|
|
Multi-layered neural networks and visualization of their structure
Drobný, Michal ; Mrázová, Iveta (advisor) ; Kukačka, Marek (referee)
The model of multi-layered neural networks of the back-propagation type is well-known for their universal approximation capability and even the standard back-propagation training algorithm used for their adjustment often provides results applicable to real-world problems. The present study deals with the issue of the multi-layered neural networks. It describes selected variants of training algorithms, mainly the standard back-propagation training algorithm and the scaled conjugate gradients algorithm, which ranks among the extremely fast second-order algorithms. One of the parts of the present study is also an application for the visualisation of the structure of multi-layered neural networks whose solution is designed with respect to its potential utilization in the education of artificial intelligence. The first part of the study introduces the subject matter and formally describes both algorithms, followed by a short description of other variants of the algorithms and their analysis. The next part discusses the selection of the appropriate programming language for the implementation of the application, specifies the goals and describes the implementation works. The conclusion summarizes the test results of the speed and implementation comparison with the selected noncommercial-based software ENCOG.
|
|
|
Methods for solving selected vehicle routing problems and their implementation.
Drobný, Michal ; Grygarová, Libuše (advisor) ; Zimmermann, Karel (referee)
Various types of transportation issues are a common practice. The issue may be approached mainly as the distribution of products from suppliers to consumers while minimising distribution costs. The difference of real transportation issues predominantly relates to the considered restrictions, such as capacities of vehicles and orders, time windows and other special distribution restrictions. Transportation issues were already defined by F.L. Hitchcock in 1941 and since then, a wide range of stochastic and non- determinist methods providing solutions to transportation issues have been developed. Nevertheless, introducing distribution restrictions in resolving real-life problems makes it difficult for such methods to be applied. This thesis provides a compilation of the well-known determinist methods that may be used to resolve transportation issues. The methods that are appropriate for resolving real issues are discussed in more detail. The solution procedure of the selected method is demonstrated using simple examples and the results are compared with the results of other methods. An analysis of the above methods is used to design and implement new methods to resolve real transportation issues, their results being compared with the methods provided by the commercial software product.
|
|
|
Methods for solving selected vehicle routing problems and their implementation.
Drobný, Michal ; Grygarová, Libuše (advisor) ; Zimmermann, Karel (referee)
Various types of transportation issues are a common practice. The issue may be approached mainly as the distribution of products from suppliers to consumers while minimising distribution costs. The difference of real transportation issues predominantly relates to the considered restrictions, such as capacities of vehicles and orders, time windows and other special distribution restrictions. Transportation issues were already defined by F.L. Hitchcock in 1941 and since then, a wide range of stochastic and non- determinist methods providing solutions to transportation issues have been developed. Nevertheless, introducing distribution restrictions in resolving real-life problems makes it difficult for such methods to be applied. This thesis provides a compilation of the well-known determinist methods that may be used to resolve transportation issues. The methods that are appropriate for resolving real issues are discussed in more detail. The solution procedure of the selected method is demonstrated using simple examples and the results are compared with the results of other methods. An analysis of the above methods is used to design and implement new methods to resolve real transportation issues, their results being compared with the methods provided by the commercial software product.
|
|
|
Multi-layered neural networks and visualization of their structure
Drobný, Michal ; Mrázová, Iveta (advisor) ; Kukačka, Marek (referee)
The model of multi-layered neural networks of the back-propagation type is well-known for their universal approximation capability and even the standard back-propagation training algorithm used for their adjustment often provides results applicable to real-world problems. The present study deals with the issue of the multi-layered neural networks. It describes selected variants of training algorithms, mainly the standard back-propagation training algorithm and the scaled conjugate gradients algorithm, which ranks among the extremely fast second-order algorithms. One of the parts of the present study is also an application for the visualisation of the structure of multi-layered neural networks whose solution is designed with respect to its potential utilization in the education of artificial intelligence. The first part of the study introduces the subject matter and formally describes both algorithms, followed by a short description of other variants of the algorithms and their analysis. The next part discusses the selection of the appropriate programming language for the implementation of the application, specifies the goals and describes the implementation works. The conclusion summarizes the test results of the speed and implementation comparison with the selected noncommercial-based software ENCOG.
|
|
|
Design and implementation of vehicle routing problem software library
Drobný, Michal ; Gála, Libor (advisor) ; Melechovský, Jan (referee)
Various types of transportation issues are a common practice. The issue may be approached mainly as the distribution of products from suppliers to consumers while minimising distribution costs. The difference of real transportation issues predominantly relates to the considered restrictions, such as capacities of vehicles and orders, time windows and other special distribution restrictions. There are some specialized commercial products for optimizing distribution costs presented on the market. These are often part of larger frameworks, which also provides other services such as vehicle tracking, warehouse location, ride statements etc. These products provide cloud services fully customized according to the customer. This work aims to design and construct an open source software library for solving vehicle routing problem, which will be able to reflect requirements of simplicity, scalability and portability. The constructed library can serve not only as an open source solution for vehicle routing problems, but also as a tool for testing new methods studied on university premises. The work is divided into several parts. In the first part we define the goals of this work, we define the vehicle routing problem and we describe the principles of the solution. In the next section, we propose the structure of the software, we discuss the choice of aids and document a design phase. The last part deals with the implementation phase, validation phase and real use of the software.
|
guest ::
RSS feed.