Resumen de Optimization methods in industrial forest plantations management: linking strategic and tactical planning scales
Pedro Henrique Bellavenutte Martins da Silva
The management planning performs a key role to the sustainability of the industrial forest plantations companies in the global market. Many kinds of optimization procedures were developed to support forest manager’s decisions. However, there still existing challenging problems, such as the harvest scheduling generation, which includes criteria from different natures and time-scales, besides which, contain spatial compounds leading to the formulation of very complex combinatorial optimization problems. Specifically, to solve the harvest scheduling problem, forest managers stablished a hierarchical decision-process consisted of strategic and tactical levels. The former is associated with extended temporal horizons, whereas the latter focus on shorter periods of time and typically encompass further spatial requirements, including the road network. The problem arises not only from the integration between the results obtained from both levels but also from the implementation of new tactical optimization approaches to improve the efficiency of these harvesting operations. Fortunately, thanks to continuous computational advances, it is suitable to develop new optimization procedures to support forest managers. This dissertation focused on these two problems previously defined: the integration of strategic and tactical results and the implementation of new tactical optimization approaches.
Initially, it was carried out a critical survey of optimization methods in industrial forest plantations management. Despite the high number of publications applying optimization methods to forest management, it has got a few associated with the management of industrial forest plantations, specifically. Thus, considering strictly the applications of optimization methods techniques in these forest systems it has been analyzed 67 articles published in journals contained in the ISI Web of Science. It has been highlighted, among other aspects, the optimization techniques employed, their evolution, their hierarchical planning levels, the objective functions and constraints considered, and the type of software deployed. Results verified the frequent usage of the strategic Model I, spatial constraints and mixed integer programming, besides the strongly employment of multi-criteria techniques in recent years.
The first addressed problem was the linkage between strategic and tactical levels of the harvest scheduling problem using multi-criteria techniques. The existence of multiple criteria of different nature is the rule rather than the exception in most of forestry scenarios. Therefore, it seems sensible to think that integrating strategic and tactical optimization models within a multi criteria context allows to approach the problem using models significantly more practical in many cases. Initially, it has been exposed the strategic and tactical theoretical models and after that, it was presented an integrated framework for dealing with this type of situation. In particular, two approaches called “Top-down” and “Integrated” has been formulated. To undertake this task, the multicriteria tool called “extended goal programming” has been resorted to. The functioning and the strengths of the integrated framework are illustrated with a case study of industrial forest plantation in Brazil. Results show that the framework allows the multiple criteria optimization and that it can achieve significant improvements including decision maker’s preferences.
The second addressed problem was the integration of tactical harvest scheduling and road network planning using the Fixed Charge Transportation Model. The problem consists of setting the stands to be harvested with specific volume production targets integrating spatial and temporal forest operations decisions, such as timber felling and extraction, road network construction/maintenance, and log transportation. Firstly, it has been presented the mathematical formulation of the optimization model followed by the model partitioning and solving procedure. Later, the functionality and performance of this optimization approach in solving the tactical harvest scheduling problem was illustrated using a case study corresponding to a large eucalyptus plantation company located in Brazil. Results show the optimization procedure allows the resolution of large-scale very complex tactical problems.
