Aplicação de meta-heurísticas no escalonamento de motoristas para o transporte de madeira

Abstract

The cargo transportation in Brazilian forest area is carried out mostly by road.Their complexity shows the importance of diagnosing the development of the activity so that we can, means operational research, more efficiently employ their resources without losing productivity. Currently, the driver scheduling problem is considered a major constraint for optimizing of transportation companies, because it has a lot of physical and technical restriction. This research aimed to solve a real problem of driver scheduling for the wood transportation from a forest company by using five metaheuristics (Genetic Algorithm – AG, Memetic Algorithm – AM, Clustering Search – CS, Greedy Randomized Adaptive Search – GRASP, and Simulated Annealing – SA), compare the methods among themselves and between the company's actual situation and propose a new scale of services. To validate the proposed approach, we used a data input generated from the qualitative research and studies time and movement. The results indicate that all methods were effective to solve the problem standing out as the best CS, followed by the SA, GRASP, AM and AG. The CS was able to meet the goals with a reduction of one third of the frame of 150 drivers and eliminates daily averages of overtime and surpluses which were 01h03min and 00h51min, respectively