making: Prioritizing the renewable energy heating technologies and the BWM-WASPAS-IN approach. Sustain Cities Soc 73: 103114. https://doi.org/10.1016/j.scs.2021.103114 doi: 10.1016/j.scs.2021.103114
|
| [26] |
Wei YM (2025) A hybrid multi-criteria decision-making framework for the strategic evaluation of business development models. Information 16: 454. https://doi.org/10.3390/info16060454 doi: 10.3390/info16060454
|
| [27] |
Haseli G, Sheikh R, Wang J, et al. (2021) A novel approach for group decision making based on the best-worst method (G-BWM): Application to supply chain management. Mathematics 9: 1881. https://doi.org/10.3390/math9161881 doi: 10.3390/math9161881
|
| [28] |
Zhiyenbayev M, Kunicina N, Mansurova M, et al. (2024) Development of aggregated sustainability indicators for quality of life evaluations in urban areas of the republic of Kazakhstan. Sustainability 16: 9483. https://doi.org/10.3390/su16219483 doi: 10.3390/su16219483
|
| [29] |
Kim YS, Kim MK, Fu N, et al. (2025) Investigating the impact of data normalization methods on predicting electricity consumption in a building using different artificial neural network models. Sustain Cities Soc 118: 105570. https://doi.org/10.1016/j.scs.2024.105570 doi: 10.1016/j.scs.2024.105570
|
| [30] |
Kabassi K (2021) Application of multi-criteria decision-making models for the evaluation cultural websites: A framework for comparative analysis. Information 12: 407. https://doi.org/10.3390/info12100407 doi: 10.3390/info12100407
|
| [31] |
Černá L, Klapita V, Bulková Z (2025) A multi-criteria evaluation framework for railway sidings supporting sustainable freight and strategic infrastructure planning. Sustainability 17: 8372. https://doi.org/10.3390/su17188372 doi: 10.3390/su17188372
|
| [32] |
Czúcz B, Keith H, Maes J, et al. (2021) Selection criteria for ecosystem condition indicators. Ecol Indic 133: 108376. https://doi.org/10.1016/j.ecolind.2021.108376 doi: 10.1016/j.ecolind.2021.108376
|
| [33] |
Petkovová L, Hartman D, Pavelka T (2020) Problems of aggregation of sustainable development indicators at the regional level. Sustainability 12: 7156. https://doi.org/10.3390/su12177156 doi: 10.3390/su12177156
|
| [34] |
Dutta B, Singha T, Goh M, et al. (2019) Post factum analysis in TOPSIS based decision making method. Expert Syst Appl 138: 112806. https://doi.org/10.1016/j.eswa.2019.07.023 doi: 10.1016/j.eswa.2019.07.023
|
| [35] |
Corrente S, Tasiou M (2023) A robust TOPSIS method for decision making problems with hierarchical and non-monotonic criteria. Expert Syst Appl 214: 119045. https://doi.org/10.1016/j.eswa.2022.119045 doi: 10.1016/j.eswa.2022.119045
|
| [36] |
McKeown S (2025) Beyond hamming distance: Exploring spatial encoding in perceptual hashes. Forensic Sci Int Digit Investig 52: 301878. https://doi.org/10.1016/j.fsidi.2025.301878 doi: 10.1016/j.fsidi.2025.301878
|
| [37] |
An D, Wang N (2025) Application of multi-objective decision-making based on entropy weight-TOPSIS method and RSR method in the analysis of orthopedic disease. Discov Artif Intell 5: 208. https://doi.org/10.1007/s44163-025-00462-y doi: 10.1007/s44163-025-00462-y
|
| [38] |
Yalcin Kavus B, Yazici Sahin K, Taskin A, et al. (2025) A hybrid IVFF-AHP and deep reinforcement learning framework for an ATM location and routing problem. Appl Sci 15: 6747. https://doi.org/10.3390/app15126747 doi: 10.3390/app15126747
|
| [39] |
Amiri M, Hashemi-Tabatabaei M, Ghahremanloo M, et al. (2021) A novel model for multi-criteria assessment based on BWM and possibilistic chance-constrained programming. Comput Ind Eng 156: 107287. https://doi.org/10.1016/j.cie.2021.107287 doi: 10.1016/j.cie.2021.107287
|
| [40] |
Alshamrani A, Majumder P, Das A, et al. (2023) An integrated BWM-TOPSIS-I approach to determine the ranking of alternatives and application of sustainability analysis of renewable energy. Axioms 12: 159. https://doi.org/10.3390/axioms12020159 doi: 10.3390/axioms12020159
|
| [41] |
Pitakaso R, Srichok T, Khonjun S, et al. (2025) A hybrid deep reinforcement learning and metaheuristic framework for heritage tourism route optimization in Warin Chamrap's old town. Heritage 8: 301. https://doi.org/10.3390/heritage8080301 doi: 10.3390/heritage8080301
|
| [42] |
Himeur Y, Elnour M, Fadli F, et al. (2023) AI-big data analytics for building automation and management systems: A survey, actual challenges and future perspectives. Artif Intell Rev 56: 4929–5021. https://doi.org/10.1007/s10462-022-10286-2 doi: 10.1007/s10462-022-10286-2
|
| [43] |
Hosseini E, Al-Ghaili AM, Kadir DH, et al. (2024) Meta-heuristics and deep learning for energy applications: Review and open research challenges (2018–2023). Energy Strategy Rev 53: 101409. https://doi.org/10.1016/j.esr.2024.101409 doi: 10.1016/j.esr.2024.101409
|
| [44] |
Tronnebati I, Jawab F, Frichi Y, et al. (2024) Green supplier selection using fuzzy AHP, fuzzy TOSIS, and fuzzy WASPAS: A case study of the moroccan automotive industry. Sustainability 16: 4580. https://doi.org/10.3390/su16114580 doi: 10.3390/su16114580
|
| [45] |
Yan D, Guan Q, Ou B, et al. (2025) Graph-driven deep reinforcement learning for vehicle routing problems with pickup and delivery. Appl Sci 15: 4776. https://doi.org/10.3390/app15094776 doi: 10.3390/app15094776
|
| [46] | Gardner J, Dutta A, Roy S, et al. (2025) Greener deep reinforcement learning: Analysis of energy and carbon efficiency across atari benchmarks. arXiv. https://doi.org/10.48550/arXiv.2509.0527 |
| [47] |
Bousnina D, Guerassimoff G (2022) Deep reinforcement learning for optimal energy management of multi-energy smart grids. In: Nicosia G, Ojha V, La Malfa E, et al. Machine Learning, Optimization, and Data Science, Eds., Cham: Springer International Publishing, 15–30. https://doi.org/10.1007/978-3-030-95470-3_2 doi: 10.1007/978-3-030-95470-3_2
|
| [48] |
Darvish M, Archetti C, Coelho LC (2019) Trade-offs between environmental and economic performance in production and inventory-routing problems. Int J Prod Econ 217: 269–280. https://doi.org/10.1016/j.ijpe.2018.08.020 doi: 10.1016/j.ijpe.2018.08.020
|
| [49] |
Khan AS, Pruncu CL, Khan R, et al. (2020) A trade-off analysis of economic and environmental aspects of a disruption based closed-loop supply chain network. Sustainability 12: 7056. https://doi.org/10.3390/su12177056 doi: 10.3390/su12177056
|
| [50] |
Coffey L, Claudio D (2021) In defense of group fuzzy AHP: A comparison of group fuzzy AHP and group AHP with confidence intervals. Expert Syst Appl 178: 114970. https://doi.org/10.1016/j.eswa.2021.114970 doi: 10.1016/j.eswa.2021.114970
|
| [51] |
Chang KH, Lai HH, Hung BJ (2024) Combining the fuzzy analytic hierarchy process method with the weighted aggregated sum product assessment method to address internet platform selection problems in an environment with incomplete information. Appl Sci 14: 4390. https://doi.org/10.3390/app14114390 doi: 10.3390/app14114390
|
Figures(1) / Tables(15)
Fadoua Tamtam, Mustapha Amzil, Wissam Jenkal, Larbi Yacoubi, Amina Tourabi. Hybrid AI-driven decision architecture for sustainable industrial planning: Integrating BWM, IVN-TOPSIS, and DRL in OCP's digital supply chain[J]. AIMS Energy, 2025, 13(6): 1538-1559. doi: 10.3934/energy.2025057
DownLoad: