Optimizing citrus production accounting for decisions’ timeframes and stochastics

Abstract

In many countries the citrus industry is an important economic factor. Unlike many supply chains, which are “demand-driven” systems, the orange industry’s supply chain is typically a “production-driven” system. Oranges are either sold directly to the market or to juice producers, which process oranges into juice. The resulting residue is either disposed of as waste or used to produce by-products that reduce waste and contribute to overall contribution margins.

In view of the changing climate and the resulting weather conditions a planting plan considering different types of orange trees is useful. For each orange variety, the planning process also includes decisions on the number of oranges that will be sold directly to final consumers, transported to juice producers, or stored and the by-products produced. The paper shows the potential of a stochastic, integrated strategic planning model that considers the temporal scope of different decisions. The proposed scenario-based linear programming model was developed to determine the amount of planted tress for each tree type and the production quantities of the different end products of the system (oranges, juice, and by-products) given the resources (fresh oranges, storage capacity) required to maximize the contribution margin of the company/system. The results of the scenario-based model are compared to the results of a worst-case approach, an on-average-approach, and an intuitive approach. In comparison, the scenario-based model always delivers the best results.