Aplicação do método Simulated Annealing em um problema de sequenciamento das atividades silviculturais

Abstract

Brazilian high forest productivity comes from numerous efforts related to research and development of silviculture. However, planning is usually not thinking before making decisions. This process is a huge and hard challenge for forest managers. The high complexity in a silviculture planning is associated with a set of activities to be performed by use the work-force over a planning horizon. In this type of problem, the exact programming does not produce an optimal solution in a viable run processing time. The heuristic algorithms are an alternative way to find excellent solutions against exact methods. In this sense, the aims of the study were to develop a mathematical programming models involving integer linear programming (ILP) and the simulated annealing metaheuristic (SA), to solve the sequence problem of silviculture activities with resource constraints and multiple modes of execution (SPSRME). The first chapter presented formal definitions and a bibliographical study about the problem at issue. In the following, chapter two, evaluated the efficiency of the SA metaheuristic compared with solutions obtained by integer linear programming formulation of the SPSRME. Different goals of planning were addressed by minimization of cost, makespan, and the worforce use oscilation in eight different scenarios of a silviculture project. The study results showed that the metaheuristic SA was efficient to solve the proposed problem, with low percentage of deviations from the ILP solutions. The presented models should be used to support the operational decision making of silviculture activities. It’s not possible to establish a single ideal scenario, since they presented different planning focus. However, a model setting for each lodged planning could be established.