Tuning of the Metaheuristic Variable Neighborhood Search for a Forest Planning Problem

Júnior, Carlos Alberto Araújo; Mendes, João Batista; Assis, Adriana Leandra de; Cabacinha, Christian Dias; Stocks, Jonathan; Silva, Liniker Fernandes da; Leite, Hélio Garcia

doi:10.1590/01047760201824032538

Cited by 1 publication

(3 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The proposed version of the VNS considers four structures to modify the neighborhood of the current solution. This implies that the prescriptions of a certain percentage of stands are randomly modified (Araújo et al, 2018) for each neighborhood structure, i.e., structures 1, 2, 3, and 4 allow changes in 1%, 2%, 3%, and 4% of stands, respectively. In this study, the structures were considered in an orderly manner during each iteration.…”

Section: Variable Neighborhood Search (Vns)mentioning

confidence: 99%

“…The models proposed by Johnson and Scheurman (1977), i.e., Model Types I and II, serve as basic models for forest planning and references for several relevant studies (Araújo Júnior et al (2018), Silva et al (2003), and ). Areas of the management unit subject to a certain silvicultural prescription can be fractionated in these models.…”

Section: Introductionmentioning

confidence: 99%

“…In several cases, the best management alternative must be defined for each stand without permission to fractionate its area, which results in binary responses (harvesting or not harvesting a stand in a specified year). This requires the use of whole mixed programming (Silva et al, 2003) or metaheuristics (Araújo Júnior et al (2018), Matos et al (2019), andFerreira et al (2018)).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Can Linear Programming Assist Metaheuristics in Forest Production Planning Problem?

Júnior

Castro

Mendes

et al. 2021

Self Cite

View full text Add to dashboard Cite

The planning of forest production requires the adoption of mathematical models to optimize the utilization of available resources. Hence, studies involving the improvement of decision-making processes must be performed. Herein, we evaluate an alternative method for improving the performance of metaheuristics when they are applied for identifying solutions to problems in forest production planning. The inclusion of a solution obtained by rounding the optimal solution of linear programming to a relaxed problem is investigated. Such a solution is included in the initial population of the clonal selection algorithm, genetic algorithm, simulated annealing, and variable neighborhood search metaheuristics when it is used to generate harvest and planting plans in an area measuring 4,210 ha comprising 120 management units with ages varying between 1 and 6 years. The same algorithms are executed without including the solutions mentioned in the initial population. Results show that the performance of the clonal selection algorithm, genetic algorithm, and variable neighborhood search algorithms improved significantly. Positive effects on the performance of the simulated annealing metaheuristic are not indicated. Hence, it is concluded that rounding off the solution to a relaxed problem is a good alternative for generating an initial solution for metaheuristics.

show abstract

Section: Variable Neighborhood Search (Vns)mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%