A multi-objective mathematical optimization approach for the forest planning problem

FRANK CECILIO, PIEDRA JIMENEZ; TASSIN, NATALIA G.; BROZ, DIEGO; NOVAS, J. MATIAS; RODRIGUEZ, MARIA ANALIA

Conferencia; XXI LATIN IBERO-AMERICAN CONFERENCE ON OPERATIONS RESEARCH; 2022

A general mathematical framework based on a Generalized Disjunctive Pro-gramming (GDP) approach for optimal forest planning problems is proposed in this work. For this purpose, a Multi-objective GDP (MO-GDP) model is con-structed and reformulated as a Multi-objective Mixed Integer Linear Program-ming (MO-MILP) model. The model determines the optimal forest management alternative (combination of silvicultural treatments), the proportion of land area to be harvested, and the flow of timber products from harvesting nodes to forest in-dustries. The proposed mathematical formulation simultaneously addresses two conflicting objectives: the maximization of the net present value and the minimiza-tion of the absolute deviations of timber assortment production between consecu-tive periods. SisPinus® growth simulator is used to estimate timber yields, and the MO-MILP developed model is solved in GAMS. In addition, two alternative iterative procedures, the so-called, ϵ-constraint and the AUGMECON methods are used to obtain the Pareto optimal solutions. The feasibility of the proposed model is tested using real data from a company located in the north of the prov-ince of Misiones. Computational results show that the designed framework serves as a decision-making tool to provide diverse solutions with different trade-offs among the considered criteria.