Wyniki wyszukiwania - Biblioteka Nauki

1

Hybrid Data Exploration Methods to Prediction Tasks Solving

100%

Sikora M. , Kozielski M.

|

2006

|

tom Vol. 18, nr 1

57-74

EN

The paper presents the data exploration issue in a context of the prediction tasks solution. There arę presented three methodologies of data models construction in the paper. The methodologies are used to solve the prediction tasks. They also deliver an easily interpretable knowledge about the modeled process. The methodologies, called hybrid methods, combine analytical algorithms (clustering, rules induction) and optimization algorithms (genetic algorithms, error backpropagation). The paper presents also a method of a data transformation between the temporal data representation and the data representation acceptable. for algorithms using learning by examples paradigm. The paper presents also the results of the experiments performed on the benchmark data and the data obtained from a methane concentration monitoring system in a coal mine.

PL

W artykule przedstawiono problem eksploracji danych w kontekście realizacji zadań związanych z predykcją. Przedstawiono i porównano trzy metodologie uzyskiwania modeli danych, które poza tym, że wykorzystywane są do rozwiązywania zadań predykcyjnych dostarczają również stosunkowo łatwo interpretowalnej wiedzy o opisywanym procesie. Przedstawione metodologie łączą algorytmy analityczne (grupowanie danych, indukcja reguł) z metodami optymalizacji (algorytmy genetyczne, wsteczna propagacja błędu) stąd nazwa metody hybrydowe. W artykule omówiono również metodę przejścia z temporalnej reprezentacji danych na reprezentację akceptowalną przez algorytmy wykorzystujące paradygmat uczenia na podstawie przykładów. Na zakończenie przedstawiono wyniki eksperymentów przeprowadzonych na zbiorze danych benchmarkowych oraz na zbiorze pochodzącym z bazy danych systemu monitorującego wydzielanie metanu w kopalniach.

2

Adaptation of symbiotic adaptive neuroevolution in assembler encoding

100%

Praczyk T.

|

2008

|

tom Vol. 20, No. 1

49-68

EN

Assembler Encoding represents 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 the whole information necessary to construct the network. To generate Assembler Encoding Programs and in consequence neural networks evolutionary techniques are used. In order to use evolutionary techniques to construct Assembler Encoding Programs it is necessary to encode them in the form of chromosomes. The simplest solution is to place the whole information necessary to construct the program into one chromosome. The paper suggests another approach. Methods proposed in the paper are an adaptation of Symbiotic Adaptive Neuro-evolution. To test the methods proposed they were used to solve a few simple optimization problems.

3

Exchange Rates: Predictable but not Explainable? Data Mining with Leading Indicators and Technical Trading Rules

80%

Brandl B. , Department of Government U. o. V.

|

2005

|

tom 192

EN

This paper presents a data mining approach to forecasting exchange rates. It is assumed that exchange rates are determined by both fundamental and technical factors. The balance of fundamental and technical factors varies for each exchange rate and frequency. It is difficult for forecasters to establish the relative relevance of different kinds of factors given this mixture; therefore the utilization of data mining algorithms is advantageous. The approach applied uses a genetic algorithm and neural networks. Out-of-sample forecasting results are illustrated for five exchange rates on different frequencies and it is shown that data mining is able to produce forecasts that perform well.

PL

W artykule przedstawiono proces eksploracji danych statystycznych w prognozowaniu kursów walutowych. Zakładamy, że kursy walutowe pozostają pod wpływem zarówno czynników o charakterze fundamentalnym, jak i czynników pozaekonomicznych. Równowaga pomiędzy tymi czynnikami różni się w zależności od rodzaju kursu walutowego i częstotliwości jego pomiaru. Prognostykom trudno jest ustalić względną siłę wpływu różnych czynników, stąd analiza polegająca na eksploracji danych ma określone zalety. W proponowanym podejściu wykorzystano algorytmy genetyczne i sztuczne sieci neuronowe. Przedstawiliśmy wyniki eksperymentów prognostycznych poza próbą statystyczną w odniesieniu do pięciu kursów walutowych, obserwowanych z różną częstotliwością. Pokazaliśmy, że metoda eksploracji danych może stanowić skuteczne narzędzie prognostyczne.

4

A hybrid approach for scheduling transportation networks

80%

Dridi M. , Kacem I.

International Journal of Applied Mathematics and Computer Science

|

2004

|

tom 14

|

nr 3

397-409

EN

In this paper, we consider a regulation problem of an urban transportation network. From a given timetable, we aim to find a new schedule of multiple vehicles after the detection of a disturbance at a given time. The main objective is to find a solution maximizing the level of service for all passengers. This problem was intensively studied with evolutionary approaches and multi-agent techniques, but without identifying its type before. In this paper, we formulate the problem as a classical one in the case of an unlimited vehicle capacity. In the case of a limited capacity and an integrity constraint, the problem becomes difficult to solve. Then, a new coding and well-adapted operators are proposed for such a problem and integrated in a new evolutionary approach.

5

USING GENETIC ALGORITHM IN DYNAMIC MODEL OF SPECULATIVE ATTACK

80%

Gawrońska-Nowak B. , Grabowski W.

Equilibrium. Quarterly Journal of Economics and Economic Policy

|

2016

|

tom 11

|

nr 2

287-306

EN

Evolution of speculative attack models shows certain progress in developing the idea of the role of expectations in the crisis mechanism. Obstfeld (1996) defines expectations as fully exogenous. Morris and Shin (1998) treat the expectations as endogenous (with respect to noise), not devoting too much attention to information structure of the foreign exchange market. Dynamic approach proposed by Angeletos, Hellwig and Pavan (2006) offers more sophisticated assumption about learning process. It tries to reflect time-variant and complex nature of information. However, this model ignores many important details like a Central Bank cost function. Genetic algorithm allows to avoid problems connected with incorporating information and expectations into agent decision-making process to an extent. There are some similarities between the evolution in Nature and currency market performance. In our paper an assumption about rational agent behaviour in the efficient market is criticised and we present our version of the dynamic model of a speculative attack, in which we use a genetic algorithm (GA) to define decision-making process of the currency market agents. The results of our simulation seem to be in line with the theory and intuition. An advantage of our model is that it reflects reality in a quite complex way, i.e. level of noise changes in time (decreasing), there are different states of fundamentals (with “more sensitive” upper part of the scale), the number of inflowing agents can be low or high (due to different globalization phases, different capital flow phases, different uncertainty levels).

6

A system identification approach to material characterisation a new method for visco-elastic characterisation using genetic algorithm

80%

Ozada N. , Mares C. , Esat I. I. , Djakab E.

|

2007

|

tom Vol. 7, nr 2(12)

264-275

EN

Material behavior is described by constitutive models. These models rely on various physical laws to describe the relationship between same input and output variables. For example stress and strain relationship can be described by a matrix formulation. The elements of matrix are described by various coefficients associated with the material modeled. This is the traditional method and bas worked effectively especially for linear material behavior. Extending such an approach to more complex material behavior requires understanding of material at macro or molecular level. It would be probably impossible for scientist to describe material behavior for every material under every possible condition. Therefore it is necessary to look for alternative methods to describe material behavior. The paper addresses this issue, presents GA as a generic problem solving tool and demonstrate how it can be used as a material characterization tool.

7

Optymalizacja sposobu rozmieszczenia latarni oświetlonych z wykorzystaniem algorytmu ewolucyjnego

80%

Gajer M.

|

2012

|

tom R. 80, nr 4

25-28

PL

Przedstawiono propozycję wykorzystania techniki obliczeń ewolucyjnych do poszukiwania optymalnego sposobu rozmieszczania słupów latarni oświetleniowych.

EN

The paper presents a proposal of using genetic calculations in quest of optimal spatial arrangement of lamp-posts.

8

Application of Genetic algorithms in graph searching problem

80%

Wrona Ł. , Jaworski B.

|

2010

|

tom T. 18

353-358

EN

Graph searching is a common approach to solving a problem of capturing a hostile intruder by a group of mobile agents. We assume that this task is performed in an environment which we are able to model as a graph G. The question asked is how many agents are needed to capture an arbitrary fast, invisible and smart intruder. This number is called the (edge) search number of G. The strategy which must be performed by agents is called the (edge) search strategy. Unfortunately calculating both the optimal search strategy and the search number is NP-hard for general graphs. Furthermore, due to the complexity of the pursuit rules, the application of heuristic solutions is not straightforward. In this paper we suggest a method of applying genetic algorithms to solve graph searching problem. The idea is based on LaPaugh's result on graph searching monotonicity and utilizes representation of a search strategy as a permutation of edges.

PL

Przeszukiwanie grafów to często stosowane podejście do problemu przechwytywania wrogiej jednostki przez grupę mobilnych agentów. Zadanie to jest realizowane w środowisku zamodelowanym za pomocą grafu G. Odpowiadamy na pytanie ile mobilnych agentów jest niezbędnych by przechwycić dowolnie szybkiego, niewidzialnego i bystrego intruza. Liczba ta jest nazywana (krawędziową) liczbą przeszukiwawczą G a strategia, którą realizują agenci - (krawędziową) strategią przeszukiwawczą. Zarówno wyznaczanie liczby przeszukiwawczej jak i strategii przeszukiwawczej jest trudne obliczeniowo. Dodatkowo, ze względu na złożoność zasad, według których odbywa się przechwytywanie, stosowanie podejścia heurystycznego jest utrudnione. W tej pracy sugerujemy metodę zastosowania algorytmów genetycznych w rozwiązywaniu omawianego problemu. W pomyśle wykorzystany jest lemat LaPaugha, dotyczący monotoniczności przeszukiwania grafów i zakłada reprezentację strategii przeszukiwawczej jako permutacji krawędzi.

9

Neurofuzzy Adaptive B-spline Circuit Performance Modelling Aided by Means of Genetic Algorithm

80%

Horbowski R. , Białko M.

|

2001

|

tom vol. 49, nr 3

445-459

EN

An application of a lattice adaptive neurofuzzy system to circuit performance modelling is presented in the paper. The investigated system makes use of the B-spline membership functions and a structure of the system is determined by means of a genetic algorithm. The presented approach possesses good modelling capabilities and, contrary to non-lattice neurofuzzy approaches, can explore structural dependencies existing in training data supporting us with valuable knowledge about the modelled circuit. This knowledge gives an insight into behaviour of the modelled performance function and makes it possible to reduce a size of the set of circuit variables, to simplify the structure of the model and hence to speed-up its evaluation and identification.

10

Structure Identification of an Associative Memory Network Using Genetic Algorithm

80%

Horbowski R. , Białko M.

|

2001

|

tom vol. 49, nr 1

125-139

EN

In the paper a novel approach to the structure identification of the lattice associative memory network is presented. In the investigated network B-spaline basis functions are used, and the network structure (i.e.a number, a distribution and a shape of the basis functions) is determined by means of a genetic algorithm. The proposed approach is partly able to make use of structural dependencies existing in training data which, in connection with high interpretability of the lattice network, can provide the modeller with valuable knowledge about the process or system being modelled. This knowledge can help to select relevant model inputs and hence the model size.

11

Optimal location of buoys in automatic radar navigational system

80%

Praczyk T.

|

2009

|

tom nr 3

CD-CD

EN

The paper is devoted to construction of a system to coastal navigation. To fix position the system uses the information about buoys surrounding a ship. Accuracy of the system is influenced on location of buoys on the given area of the sea. The paper suggests using genetic algorithms to arrange the buoys. The solution proposed was tested and results of the tests are presented at the end of the paper.

PL

Artykuł poświęcony jest konstrukcji automatycznego, zapasowego systemu nawigacji radarowej. Jedna z koncepcji wspomnianego systemu zakłada wykorzystanie systemu pław do wyznaczenia pozycji jednostki. Dokładność takiego systemu zależy od rozmieszczenia pław na akwenie. W artykule zaproponowano algorytmy genetyczne do określenia optymalnej lokalizacji pław.

12

Procedure application in assembler encoding

80%

Praczyk T.

|

2007

|

tom Vol. 17, no. 1

71-91

EN

In order to use evolutionary techniques to search for optimal neural networks it is necessary to encode the latter in the form of chromosome or a set of chromosomes. In the paper a new neural network encoding method is presented - assembler encoding (AE). It assumes neural network encoded in the form of linearly organized structure similar to assembler program with code part and with data part. The task of assembler code is to create connectivity matrix which in turn can be transferred into neural network with any architecture. In the article the variant of AE in which we deal with application of procedures is discussed. Assembler encoding programs consisting of many procedures are used to solve optimization problem. Results of tests conducted are included in the paper.

13

Parallel genetic algorithm finding optimal initial mesh for three dimensional self-adaptive hp finite element metho

80%

Paszyńska A. , Paszyński M.

|

2011

|

tom Vol. 11, No. 1

34-40

EN

In this paper we investigate the parallel version of the hierarchical chromosome based genetic algorithm (HCBGA) for finding the optimal initial mesh for self-adaptive hp-Finite Element Method (hp-FEM). The HCBGA algorithm solves the global optimization problem of r refinements in order to provide optimal starting mesh for the hp-FEM that will fit material data and singularities, and will result in the exponential convergence of the hp-FEM. The parallel algorithm is tested on the damaged Step-and-Flash Imprint Lithography problem, modeled by linear elasticity with thermal expansion ':, coefficient.

PL

W artykule tym przedstawiono wersję równoległą algorytmu genetycznego bazującego na hierarchicznym chromosomie (AGBHC) służącego do znajdowania optymalnych siatek począt-kowych dla hp adaptacyjnej metody elementów skończonych (hp-MES). Zaproponowany algorytm AGBHC rozwiązuje problem r adaptacji. Jest to problem globalnej optymalizacji polegający na znalezieniu optymalnej siatki początkowej dla algorytmu automa-tycznej hp adaptacji. Poszukiwana siatka początkowa powinna pasować do przyjętych stałych materiałowych oraz do osobliwości rozwiązania. W efekcie algorytm automatycznej hp adaptacji uruchomiony na takiej optymalnej siatce początkowej powinien dostarczyć eksponencjalną zbieżność dokładności rozwiązania względem rozmiaru siatki obliczeniowej. Algorytm równoległy testowany jest na problemie nanolitografii poprzez wyciskanie i naświetlanie, modelowanym z pomocą liniowej sprężystości ze współczynnikiem rozszerzalności cieplnej.

14

Multiobjective optimization of a fuzzy PID controller

80%

Soukkou A. , Khellaf A. H. , Leulmi S.

|

2006

|

tom Vol. 16, no. 4

445-461

EN

A fuzzy logic controller with multilayer neutral network whose synaptic weights represent the fuzzy knowledge base and its application to the highly nonlinear systems is presented in this work. The scaling factors of the input variables, membership functions and the rule sets are optimized by the use of the multiobjective genetic algorithms. The fuzzy network structure is specified by a combination of the mixed Takagi-Sugeno's and Mamdani's fuzzy reasoning. The mixed, Binary-Real-Integer, optimal coding is utilized to construct the chromosomes, which define the same of necessary prevai;ling parameters for the conception of the desired controller. This new controller stands out by a non-standard gain, which varies lineary with the fazzy inputs. Under certain conditions, it becomes similar to the conventional PID controller with non-linearly variable coefficients. Computer simulation indicates that the designed fuzzy controller is satisfactory in control of a nonlinear system "Inverted Pendulum".

15

Nowe spojrzenie na problem strojenia regulatora PID - charakterystyka metod i narzędzi komputerowego modelowania

80%

Giernacki W. , Ziętkiewicz J.

|

2007

|

tom T. 32

75-84

PL

W niniejszej pracy przedstawiono nowe spojrzenie na problem strojenia regulatora PID. Wykorzystano algorytmy genetyczne jako alternatywę dla metody Zieglera-Nicholsa i nieoptymalnych metod bazujących na wiedzy eksperckiej. Problem poszukiwania nastaw regulatora został sprowadzony do problemu optymalizacji, w którym całkowe wskaźniki jakości (IAE, ISE, ITAE) stanowią optymalizowane funkcje (funkcje przystosowania). Ponadto, krótko omówiono narzędzia komputerowego modelowania dostępne w MATLAB 7.0/SIMULINK. Zaprezentowano wyniki symulacyjne oraz analizę porównawczą metod strojenia regulatora dla wybranego obiektu regulacji. Strojenie regulatora metodą algorytmów genetycznych dawało w efekcie najlepszą jakość regulacji.

EN

A new look at the problem of PID controller tuning is shown in the article. Genetic algorithms were used as an alternative for Ziegler-Nichols method and for suboptimal methods basing on expert's knowledge. The search problem of optimal controller parameters was formulated to an optimization problem, in which the quality integral indices (IAE, ISE, ITAE) establish the optimized functions (fitness functions). Furthermore, the methods implemented in MATLAB 7.0/SIMULINK environment were concisely discussed. It demonstrates tests results and comparative analysis of controller tuning metods with selected control plant. PID controller tuning by genetic algorithms method generated the best control quality.

16

Evolutionary learning of rich neural networks in the Bayesian model selection framework

80%

Matteucci M. , Spadoni D.

International Journal of Applied Mathematics and Computer Science

|

2004

|

tom 14

|

nr 3

423-440

EN

In this paper we focus on the problem of using a genetic algorithm for model selection within a Bayesian framework. We propose to reduce the model selection problem to a search problem solved using evolutionary computation to explore a posterior distribution over the model space. As a case study, we introduce ELeaRNT (Evolutionary Learning of Rich Neural Network Topologies), a genetic algorithm which evolves a particular class of models, namely, Rich Neural Networks (RNN), in order to find an optimal domain-specific non-linear function approximator with a good generalization capability. In order to evolve this kind of neural networks, ELeaRNT uses a Bayesian fitness function. The experimental results prove that ELeaRNT using a Bayesian fitness function finds, in a completely automated way, networks well-matched to the analysed problem, with acceptable complexity.

17

Comparison of genetic and ant algorithms

80%

Boryczka U.

|

2000

|

tom T. 12, z. 4

249-264

EN

Genetic and ant algorithms apply stochastic searching, parallel investigation as well as autocatalitic process (or stigmergy) to solve optimization problems. This paper concentrates on the Traveling Salesman Problem (TSP) solved by genetic and ant algorithms. We consider the sets of parameters and operators which influence the acting of these algorithms. Two algorithmic structures emphasizing the selection problem are discussed. We describe the TSP experiments performed for 50 cities. The aim of the comparison is the conclusion that the evolution, which is exploited in genetic algorithms, can improve the performance of ant algorithms.

PL

Artykuł porównuje możliwości algorytmów genetycznych i mrówkowych na przykładzie problemu komiwojażera. Rozważono szereg parametrów mających wpływ na funkcjonowanie wymienionych typów algorytmów. Problem komiwojażera był rozważany dla sieci 50 miast. Celem porównania jest pokazanie, że ewolucja, która jest podstawą funkcjonowania algorytmów genetycznych, zastosowana do algorytmów mrówkowych może zwiększyć ich wydajność.

18

Application of genetic algorithms to computer assignment problem in distributed hard real-time systems

80%

Martyna J.

|

1998

|

tom z. 8

45-61

EN

This paper is devoted to a fundamental problem arising during a design of distributed hard real-time systems - a cost minimization of located computers in a network working in a time-constrained environment. As regard our problem, each computer location must take into consideration a prper class of processor. In contrast to the existing traditional solutions of this problem , we used genetic algorithms for a required optimization. We present its application in obtaining all main parameters of designed systems. Genetic algorithms are shown to be effective for the solution of computer assignment problem in distributed hard real-time systems.

19

An efficient genetic algorithm for the uncapacitated multiple allocation p-hub median problem

80%

Stanimirovic Z.

|

2008

|

tom Vol. 37, no 3

669-692

EN

In this paper the Uncapacitated Multiple Allocation p-hub Median Problem (the UMApHMP) is considered. A new heuristic method based on a genetic algorithm approach (GA) for solving UMApHMP is proposed. The described GA uses binary representation of the solutions. Genetic operators which keep the feasibility of individuals in the population are designed and implemented. The mutation operator with frozen bits is used to increase the diversibility of the genetic material. The running time of the GA is improved by caching technique. Proposed GA approach is bench-marked on the well known CAB and AP data sets and compared with the existing methods for solving the UMApHMP. Computational results show that the GA quickly reaches all previously known optimal solutions, and also gives results on large scale AP instances (up to n=200, p=20) that were not considered in the literature so far.

20

Topology and shape optimization of continuum structures by genetic algorithm and BEM

80%

Kita E. , Tamaki T. , Tanie H.

|

2004

|

tom Vol. 11, No. 1

63-75

EN

This paper describes the topology and shape optimization scheme of continuum structures by using genetic algorithm (GA) and boundary element method (BEM). The structure profiles are defined by using the spline function surfaces. Then, the genetic algorithm is applied for determining the structure profile satisfying the design objectives and the constraint conditions. The present scheme is applied to minimum weight design of two-dimensional elastic problems in order to confirm the validity.