About us
Goplan Goplan is a forest optimization and planning team hold in LEMAF (Forest Management Research Lab)/UFLA(Federal University of Lavras), which has a mission to explore operational research methods for helping the forest decision-making processes. We have a Professor leader and 6 students (undergraduate and postgraduate) to design numerical experiments, formulate problems, and develop computational codes. Our interests research area is wood supply chain; forest planning (operational, tactical and strategic); routing problem; artificial intelligence (heuristics and meta-heuristics); forest management (even and un-even aged); habitat’s species prediction and recognition pattern algorithms; resource and facilities location; prediction models (regression analysis and Random Forest algorithm). Let's check out our website to discover more about our researches findings!
OPERATIONAL RESEARCH
LINEAR PROGRAMMING
Forest planning
Wood suply chain
Land use soil
Routing problem
Forest management problems
Harvest scheduling problem
COMPUTATIONAL INTELLIGENCE
Heuristics
Meta-heuristics
Random forest
Researches board
WOOD SUPPLY OPTIMIZATION IN BRAZILIAN PULP INDUSTRY
Matheus Andrade Ferreira, Carolina Souza Jarochinski e Silva
Lucas Gomide
Forest outgrower scheme is a partnership between producers and forest companies, without land acquisition by the companies. Forest producers are recognized as an important source of growth for forest production. In order to achieve successful partnerships, the company needs to plan its hiring. In this study, the partnerships are made through formal contracts and the model does not allow breach of contract. Six scenarios were modeled and evaluated based on the total cost of the objective function, average wood cost (USD/m³), average transport distance (km) and processing time. The mathematical model showed its effi ciency, eff ectiveness, and fl exibility in generating scenarios to support company managers in the decision-making and to avoid breach of contract. Hiring planning benefi ts both sides as it ensures the company’s production and provides fi nancial security to forest producers...
search techniques to estimate Weibull function
Lais Araujo,
Lucas Gomide
The Weibull function, a continuous probability distribution, is widely used for diameter distribution modelling, in which parameter estimation performance is afected by stand attributes and ftting methods. The Weibull cumulative distribution function is nonlinear, and classical ftting methods may provide a not optimal solution. Invoking the use of artifcial intelligence by metaheuristics is reasonable for this optimisation task. Therefore, aimed and compared (1) the metaheuristics genetic algorithm and simulated annealing performance over the moment and percentile methods; (2) the hybrid strategy merging the metaheuristics tested and the percentile method and, (3) the metaheuristics ftness functions under basal area and density errors. A long-term experiment in a Pinus taeda stand subjected to crown thinning was used. According to our fndings, all methods have a similar performance, independent of the thinning regimes and age...
Amazon forest management - reduced logging impact
Paulo Henrique da Silva, Lucas Gomide
Natural uneven-aged forests represent a population in different age levels with diverse diameter structure and floristic composition resulting from the interaction between edaphic and climatic factors. In this work we developed an integer linear programming model to minimize the distance between harvested trees and log landings. The processes were established according to tree diameter structure balance, economy aspect, adjacency constraints and log production. Data are from a dense ombrophilous forest in the Western Amazon…
A HYBRID METHOD FOR FITTING NONLINEAR MODELS
Cássio Augusto Ussi Monti
Lucas Gomide
Regression analysis is widely applied in many fields of science and engineering to predict any hard to determine variable. In general, nonlinear regression models are more complex than linear regression. The former models request initial parameters to be fitted and the procedure to estimate these values is not simple. We propose a new hybrid method to fit nonlinear regression models using height diameter models. The genetic algorithm (GA) to prediction the initial parameters and Gauss-Newton, NL2SOL or Levenberg- Marquardt...
Lucas Gomide
Automated Selective thinning and artificial intelligence
Thinning is a usual silvicultural practice designed to supply the market with high quality and dimension of logs. This activity is described as an operational process applied in a stand when some trees are cut off. As result, the space grows between the remaining trees and it enhances individual growth. The choice of the remaining trees includes not only the threshold diameter size, but also a set of evaluations covering stem quality, spatial location of the tree, vitality and crown distribution. On the other hand, the decision making of the process should be direct and quick in the field. The question arises is how to automate the selection process of trees to be thinned in a stand and following a technical criterion for its standardization...
Habitat sutability modelling tree 1
Mônica CArvalho,
Lucas Gomide
Process of modeling potential specie’s distribution aims to predict and reconstruct non-randomly the relationship between environmental conditions and the occurrence of the species, based on known locations. This process associates specie’s occurrence data (geographical coordinates) to biotic and abiotic local variables (such as temperature, rainfall, elevation, vegetation indexes, individual characteristics, etc.). Thus, these models perform prediction suitable environments where, in theory, a population may remain viable. We modeled the potential distribution of four native three species in Minas Gerais using machine learning techniques...
Habitat sutability modelling tree 2
Mônica CArvalho,
Lucas Gomide
Eremanthus erythropappus is a commercially-important tree which has a long history of exploitation in the Brazilian State of Minas Gerais. The knowledge on the potential geographical distribution of E. erythropappus is therefore critical for the species sustainability. Thus, the aim of this study was to estimate and map current and future ecological niche for E. erythropappus in Minas Gerais. We used the Random Forests algorithm to model the ecological niche for current and future climates scenarios. Our predictions indicate Espinhaço, Mantiqueira, and Canastra mountain ranges as core areas of distribution and forecast drastic reductions in potential areas under all climate scenarios. Based on our results, we highlight that the continual harvesting of naturally-occurring E. erythropappus populations will not be sufficient to supply the market demand. Silviculture practices would likely serve as an economically viable and ecological sustainable alternative to harvesting natural populations.
Multiobjective forest transportation problem
Cássio Augusto Ussi Monti
Lucas Gomide
Forest transportation is one of the most expensive operations and often requests high level of skills and attention. The main problem is usually defined as logistic model with thousand of constraints. We adopted the optimization techniques and computing intelligence to solve the real forest problem. The transportation planning was designed to supply the industry with 21,881.82 tons of log. The resources applied were 10 Eucalyptus stands, 48 trucks and 5 cranes to be almost finished in 6 days...
Eucalyptus clonal mosaics in forest planning
THAYRINE MARTINS
Lucas Gomide
Currently a new concept of forest planning has become popular, spatial forest planning. In it, the stand is seen as part of a whole and it is considered its size, shape and distribution in relation to the others. For this problem, we used constraints of traditional planning, spatial constraints of the URM (Unit Restriction Model) type and restrictions of clonal adjacency, a new application of URM in forestry. This study aimed to create a cost minimization model using clonal material adjacent spatial restriction. This was done with the intention of reducing the size of homogenous areas in eucalypt forests and thus hinder the proliferation of pests and diseases, as well as contribute to the decision making of the forest manager...
Project schedulling problem
Nathalia Mendonça, Lucas Gomide
The scheduling of silvicultural activities sequencing is usually hard to solve, it usually involve resource constraints and multiple execution modes (SPSREM). The problem is classified as combinatorial optimization in programming projects area. The meta-heuristics are widely applied in problems of combinatorial optimization and its solutions are consistent in most of cases...
urban forest inventory routing problem
henrique haddad,
Lucas Gomide
Challenge is a key point of every scientist and routing problem is the hard one. We solve an arc routing problem to plan a forest urban inventory . The project was hold in Belo Horizonte city, capital of Minas Gerais state – Brazil. Integer linear programming model was formulated to solve and get the exact solution for every scenario. These solutions were compare with simulated annealing…
spatial stratification for forest harvest plannig
lucas Gomide
The aim was to evaluate the use of simulated annealing metaheuristic to solve P-median problem (capacitated and non-capacitated) and comparing their results with the branch and bound algorithm. To analyze the spatial stratum of the stand into compartments to be applied for the forest harvesting planning. The data was obtained from a forest project which containing technical information to build 2353 grids and 65 storage piles of wood....
drawing a real 3d forest with mathematical functions
matheus ferreira,
Lucas Gomide
Digital era is a reality for every professional and can be useful for our advance. Nowadays, Google is a great example where we are invited to explore and integrate our lives on maps and satellite images with high resolution. These actions promote a positive experience of being in some place virtually before travelling and help us to make better choices. Foresters used to quantify log volume according to aerial or terrestrian way. All information required is applied to predict or plan the forest resource and we employ statistical methods. Mathematical functions are most common to describe or estimate some hard and expensive variables to mensure. A great example is taper functions which is able to predict diameter in every correspondent height over the stem. Crown is another variable used to estimate the inter-tree competition and canopy cover. Based on several forest models and a range of GIS procedures by Model builder functions we have drawing a 3D forest of Eucaliptus sp in Brazil…