CONICET | Buscador de Institutos y Recursos Humanos

Log transportation in forest industry is a resource-intensive operation and represents a great challenge for logistic planners. In the Argentinean Northeast (NEA) region, forest production represents the main economic activity, so enhancing efficiency in the transportation operation would imply a substantial improvement for this economy. The problem addressed in this work is how to distribute logs from the harvest areas to different plants, at a minimum cost, as long as the demand of plants is met. The main decisions that have to be taken at the operational level are related to the allocation of the logs to the demanding plants and the routing of the truck fleet (in the Argentine context, the availability of logs in the harvest areas allows the trucks to be fully loaded). Since the loadings resources (at plants and harvest areas) are limited, the sequencing of trucks arrivals in each node of the network becomes critical to avoid unproductive waiting times.This kind of problems is classified as NP-hard and different approaches are used to deal with it, most of them are heuristic based (El Hachemi et al., 2013; Haridass et al., 2014; Rey et al., 2009; among others). Since routing and scheduling problems are difficult to solve jointly, three solution approaches based on a previous work developed by Bordón et al. (2018) are presented.The first one involves two stages that are iteratively solved: product allocation, trip composition and truck routing decisions are jointly considered in a Mixed-Integer Linear Programming model (MILP), while in the second stage, fixing the route for each truck (according to the results of the previous step), a MILP model for the scheduling of truck arrivals at plants is considered. If no feasible solution for the scheduling problem is obtained, then an integer cut is applied in order to exclude from the search space truck routes already explored in previous iterations.The second methodology is similar to the first one, where actualization of some bounds in each iteration is added, i.e., the total traveled distances (in the routing model) must be greater than in the previous iteration and the maximum exceeded routing time (in the scheduling model) must be less than in the previous phase. This approach defines a heuristic solution, unlike the first methodology in which the problem is solved exactly.Finally, the last methodology consists of increasing the number of trucks used (in the routing model) with respect to the previous iteration, and it also defines a heuristic solution.The resolution approaches are tested in a case study representative of the Argentine context and conclusions are detailed.