Efficient multi-objective metaheuristic algorithm for sustainable harvest planning problem

Abstract

The shift towards sustainable and regenerative agriculture is being propelled by global farmers due to increasing awareness of social inequalities and climate change. To make this transition a reality, farmers should consider various sustainability factors including all economic, environmental, and social factors, and tackle the complexity of the harvest planning. This study introduces a new multi-objective optimization approach that employs fuzzy logic and multiple objectives to facilitate sustainable harvest planning in the face of various sources of uncertainties such as changes in commodity prices, weather conditions, crop ripening patterns, and productivity fluctuations. The model seeks to optimize profit while minimizing greenhouse gas emissions and wastes generated by harvesting machines as the economic and environmental dimensions. To incorporate social sustainability, we define the farmer's working days on each block as a constraint set in our model. To address the complexity of this optimization model in large-scale networks, this paper proposes a revised version of the non-dominated sorting genetic algorithm (NSGA-II) using the genetic engineering concept, called the non-dominated sorting genetic engineering algorithm (NSGEA). This article showcases the outcomes of a case study that employed the blueberry industry in Canada. The findings indicate that the NSGEA algorithm, which was proposed in the study, is effective in addressing our multi-objective optimization model in comparison to other metaheuristic algorithms and the epsilon constraint method. This paper concludes by discussing theoretical contributions and managerial insights that emphasize the advantages of the proposed multi-objective harvest planning problem for achieving sustainable blueberry agriculture in Canada.