Multi-Objective Metaheuristic Solution Approach for the Crop Plant Scheduling Problem

Abstract

Accurate and optimized harvest schedules should improve the subsequent harvest performance. In this work, we formulate a bi-objective optimization model used to improve harvest effectiveness, measured by the amount of unnecessarily produced waste, and its efficiency, measured by the time required for harvest completion. The proposed model considers a farm field divided into management units that are treated and harvested independently and assume storage and processing limits that dictate the pace of the optimal harvest. The optimization problem is solved using 2 distinctive metaheuristic techniques. We examined adaptive large neighborhood search (ALNS) as a model-driven single-solution-based metaheuristic and non-dominated sorting genetic algorithm II (NSGA-II) as a data-driven population-based approach. Optimization techniques were tested on 20 synthesized test cases. All obtained results emphasize the significant and consistent dominance of the ALNS over the NSGA-II.