|
|
Management of production populations of forest tree species utilizing methods of mathematical programing
Hořejší, Matouš ; Lstibůrek, Milan (advisor) ; Stejskal, Jan (referee)
Breeding of forest tree species is a great opportunity for the entire forestry sector to succeed in the present modern World. Genetically improved forest trees have much better growing quality and they can sustain higher stress and unfavorable climatic conditions.
Efficient algorithms are necessary for optimizing and establishing seed orchards. These should optimally select clones as parents in future seed orchards. In the present thesis, we developed model, based on mathematical programming methods, which creates mix of clones with the maximum breeding value and constraints on genetic diversity.
|
|
|
The spatial final cut optimisation
Kašpar, Jan ; Marušák, Róbert (advisor) ; Lstibůrek, Milan (referee)
For more than 150 years, timber indicators have been used for forest harvest scheduling in the region of current Czech Republic. With social changes and increased importance of forest yields and functions, there has been significant changes in forest management during this time. The changes in ownership structure after 1989 are most problematic from a sustainability standpoint. Also, the timber indicators cannot consider forest age structure, which is not ideal because even small changes related to abiotic and biotic disturbances may inhibit forest planning efforts. These changes to forest ownership and structure combined with current timber indicators make it difficult to ensure sustainable harvest levels over long time horizons. Methods of operations research, such as mathematical programming and heuristic methods, provide alternatives to historical harvest scheduling methods because they can account for forest age structure and spatial relationships over long scheduling horizons.
|
|
|
Proposal of a new seed orchard layout assuming panmictic model
Chaloupková, Kateřina ; Lstibůrek, Milan (advisor) ; Vladimír, Vladimír (referee)
The new panmictic model of seed orchard scheme for forests trees is described in this thesis. It can be used either alone or in combination with existing designs without losing efficiency. In this case it iscombined with Minimum Inbreeding design. The establishment of seed orchards and importance of achieving panmixy and minimizing inbreeding in the population of seed orchard are described in the theoretical part. The resultant algorithm defines panmictic situation as a situation where close neighborhoods of all possible combinations of clones occur within the seed orchard grid with the same frequency. To achieve this we developed a formula for variance that is minimized as a criterion function. The algorithm has been programmed in R and it was tested for different tasks including balanced or unbalanced numbers of clones and different sizes or shapes (rectangular or non-rectangular shape) of orchard. The model is able to solve all of these situations, and it is very effective especially in the case of balanced numbers of clones. Optimal solution of the given task was often found at the first iteration. To evaluate the relative quality of resulting schemes theoretic proxy to level of maximum panmixy was defined for every specific task- the Theoretical minimum variance. The effectiveness of MI scheme combined with clonal rows was also evaluated. The result is a hybrid scheme with predefined MI clonal rows surrounded by panmictic algorithm, while close neighborhood of clonal rows was also optimized. This scheme will be used for the establishment of Abies fraseriseed orchard in North Carolina under NCSU Christmas Tree Genetics Program. Another possible modifications of the algorithm, such as the inclusion of local separation zones or assortative mating are summarized in conclusions.
|
guest ::
