Applying machine learning approach an‎d mathematical models to design sustainable photovoltaic supply chain: A case study in Iran

9/20/2026 12:00:00 AM

Given the environmental an‎d economic consequences of the excessive consumption of fossil fuels, the transition to renewable energy sources, especially solar energy, has been proposed as an inevitable necessity in macro-energy policies. The increasing deman‎d for photovoltaic systems an‎d the increasing growth of installed capacity in developing an‎d developed countries have doubled the necessity of designing efficient an‎d sustainable supply chains. In the meantime, paying attention to recycling an‎d product life cycle plays a key role in improving the efficiency an‎d sustainability of these technologies. In this research, a new approach is presented for designing a sustainable supply chain of solar systems. The proposed framework consists of two main steps: first, selec‎ting the optimal solar panel technology using reinforcement learning algorithms, taking into account technological advances, technology diversity, an‎d environmental an‎d economic uncertainties; an‎d second, designing a closed-loop supply chain network based on a multi-objective optimization model. In this model, economic, environmental an‎d social dimensions are simultaneously considered to provide a comprehensive approach to sustainability management. In order to increase accuracy an‎d adapt to market fluctuations, solar energy deman‎d forecasting is also incorporated into the model structure using time series models. This forecasting allows for better planning an‎d reduces deman‎d uncertainties. To validate the proposed framework, a case study was conducted in Iran. The results showed that the reinforcement learning algorithm identifies crystalline silicon panels as the optimal option. The deman‎d forecasting model also showed an increasing trend in the need for solar energy. Finally, the multi-objective optimization model succeeded in providing balanced configurations that allow for the simultaneous management of economic costs, environmental impacts an‎d social benefits in supply chain design.

انرژي خورشيدي , يادگيري تقويتي , تحليل سري‌هاي زماني , بهينه‌سازي چندهدفه , زنجيره تأمين فتوولتائيك

Solar energy , Reinforcement learning , Time series analysis , Multi-objective optimization , Photovoltaic supply chain

Reihaneh Ahadzadeh

Ehsan Dehghani