A multi-objective optimization of by-products and post-harvest wastes in agricultural circular supply chain considering perishability

• Main Author: Chan, Aaron C.
• Format: text
• Language: English
• Published: Animo Repository 2024
• Subjects: Agricultural wastes; Business logistics; Circular economy; Industrial Engineering
• Online Access: https://animorepository.dlsu.edu.ph/etdm_induseng/9; https://animorepository.dlsu.edu.ph/context/etdm_induseng/article/1013/viewcontent/2023_Chan_A_multi_objective_optimization_of_by_products_and_post_harvest_Full_text.pdf
• Institution: De La Salle University

While many studies employ mathematical models to support decision-making, recent attention has shifted towards sustainable management of perishable supply chains like agriculture as these supply chains generate significant waste. However, there are still areas where these models can be improved. Current models in sustainable agricultural supply chains have adapted the closed-loop concept from traditional modular supply chains. They use waste to create secondary composts but often treat post-harvest waste and by-products as identical. Distinguishing between these items can create opportunities for higher profits. Another limitation is the oversimplified representation of perishability in these models, which hampers their accuracy and timeliness in decision making. To address these gaps, this study proposes a multi-objective mixed-integer nonlinear programming model for an agricultural circular supply chain. The model optimizes production, inventory, distribution, and recovery considering profit maximization, waste reduction, and average freshness as objectives. The model was validated with CPLEX solver in GAMS with an alternative approach to address the nonlinearities between continuous variables. Specifically, the product of a binary variable and a vector of possible values was used to estimate one of the continuous variables, transforming the product between two continuous variables to the product between a continuous and a binary variable. Additionally, a normalized goal program with systematic target setting was employed to handle multiple objectives. To consider uncertain perishability, a simulation-optimization approach was also implemented. The sensitivity analysis and scenario analysis performed revealed key insights: (1) Setting the minimum planting requirement too high can negatively impact the supply chain. A more conservative approach is recommended. (2) Harvest period length significantly affects production. Shorter periods can lead to higher costs and freshness compromise. (3) Maximizing raw material inputs can enhance production when weather significantly impacts yields. (4) Demand shapes supply chain decisions. Lower demand leads to less waste, while increasing demand under a larger minimum planting requirement results in excess inventory and waste. (5) Establishing a reverse system, even if unprofitable, contributes significantly to environmental and social objectives. (6) Incorporating shelf life uncertainty results in more resilient decisions but sacrifices average freshness to meet demand and reduce waste. For future studies, it is advised to expand the scope by treating some assumed constants as variables. Alternatively, some relaxed assumptions can be revisited to explore a more holistic system.