Solar GenAI: An Offline LLM-Based Generative AI Agent for Solar Photovoltaics in Malaysia
DOI:
https://doi.org/10.11113/ijic.v16n1-2.689Keywords:
Generative AI, Large Language Model, Local LLM, Solar PhotovoltaicAbstract
The adoption of solar photovoltaic (PV) systems in Malaysia is often limited by fragmented information and restricted access to reliable, localized knowledge. To address this challenge, Solar GenAI is introduced as an offline-capable, generative AI agent built on the Mistral-7B-Instruct model and deployed entirely on local hardware using the Ollama platform. Designed to function without internet connectivity, Solar GenAI ensures data privacy, institutional autonomy, and broad accessibility, particularly in bandwidth-limited or secure environments. The system integrates a Retrieval-Augmented Generation (RAG) pipeline grounded in a curated corpus of Malaysia-specific solar policies and technical documents. This architecture eliminates hallucinations and achieves 100% factual accuracy across representative queries, compared to a 60% hallucination rate in an LLM-only baseline. A caching mechanism further enhances responsiveness, reducing average query latency by 35%. The user interface was developed based on Nielsen’s usability principles to support intuitive interaction across diverse user groups. By combining fully offline LLM deployment, grounded retrieval, and performance optimization, Solar GenAI provides a secure, accessible, and scalable platform to support solar PV awareness and education, institutional applications, and policy implementation in Malaysia.
References
Sustainable Energy Development Authority Malaysia (SEDA). (2021). Malaysia renewable energy roadmap: Pathway towards a low carbon energy system. https://www.seda.gov.my/reportal/myrer/
World Bank Group, Energy Sector Management Assistance Program (ESMAP), & Solargis. (2019). Photovoltaic power potential: Malaysia.
Teoh, A. N., Go, Y. I., & Yap, T. C. (2020). Is Malaysia ready for sustainable energy? Exploring the attitudes toward solar energy and energy behaviors in Malaysia. World, 1(1), 90–103.
Varshney, T. (2023). Introduction to LLM agents. NVIDIA Technical Blog. http://developer.nvidia.com/blog/introduction-to-llm-agents
Aydin, O., Karaarslan, E., Erenay, F. S., & Bacanin, N. (2025). Generative AI in academic writing: A comparison of DeepSeek, Qwen, ChatGPT, Gemini, Llama, Mistral, and Gemma (arXiv Preprint arXiv:2503.04765).
Zhang, L., & Chen, Z. (2025). Opportunities of applying large language models in the building energy sector. Renewable and Sustainable Energy Reviews, 214, 115558.
Liang, M., Hu, Y., Weng, H., Xi, J., & Yin, B. (2025). EnergyGPT: Fine-tuning large language model for multi-energy load forecasting. Renewable Energy, 123313.
Cho, K., & Bahn, H. (2020). A cost estimation model for cloud services and applying to PC laboratory platforms. Processes, 8(1), 76.
Irugalbandara, C., Mahendra, A., Daynauth, R., Arachchige, T. K., Dantanarayana, J., Flautner, K., Tang, L., Kang, Y., & Mars, J. (2024). Scaling down to scale up: A cost-benefit analysis of replacing OpenAI’s LLM with open-source SLMs in production. In 2024 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS) (pp. 280–291).
Government of Malaysia. (2022). National Energy Policy (DTN) 2022–2040. https://myenergystats.st.gov.my/new-energy-tech1
Mekhilef, S., Safari, A., Mustaffa, W., Saidur, R., Omar, R., & Younis, M. (2012). Solar energy in Malaysia: Current state and prospects. Renewable and Sustainable Energy Reviews, 16(1), 386–396.
Chua, S. C., & Oh, T. H. (2012). Solar energy outlook in Malaysia. Renewable and Sustainable Energy Reviews, 16(1), 564–574.
Wahid, S. S. A., Ramli, M. S., Noorden, Z. A., Hassan, K. K., & Azli, S. A. (2017). A review on highlights and feasibility studies on solar energy utilization in Malaysia. In AIP Conference Proceedings (Vol. 1875, Article 030014).
Vaka, M., Walvekar, R., Rasheed, A. K., & Khalid, M. (2020). A review on Malaysia’s solar energy pathway towards carbon-neutral Malaysia beyond the COVID-19 pandemic. Journal of Cleaner Production, 273, 122834.
Lau, L.-S., Choong, Y.-O., Ching, S.-L., Wei, C.-Y., Senadjki, A., Choong, C.-K., & Seow, A.-N. (2022). Expert insights on Malaysia’s residential solar-energy policies: Shortcomings and recommendations. Clean Energy, 6(5), 619–631.
Koerner, S. A., Siew, W. S., Salema, A. A., Balan, P., Mekhilef, S., & Thavamoney, N. (2022). Energy policies shaping the solar photovoltaics business models in Malaysia with some insights on COVID-19 pandemic effect. Energy Policy, 164, 112918.
Kim, H., Kong, Y.-J., Kim, W.-S., Shafian, S., & Kim, K. (2024). Enhancing reproducibility in organic solar cell fabrication via static sequential deposition with cross-linked polymer donor and nonfullerene acceptor. ACS Applied Polymer Materials, 6(8), 5814–5821.
Ibrahim, N. A., Alwi, S. R. W., Abd Manan, Z., Mustaffa, A. A., & Kidam, K. (2024). Climate change impact on solar systems in Malaysia: Techno-economic analysis. Renewable and Sustainable Energy Reviews, 189, 113901.
Shafian, S., Mohd Salehin, F. N., Lee, S., Ismail, A., Mohamed Shuhidan, S., Xie, L., & Kim, K. (2025). Development of organic semiconductor materials for organic solar cells via the integration of computational quantum chemistry and AI-powered machine learning. ACS Applied Energy Materials, 8(2), 699–722.
Zhang, X., Wang, L., Shafian, S., Wang, P., Zhao, Y., Wang, P., Wu, B., Zhai, J., Chen, J., & Sun, L. (2025). Crosslinking-driven chemical homogeneity enhances performance of pre-seeded perovskite solar cells. Small, 21, 2408362.
Salehin, F. N. M., Chelvanathan, P., Goje, A. A., Ludin, N. A., Ibrahim, M. A., & Shafian, S. (2025). Design of blue, green and red colorful semitransparent films using Ag/SnO₂/Ag color filter for integration into solar cells. Results in Physics, 70, 108172.
Lee, S., Yoon, Y. S., Shafian, S., Kim, J. Y., & Kim, K. (2025). Sequential co-deposition of perovskite film: An effective way of tailoring bandgap in all-vacuum processed perovskite solar cells. Small Methods, 9, 2500104.
Shafian, S., Husen, M. N., Xie, L., & Kim, K. (2025). Predicting high-performance perovskite solar cells using AI-based machine learning models. Materials Today Sustainability, 31, 101176.
Zhai, J. X., Xie, L., & Shafian, S. (2025). Advancements in perovskite/CIGS tandem solar cells: Material synergies, device configurations, and economic viability for sustainable energy. Nanotechnology Reviews, 14, 20250196.
Martino, A., Iannelli, M., & Truong, C. (2023). Knowledge injection to counter large language model (LLM) hallucination. In Proceedings of the European Semantic Web Conference (pp. 182–185).
Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N., Küttler, H., Lewis, M., Yih, W.-T., & Rocktäschel, T. (2020). Retrieval-augmented generation for knowledge-intensive NLP tasks. Advances in Neural Information Processing Systems, 33, 9459–9474.
Singh, S., Fore, M., Karatzas, A., Lee, C., Jian, Y., Shangguan, L., Yu, F., Anagnostopoulos, I., & Stamoulis, D. (2024). LLM-dCache: Improving tool-augmented LLMs with GPT-driven localized data caching. In 2024 31st IEEE International Conference on Electronics, Circuits and Systems (ICECS) (pp. 1–4).
Zou, L., Liu, T., Chen, K., Kong, J., & Deng, Y. (2024). InstCache: A predictive cache for LLM serving (arXiv Preprint arXiv:2411.13820).
Nielsen, J. (2005). Ten usability heuristics.
Martinez, W. L. (2011). Graphical user interfaces. Wiley Interdisciplinary Reviews: Computational Statistics, 3(2), 119–133.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 International Journal of Innovative Computing
This work is licensed under a Creative Commons Attribution 4.0 International License.
