PL
 |
EN
Preferencje help
Polish version English version Język
Widoczny [Schowaj] Abstrakt
Liczba wyników

Znaleziono wyników: 16

Liczba wyników na stronie
first rewind previous Strona / 1 next fast forward last
Wyniki wyszukiwania
Wyszukiwano:
w słowach kluczowych:  evolutionary neural networks
help Sortuj według:

help Ogranicz wyniki do:
first rewind previous Strona / 1 next fast forward last
1
Content available Collectively intelligent prediction in evolutionary multi-agent system
Kijak J., Martyna P., Byrski A., Faber Ł., Piętak K., Kisiel-Dorohinicki M.
International Journal of Microelectronics and Computer Science
|
2017
|
Vol. 8, nr 3
85--96
EN
In the paper a summary of our previously realized and published work connected with constructing collective intelligent evolutionary multi-agent systems for time series prediction, based on multi-layered perceptrons is shown. Besides recalling our past papers, we describe the whole concept, present an implementation in a contemporary, componentoriented software framework AgE 3.0 and we conduct a number of experiments, finding different optimal parametrization for the considered instances of the problems (popular Mackey-Glass chaotic time series). The paper may be useful for a practitioner willing to use our meatheuristic algorithm (EMAS) along with the idea of collective agent-based system in order to realize prediction tasks.
2
Content available remote Assembler Encoding with Evolvable Operations
Praczyk T.
Computational Methods in Science and Technology
|
2015
|
Vol. 21, No. 3
123--139
EN
Assembler Encoding is a neuro-evolutionary method which represents a neural network in the form of a linear program. The program consists of operations and data and its goal is to produce a matrix including all the information necessary to construct a network. In order for the programs to produce effective networks, evolutionary techniques are used. A genetic algorithm determines an arrangement of the operations and data in the program and parameters of the operations. Implementations of the operations do not evolve, they are defined in advance by a designer. Since operations with predefined implementations could narrow down applicability of Assembler Encoding to a restricted class of problems, the method has been modified by applying evolvable operations. To verify effectiveness of the new method, experiments on the predator-prey problem were carried out. In the experiments, the task of neural networks was to control a team of underwater-vehicles-predators whose common goal was to capture an underwater-vehicle-prey behaving by a simple deterministic strategy. The paper describes the modified method and reports the experiments.
3
Content available remote Diverse Neural Architectures in Assembler Encoding
Praczyk T.
Computational Methods in Science and Technology
|
2014
|
Vol. 20, No. 1
21--34
EN
The paper presents a neuro-evolutionary method called Assembler Encoding (AE) and proposes its several modifications. The main goal of the modifications is to ensure AE greater freedom in generating diverse neural architectures. To compare the modifications with each other and with the original method the particular case of the predator-prey problem has been discussed.
4
Content available Using Assembler Encoding to build neuro-controllers for a team of autonomous underwater vehicles
Praczyk T., Szymak P.
Control and Cybernetics
|
2013
|
Vol. 42, no. 1
267--286
EN
The paper compares a neuro-evolutionary metod called Assembler Encoding with two other methods from the area of neuro–evolution. As a testbed for the methods a variant of the predator–prey problem with Autonomous Underwater Vehicles (AUV) operating in an environment with the sea current was used. In the experiments, the task of vehicles–predators controlled with evolutionary neural networks was to capture a vehicle–prey behaving according to a simple deterministic strategy. All the experiments were carried out in simulation, and in order to simplify calculations in the two–dimensional environment – AUVs moved on a horizontal surface under the water.
5
Content available Real-time training algorithms in neuroevolutionary navigational decision support system
Łącki M.
Zeszyty Naukowe Akademii Marynarki Wojennej
|
2012
|
R. 53 nr 4 (191)
93-104
EN
The paper presents the idea of using advanced machine learning algorithms to aid decision making in ship manoeuvring in real time. Evolutionary neural networks are used in this purpose. In the simulated model of manoeuvring ship a helmsman is treated as an individual in population of competitive helmsmen, which through environmental sensing and evolution processes learn how to navigate safely through restricted waters.
PL
Artykuł przedstawia koncepcję wykorzystania zaawansowanych algorytmów uczenia się maszyn dla wsparcia podejmowania decyzji manewrowania okrętem w czasie rzeczywistym. Do tego celu wykorzystywane są ewolucyjne sieci neuronowe. W symulowanym modelu manewrowania okrętem sternik jest traktowany jako jednostka w populacji konkurencyjnych sterników, którzy poprzez wyczuwanie środowiskowe i procesy ewolucyjne uczą się jak prowadzić nawigację bezpiecznie po ograniczonych akwenach.
6
Content available remote Assembler Encoding Improved
Praczyk T.
Computational Methods in Science and Technology
|
2012
|
Vol. 18, No. 1
11-24
EN
Assembler Encoding is a neuro-evolutionary method which was used to produce a neural decision system for a team of autonomous underwater vehicles. Since results accomplished during experiments with the classic variant of Assembler Encoding appeared to be unsatisfactory, the method has been appropriately improved. The paper presents modifications to Assembler Encoding and reports experiments whose main goal was to test effectiveness of each of them.
7
Content available Properties of Assembler encoding
Praczyk T.
Zeszyty Naukowe Akademii Marynarki Wojennej
|
2010
|
R. 51 nr 1 (180)
51-63
EN
Assembler Encoding is a new neuro-evolutionary method. In the paper, characterization of the method is given. To characterize Assembler Encoding, the following properties were taken into consideration: Completeness, Closure, Compactness, Scalability, Multiplicity, Ontogenetic Plasticity, Modularity, and Redundancy.
PL
Kodowanie Assembler jest metodą neuro-ewolucyjną. W artykule scharakteryzowano ją, biorąc pod uwagę następujące własności: kompletność, zamknięcie, kompaktowość, mierzalność, plastyczność ontogenetyczną, modularność oraz redundancję.
8
Content available remote Searching for optimal size neural networks in Assembler Encoding
Praczyk T.
Control and Cybernetics
|
2010
|
Vol. 39, no 4
1193-1215
EN
Assembler Encoding represents a neural network in the form of a simple program called Assembler Encoding Program. The task of the program is to create the so-called Network Definition Matrix, which maintains all the information necessary to construct a network. To generate the programs and, in consequence, neural networks, evolutionary techniques are used. One of the problems in Assembler Encoding is to determine an optimal number of neurons in a neural network. To deal with this problem a current version of Assembler Encoding uses a solution that is time consuming and hence rather impractical. The paper proposes four other solutions to the problem mentioned. To test them, experiments in a predator-prey problem were carried out. The results of the experiments are included at the end of the paper.
9
Content available remote Improvements to Assembler Encoding
Praczyk T.
Theoretical and Applied Informatics
|
2009
|
Vol. 21, No. 2
95-105
EN
Assembler Encoding is a neuro-evolutionary method which was used to produce neural controllers for a group of autonomous underwater vehicles. Since results accomplished during experiments with the classic variant of Assembler Encoding appeared to be unsatisfactory the method has been appropriately improved. The paper presents modifications made to Assembler Encoding and reports experiments whose the main goal was to test effectiveness of the modifications mentioned.
PL
Kodowanie Asemblerowe jest metodą neuro-ewolucyjną w której sieć neuronowa reprezentowana jest w postaci prostego programu składającego się z ciągu operacji i danych. Zadaniem wspomnianego programu jest utworzenie macierzy zawierającej pełną informację potrzebną do konstrukcji sieci neuronowej. W Kodowaniu Asemblerowym tworzenie programów odbywa się z wykorzystaniem technik ewolucyjnych. Do tej pory Kodowanie Asemblerowe zostało sprawdzone eksperymentalnie w szeregu problemach. Jednym z nich jest problem typu predator-prey, w którym zarówno drapieżnicy jak i ofiara są autonomicznymi pojazdami podwodnymi. W trakcie eksperymentów zadaniem Kodowania Asemblerowego było tworzenie neuro-kontrolerów sterujących poczynaniem zespołu pojazdów-drapieżników, których głównym celem było pochwycenie pojazdu-ofiary postępującego zgodnie z pewną prostą deterministyczną strategią. Podczas wstępnych badań okazało się, że wyniki sieci neuronowych tworzonych za pomocą Kodowania Asemblerowego są niesatysfakcjonujące i wymagają poprawy. Aby zwiększyć skuteczność sieci opracowano szereg modyfikacji metody ich tworzenia. Niniejszy artykuł prezentuje każdą z modyfikacji oraz przedstawia wyniki badań uzyskane w rezultacie zastosowania każdej z nich.
10
Content available remote Assembler Encoding Versus Connectivity Matrix Encoding in the Inverted Pendulum Problem with a Hidden State
Praczyk T.
Computational Methods in Science and Technology
|
2009
|
Vol. 15, nr 2
177-183
EN
Assembler Encoding is the Artificial Neural Network encoding method. To date, Assembler Encoding has been tested in the optimization problem and in the so-called predator-prey problem. The paper reports experiments in a next test problem, i.e. in the inverted pendulum problem. To compare Assembler Encoding with other Artificial Neural Network encoding methods in the experiments, two direct encodings were also tested.
11
Content available Organization of the evolutionary process responsible for creating neural networks in assembler encoding
Praczyk T.
Biuletyn Wojskowej Akademii Technicznej
|
2009
|
Vol. 58, nr 2
103-122
EN
Assembler Encoding (AE) represents Artificial Neural Network (ANN) in the form of a simple program called Assembler Encoding Program (AEP). The task of AEP is to create the socalled Network Definition Matrix (NDM) including all the information necessary to construct ANN. AEPs and in consequence ANNs are formed by means of evolutionary techniques. To make AE an effective tool for creating ANNs it is necessary to appropriately organize all the evolutionary processes responsible for generating AEPs, i.e., it is necessary to properly select values of different parameters controlling the evolutionary process mentioned. To determine optimal conditions of the evolution in AE, experiments in a predator-prey problem were performed. The results of the experiments are presented at the end of the paper.
PL
Kodowanie asemblerowe jest metodą wykorzystującą metody ewolucyjne do tworzenia sieci neuronowych. W kodowaniu asemblerowym sieci neuronowe ewoluują w wielu oddzielnych populacjach. Stworzenie pojedynczej sieci neuronowej wymaga połączenia elementów pochodzących z różnych populacji. Aby sieci neuronowe tworzone w ten sposób były wysokiej jakości konieczne jest odpowiednie sterowanie ewolucją w każdej populacji. Artykuł prezentuje wyniki badań, których głównym celem było określenie zasad prowadzenia ewolucji w Kodowaniu Asemblerowym.
12
Content available Wspomaganie sterowania statkiem za pomocą ewolucyjnych sieci neuronowych
Łącki M.
Zeszyty Naukowe / Akademia Morska w Szczecinie
|
2008
|
nr 14 (86)
34-37
PL
W artykule przedstawiono koncepcję zastosowania ewolucyjnych sieci neuronowych we wspomaganiu procesów podejmowania decyzji podczas manewrowania statkiem na ograniczonym obszarze. Rozważane są wybrane algorytmy, operacje genetyczne, metody kodowania i selekcji oraz struktury ewolucyjnych sieci neuronowych.
EN
This paper describes a concept of evolutionary neural networks application in decision process support during vessel manoeuvring in a restricted area. Selected algorithms, genetic operations, methods of coding and selection, and structures of evolutionary neural networks are considered in the paper.
13
Content available Assembler Encoding : a new Artificial Neural Network encoding method
Praczyk T.
Biuletyn Wojskowej Akademii Technicznej
|
2008
|
Vol. 57, nr 3
395-429
EN
The main goal of the paper is to outline a new Artificial Neural Network (ANN) encoding method called Assembler Encoding (AE). In AE, ANN is encoded in the form of a program (Assembler Encoding Program - AEP) of linear organization and of a structure similar to the structure of a simple assembler program. The task of AEP is to create the so-called Network Definition Matrix (NDM) including the whole information necessary to produce ANN. To create AEPs, and in consequence ANNs, genetic algorithms are used.
PL
Głównym celem artykułu jest przedstawienie Kodowania Asemblerowego czyli nowej metody kodowania sztucznych sieci neuronowych. W Kodowaniu Asemblerowym sieć neuronowa jest zakodowana w postaci programu (AEP - Assembler Encoding Program) o liniowej organizacji i o strukturze podobnej do struktury prostego programu asemblerowego. Zadaniem AEP jest stworzenie tzw. Macierzy Definicji Sieci (NDM - Network Definition Matrix) zawierającej całą informację potrzebną do stworzenia sieci. Tworzenie AEP i w konsekwencji sieci neuronowych odbywa się z wykorzystaniem technik ewolucyjnych.
14
Content available remote Selecting genetic algorithms for assembler encoding
Praczyk T.
Archives of Control Sciences
|
2008
|
Vol. 18, no. 4
477-495
EN
Assembler Encoding makes it possible to use genetic algorithms to construct neural networks. Assembler Encoding represents neural network in a form of the so-called Assembler Encoding Program. The task of the program is to create Network Definition Matrix maintaining all the information necessary to construct the network. In Assembler Encoding different components of Assembler Encoding Programs evolve in separate populations. The evolution in each population can be controlled by a different genetic algorithm. In the experiments reported in the paper the following genetic algorithms were used to control the evolution of programs: Canonical Genetic Algorithm, Steady State Genetic Algorithm and Eugenic Algorithm. The programs created by means of the specified algorithms were used to create neural networks. Then, the networks were tested in the so-called predator-prey problem. The results of the experiments are presented at the end of the paper.
15
Content available remote Concepts of learning in assembler encoding
Praczyk T.
Archives of Control Sciences
|
2008
|
Vol. 18, no. 3
323-337
EN
Assembler Encoding (AE) represents Artificial Neural Network (ANN) in the form of a simple program called Assembler Encoding Program (AEP). The task of AEP is to create the so-called Network Definition Matrix (NDM) maintaining the whole information necessary to construct ANN. To generate AEPs and in consequence ANNs genetic algorithms are used. Using evolution is one of the methods to create optimal ANNs. Another method is learning. During learning parameters of ANN, e.g. weights of interneuron connections, adjust to the task performed by ANN. Usually, combining both methods accelerates generating optimal ANNs. The paper addresses the problem of simultaneous use of the evolution and learning in AE.
16
Evolution of neural networks in a multi-agent world
Kisiel-Dorohinicki M., Klapper-Rybicka M.
Zeszyty Naukowe Uniwersytetu Jagiellońskiego. Prace Informatyczne
|
2000
|
z. 10
37-50
EN
The concept of decentralised evolutionary computation realised as evolutionary multi-agent system (EMAS) is described in the paper. Also agent-based evolutionary approach to neural network architecture optimisation is presented. Then the problem of tme-series prediction and a general idea of evolutionary neural multi-agent predicting system is introduced. Selected design issues together with preliminary simulation results conclude the work.
first rewind previous Strona / 1 next fast forward last
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.