An Integer Linear Programming Model for the Daily Transportation Planning in the Sugarcane Industry

MELCHIORI, LUCIANA; NASINI, GRACIELA; MONTAGNA, JORGE MARCELO; CORSANO, GABRIELA

Buenos Aires

Conferencia; Latin-Iberoamerican Conference on Operations Research (CLAIO 2022); 2022

CLAIO

This work aims to develop a tool to solve the daily sugarcane transport planning (DSTP) in the context of the Argentine industry. The tasks involved in the sugarcane harvest scheduling are cane cutting, in-fields loading, transportation to the mill (sugarcane processing plant) and unloading at the mill. These activities represent a considerable percentage of the total sugar supply chain cost (Ministerio de Hacienda de la Naci´on Argentina (2018)), and have significant influence on supply chain efficiency due to the cane deterioration. Thus, it is important to get a daily scheduling program to avoid this drawback.In Argentina, the sugar harvest time is about 150-180 days at year and through strategic andtactical planning analysis, the set of sugarcane providers for each day is prior determined. Therefore, from operational point of view, the DSTP can be stated as: given a set of suppliers, each one offering a maximum number of sugarcane full-truckloads and with time windows for the pickup activity, the sugar mill demand expressed in number of full-truckloads, and a truck fleet hosted at the sugar mill, the goal is to determine the truck routing and the scheduling arrivals at harvest areas and the mill in order to satisfy the mill demand at minimum cost.The problem involved in the DSTP can be modeled as a particular case of the Unpaired Full truck-load Pickup and Delivery Problem with Resource Synchronization (U-FTL-PDP-RS) studied in Melchiori et al. (2022). In the present work, the exact approach formulated in this previous articlethrough an Integer Linear Programm model is applied to for the particular case of the DSTP. Theresolution of the proposed model integrates valid and symmetry-breaking constraints, and primal and dual bounds based on the solution of a combinatorial relaxation of the problem, ignoring scheduling restrictions (see Melchiori et al. (2020)). The presented approach is evaluated in several scenarios associated with the Argentine sugar industry, demonstrating in all of them good performance to solve the hard and complex DSTP.