Integrated forward/reverse logistics thin-film photovoltaic power plant supply chain network design with uncertain data

Unique features of solar energy, such as the low cost of operation and low pollution, make photovoltaic power plants attractive. Despite the advantages of these power plants, such as low cost of operation and operating without pollution, their high establishment cost compared to that of the conventional fossil fuel-fired power plants prevents their rapid growth. Therefore, designing and optimizing an efficient supply chain network will help to develop the photovoltaic industry. This study proposes a bi-objective model to design and optimize a thin-film photovoltaic power plant supply chain network integrating reverse logistics. The first objective minimizes the costs of supply chain network design, which optimizes decisions from suppliers to disposal centers. The second objective maximizes the efficiency of the data envelopment analysis model to find suitable locations of facilities. As there is uncertainty in electricity demand prediction and perfect foresight is not possible, a fuzzy-robust programming approach is devised to cope with the uncertain parameters. The validity and practicality of the proposed approach are investigated by a case study based on real-world data. Finally, K-Means clustering is incorporated into the model to analyze the installed photovoltaic capacity of different cities in each period. The results show that using an optimistic approach to deal with uncertainty reduces the total cost by 20% in comparison with using a pessimistic approach. The results also provide helpful insights for decision-makers and policymakers in allocating and optimizing the operation of facilities by considering the real-world characteristics of the candidate locations.