TACTICAL CROP PLANNING: AN APPROACH WITH SINKHORN ALGORITHM AND MULTIOBJECTIVE OPTIMIZATION

Agriculture plays a central role in the global economy and in combating food insecurity. However, agricultural planning faces significant logistical uncertainties, particularly regarding the time interval between planting and harvesting, whose variability affects production efficiency. This challenge is especially relevant in the Andean slopes of Peru, where temporary crops are predominant. In this context, this study proposes a decision-support tool that integrates multiobjective optimization and optimal transport theory to support agricultural management. Instead of assuming fixed crop cycles, the Sinkhorn algorithm is applied to historical data to estimate the actual lag between recorded cultivation cycles. The resulting information is used to adjust temporal data and construct a productivity function embedded into the optimization model. The proposed approach generates planting and harvesting plans aimed at maximizing food production while minimizing the number of simultaneous harvesting fronts.