• Journal of the Korea Society of Computer and Information
• /
• v.18 no.12
• /
• pp.197-205
• /
• 2013

In this study, we developed a program of optimizing stand-level carbon stock using a stand-level yield model and the Simulated Annealing (SA) heuristic method to derive a optimized forest treatment solution. The SA is one of the heuristic algorithms that can provide a desirable management solution when dealing with various management purposes. The SA heuristic algorithm applied 'thermal equilibrium test', a thresholds approach to solve the phenomenon that does not find an optimum solution and stays at a local optimum value during the process. We conducted a sensitivity test for the temperature reduction rate, the major parameter of the thermal equilibrium test, to analyze its influence on the objective function value and the total iteration of the optimization process. Using the developed program, three scenarios were compared: a common treatment in forestry (baseline), the optimized solution of maximizing the amount of harvest(alternative 1), and the optimized solution of maximizing the amount of carbon stocks(alternative 2). As the results, we found that the alternative 1 showed provide acceptable solutions for the objectives. From the sensitivity test, we found that the objective function value and the total iteration of the process can be significantly influenced by the temperature reduction rate. The developed program will be practically used for optimizing stand-level carbon stock and developing optimized treatment solutions.

• Journal of Bio-Environment Control
• /
• v.31 no.1
• /
• pp.1-7
• /
• 2022

Strawberry is a stand-out cultivating fruit in Korea. The optimum production of strawberry is highly dependent on growing environment. Smart farm technology, and automatic monitoring and control system maintain a favorable environment for strawberry growth in greenhouses, as well as play an important role to improve production. Moreover, physiological parameters of strawberry plant and it is surrounding environment may allow to give an idea on production of strawberry. Therefore, this study intends to build a machine learning model to predict strawberry's yield, cultivated in greenhouse. The environmental parameter like as temperature, humidity and CO₂ and physiological parameters such as length of leaves, number of flowers and fruits and chlorophyll content of 'Seolhyang' (widely growing strawberry cultivar in Korea) were collected from three strawberry greenhouses located in Sacheon of Gyeongsangnam-do during the period of 2019-2020. A predictive model, Lasso regression was designed and validated through 5-fold cross-validation. The current study found that performance of the Lasso regression model is good to predict the number of flowers and fruits, when the MAPE value are 0.511 and 0.488, respectively during the model validation. Overall, the present study demonstrates that using AI based regression model may be convenient for farms and agricultural companies to predict yield of crops with fewer input attributes.

• Journal of Korean Society of Forest Science
• /
• v.98 no.3
• /
• pp.247-254
• /
• 2009

In this study, linear programming was applied to a case study in Gwangreung Experimental Forest of Korea Forest Research Institute investigating the effect of the desired future conditions on forest management. Considering the social, economic and ecological demands of people from the forest, the forest functions were classified into four including natural conservation, timber production, water yield and scenic conservation. The forest land areas were divided into four-types of forest functional zones and forest management prescriptions including the desired future conditions by the forest function type were established. The Model II linear programming was used in optimizing the forest management planning. The model includes management policies, as the constraints, for non-declining yield, allowable cutting area, allowable % age class distribution and allowable % species allocation as well as the land and other accounting regimes. Maximization of timber production was used the objective function. Based on the Model II formulations, the effects of the desired future conditions by the forest function type on forest management planning were investigated in terms of timber production, net present value and stand structures over time.

• Research in Plant Disease
• /
• v.7 no.1
• /
• pp.20-24
• /
• 2001

As stand-missing rate (SMR) of ginseng plants in fields are directly related to the ginseng root rot, the forecasting model for the root rot progress in ginseng fields was developed, using the estimated SMRs by disease incidence (DI) of ginseng seedling in the soil-indexing bioassay and the estimate of DI derived from soil environmental factors or rhizoplane microflora. For fitness analysis of the forecasting model, simple correlation and linear regression between SMRs at different planting ages in fields and their estimates by 3 factors of the model were evaluated.The SMR estimated from the factor of DI in the bioassay had much higher fitness to the SMR observed in fields than that from the factors of soil environments and rhizoplane microflora. The estimated SMRs in young and aged ginseng fields by DI in the bioassay were significantly correlated with the observed SMRs in 3- and 5-year-old ginseng fields, respectively (p=0.01). this implicates that indexing preplanting field soils with the forecasting model using DI in the bioassay can provide an information to determine the suitability of the fields for ginseng cultivation, and that indexing cultivating field soils can be helpful to determine the time of harvesting to reduce further yield loss by root rot in continuous cultivation in the next year.

• Journal of Korean Society of Forest Science
• /
• v.96 no.1
• /
• pp.40-47
• /
• 2007

The objectives of this study were to derive site index and stem volume prediction equation based on stem analysis data for Larix leptolepis in Jinan region. The function for site index was developed by algebraic difference equation method. Polymorphic site index family curves with base age of 40 were presented based on the Schumacher height equation. The best stem volume prediction equation was suggested as $V=0.00260+0.00000399D^2H$. The simultaneous F-test using this equation showed that the estimated tree stem volumes were not significantly different (${\alpha}=0.05$ level) from the observed stem volumes for model evaluation. Therefore, site index and volume prediction equations prepared in this study could provide an indication of site quality and basic information for making of yield table, and could be used for rational forest management of Larix leptolepis stands grown in Jinan region.

• Journal of agriculture & life science
• /
• v.52 no.6
• /
• pp.37-47
• /
• 2018

The objective of this study was to analyze the influences of forest management activity on the diameter distribution of Larix kaempferi Carriere stands in South Korea. We used 232 managed stands data, 47 unmanaged stands data of National Forest Inventory for this study. We employed the Weibull distribution function for estimating diameter based on percentiles and parameter recovery method. The results revealed that the average diameter breast height movements and growth of tree in the managed stands higher than the unmanaged stands according to the scenario: age, site index, and tree density change. The finding shows the percentage of the total amount of large class diameter was also high in the managed stands. The results of this study could be apply for the estimation of multi-products of timbers per diameter classes and stand structure development for Larix kaempferi Carriere stands in South Korea.

• Korean Journal of Agricultural Science
• /
• v.4 no.2
• /
• pp.344-390
• /
• 1977

The purpose of this paper was to schedule optimum cutting strategy which could maximize the total yield under certain restrictions on periodic timber removals and harvest areas from an industrial forest, based on a linear programming technique. Sensitivity of the regulation model to variations in restrictions has also been analyzed to get information on the changes of total yield in the planning period. The regulation procedure has been made on the experimental forest of the Agricultural College of Seoul National University. The forest is composed of 219 cutting units, and characterized by younger age group which is very common in Korea. The planning period is devided into 10 cutting periods of five years each, and cutting is permissible only on the stands of age groups 5-9. It is also assumed in the study that the subsequent forests are established immediately after cutting existing forests, non-stocked forest lands are planted in first cutting period, and established forests are fully stocked until next harvest. All feasible cutting regimes have been defined to each unit depending on their age groups. Total yield (Vi, k) of each regime expected in the planning period has been projected using stand yield tables and forest inventory data, and the regime which gives highest Vi, k has been selected as a optimum cutting regime. After calculating periodic yields and cutting areas, and total yield from the optimum regimes selected without any restrictions, the upper and lower limits of periodic yields(Vj-max, Vj-min) and those of periodic cutting areas (Aj-max, Aj-min) have been decided. The optimum regimes under such restrictions have been selected by linear programming. The results of the study may be summarized as follows:- 1. The fluctuations of periodic harvest yields and areas under cutting regimes selected without restrictions were very great, because of irregular composition of age classes and growing stocks of existing stands. About 68.8 percent of total yield is expected in period 10, while none of yield in periods 6 and 7. 2. After inspection of the above solution, restricted optimum cutting regimes were obtained under the restrictions of Amin=150 ha, Amax=400ha, $Vmin=5,000m^3$ and $Vmax=50,000m^3$, using LP regulation model. As a result, about $50,000m^3$ of stable harvest yield per period and a relatively balanced age group distribution is expected from period 5. In this case, the loss in total yield was about 29 percent of that of unrestricted regimes. 3. Thinning schedule could be easily treated by the model presented in the study, and the thinnings made it possible to select optimum regimes which might be effective for smoothing the wood flows, not to speak of increasing total yield in the planning period. 4. It was known that the stronger the restrictions becomes in the optimum solution the earlier the period comes in which balanced harvest yields and age group distribution can be formed. There was also a tendency in this particular case that the periodic yields were strongly affected by constraints, and the fluctuations of harvest areas depended upon the amount of periodic yields. 5. Because the total yield was decreased at the increasing rate with imposing stronger restrictions, the Joss would be very great where strict sustained yield and normal age group distribution are required in the earlier periods. 6. Total yield under the same restrictions in a period was increased by lowering the felling age and extending the range of cutting age groups. Therefore, it seemed to be advantageous for producing maximum timber yield to adopt wider range of cutting age groups with the lower limit at which the smallest utilization size of timber could be produced. 7. The LP regulation model presented in the study seemed to be useful in the Korean situation from the following point of view: (1) The model can provide forest managers with the solution of where, when, and how much to cut in order to best fulfill the owners objective. (2) Planning is visualized as a continuous process where new strateges are automatically evolved as changes in the forest environment are recognized. (3) The cost (measured as decrease in total yield) of imposing restrictions can be easily evaluated. (4) Thinning schedule can be treated without difficulty. (5) The model can be applied to irregular forests. (6) Traditional regulation methods can be rainforced by the model.