Hybrid heuristic strategies for planning and scheduling forest harvest and transportation activities in short term

Author(s)

Rafael Augusto Scaraficci

• Document type: Master's Dissertation
• Press: Campinas, SP.
• Institution: Universidade Estadual de Campinas (UNICAMP). Instituto de Computação
• Defense date: 2008-07-25

Examining board members

Arnaldo Vieira Moura; Nizam Omar; João Meidanis

• Advisor: Arnaldo Vieira Moura

Abstract

This thesis aimed at studying and solving a planning and scheduling problem stemming from forest harvest and wood transportation activities. Our approach treated a real problem faced by large pulp and paper companies in Brazil. It consists in planning, for a short-term horizon, the harvesting operations in different forest areas and the transportation of the logs to a processing unit, while satisfying a complex set of constraints, which includes constraints related to the structure of the harvest areas, some properties of the logs, the organization of the harvest teams and the degradation of dirt roads during rainy periods. In this research, we developed and evaluated some hybrid algorithmic strategies based on some principles of the GRASP (Greedy Randomized Adaptive Search Procedure), combined with advanced techniques such as long term memory, solution recombination methods and linear models. Computational tests proved that our strategies are robust and able to produce high quality solutions in a short amount of time. (AU)