Uma abordagem heurística construtiva para agendar e sequenciar atividades silviculturais em janelas de tempo restritas

Abstract

Brazil plays a key role in the expansion of the forestry sector in South America, recording an impressive growth of 52.9% over 14 years. However, maintaining this performance faces significant challenges. External factors, such as climate variations and rising input costs, combined with internal factors, such as the complexity of sequential operations conducted within pre-established timeframes, demand a high level of organization. These operations, carried out by teams that may or may not be exclusive to each activity, become even more critical when non-execution or delayed implementation results in significant volumetric losses during planting. The optimal sequencing of silvicultural operations, therefore, is a highly complex problem. It requires planning a large number of activities across various production units (plots). As the complexity of forestry problems increases, characterized by multiple variables and combinations, heuristic or approximative methods emerge as viable solutions for highly complex problems, reducing the computational effort needed to find optimal solutions. This study proposes a deterministic optimization model for operational micro-planning in plots, based on a constructive heuristic. The model spans an annual planning horizon with daily periodicity, aiming to schedule and sequence activities for the first year of operation. The goal is to respect the time windows of each operation, minimize travel between plots, and reduce team idleness. The first chapter presents the definitions and theoretical foundation of the problem. The second chapter details the SIOP simulator, capable of generalizing the problem, reproducing the initial status of the plots, and optimizing routes. Additionally, the CHESS-W model, developed for scheduling and sequencing 16 silvicultural activities in the first year post- harvest, demonstrates efficiency by operating with specialized teams, respecting mandatory sequences and time windows. Finally, the study compares the efficiency of the proposed model with real scenarios, highlighting its contributions to silvicultural planning.