Vehicle Sequencing Strategies in Optimization of Forestry Logistics

Abstract

Timber transportation is a hard task for any forest company. There are several efforts to reduce and control these costs, considering efficient equipment and optimization frameworks. Due to the pattern of combinatorial problems, the deterministic methods have a high computational effort of processing time, which generally makes their use unfeasible. An alternative procedure is applying approximation algorithms, which have efficient searches for finding feasible solutions. The present study evaluated algorithms for solving the forest transport sequencing problem. The Simulated instances were designed to highlight three instances for 10, 20, and 30 stands with a realistic pattern. According to the operational set, four solving strategies were proposed, considering three algorithms (Simulated Annealing, Greedy, and Greedy-Simulated Annealing). The computational code, processing, and analysis of the data were performed using RStudio software. The results show feasible solutions from all tested algorithms, highlighting the hybrid Greedy-Simulated Annealing algorithm. With proven normality and homogeneity of variance, the algorithms were tested with the Tukey test at a level of 5%. The blocks, stand random selection, and multiple flow strategies produced the best results of the tested instances. As a conclusion, the proposed algorithms are efficient for solving the forest transportation problem and may be helpful at the operational planning level.