Wyniki wyszukiwania - BazTech

1

Stochastic schemata exploiter-based optimization of hyper-parameters for XGBoost

Makino Hiroya, Kita Eisuke

Computer Assisted Methods in Engineering and Science

|

2024

|

Vol. 31, no. 1

113--132

EN

XGBoost is well-known as an open-source software library that provides a regularizing gradient boosting framework. Although it is widely used in the machine learning field, its performance depends on the determination of hyper-parameters. This study focuses on the optimization algorithm for hyper-parameters of XGBoost by using Stochastic Schemata Exploiter (SSE). SSE, which is one of Evolutionary Algorithms, is successfully applied to combinatorial optimization problems. SSE is applied for optimizing hyper-parameters of XGBoost in this study. The original SSE algorithm is modified for hyper-parameter optimization. When comparing SSE with a simple Genetic Algorithm, there are two interesting features: quick convergence and a small number of control parameters. The proposed algorithm is compared with other hyper-parameter optimization algorithms such as Gradient Boosted Regression Trees (GBRT), Tree-structured Parzen Estimator (TPE), Covariance Matrix Adaptation Evolution Strategy (CMA-ES), and Random Search in order to confirm its validity. The numerical results show that SSE has a good convergence property, even with fewer control parameters than other methods.

2

Using the Barnacles Mating optimizer for economic emission load dispatch problems

Sulaiman Mohd Herwan, Mustaffa Zuriani, Saari Mohd Mawardi

Przegląd Elektrotechniczny

|

2020

|

R. 96, nr 7

141--144

EN

This paper proposes a recent nature inspired algorithm namely Barnacle Mating Optimizer (BMO) for solving the economic emission load dispatch (EELD) problems. BMO is based on the mating behaviour of barnacles and is treated as an evolutionary computation algorithm in solving the optimization problems. Three cases have been tested using the proposed BMO: 3-units, 10 -units and 40-units system and the performances of BMO are compared with other recent selected algorithms to show the effectiveness of the proposed BMO in solving the EELD problems.

PL

W artykule zaprezentowany inspirowany naturą algorytm nazwany Barnacle Mating przeznaczony do rozwiązywania problemu optymalizacji dystrybucji energii uwzglęa)dniające emisję. Analizowano trzy przypadki – system trzech, dziesięciu I czterdziestu jednostek.

3

Computational Intelligence for Life Sciences

Besozzi Daniela, Manzoni Luca, Nobile Marco S., Spolaor Simone, Castelli Mauro, Vanneschi Leonardo, Cazzaniga Paolo, Ruberto Stefano, Rundo Leonardo, Tangherloni Andrea

Fundamenta Informaticae

|

2020

|

Vol. 171, nr 1-4

57--80

EN

Computational Intelligence (CI) is a computer science discipline encompassing the theory, design, development and application of biologically and linguistically derived computational paradigms. Traditionally, the main elements of CI are Evolutionary Computation, Swarm Intelligence, Fuzzy Logic, and Neural Networks. CI aims at proposing new algorithms able to solve complex computational problems by taking inspiration from natural phenomena. In an intriguing turn of events, these nature-inspired methods have been widely adopted to investigate a plethora of problems related to nature itself. In this paper we present a variety of CI methods applied to three problems in life sciences, highlighting their effectiveness: we describe how protein folding can be faced by exploiting Genetic Programming, the inference of haplotypes can be tackled using Genetic Algorithms, and the estimation of biochemical kinetic parameters can be performed by means of Swarm Intelligence. We show that CI methods can generate very high quality solutions, providing a sound methodology to solve complex optimization problems in life sciences.

4

A comparison of evolutionary and simulated annealing algorithms for bi-criteria location-scheduling problem

Ławrynowicz Mirosław, Filcek Grzegorz

Annals of Computer Science and Information Systems

|

2020

|

Vol. 21

251--255

EN

A comparison of two heuristic algorithms solving a bi-criteria joint location and scheduling (ScheLoc) problem is considered. In this strongly NP-hard problem the sum of job completion times and location investment costs are used to evaluate the solution. The first solution algorithm (EV) uses an evolutionary approach, and the second more time-efficient algorithm (SA) is based on Simulated Annealing.

5

Evolutionary multi-agent system with crowding factor and mass center mechanisms for multiobjective optimisation

Różański Mateusz, Siwik Leszek

Computer Science

|

2019

|

Vol. 20 (3)

343--367

EN

This work presents some additional mechanisms for Evolutionary Multi-Agent Systems for Multiobjective Optimisation trying to solve problems with population stagnation and loss of diversity. Those mechanisms reward solutions located in a less crowded neighborhood and on edges of the frontier. Both techniques have been described and also some preliminary results have been shown.

6

Genetic minimisation of peak-to-peak level of a complex multi-tone signal

Chruszczyk Ł.

Bulletin of the Polish Academy of Sciences. Technical Sciences

|

2019

|

Vol. 67, nr 3

621--629

EN

This paper presents results of evolutionary minimisation of peak-to-peak value of a?multi-tone signal. The signal is the sum of multiple tones (channels) with constant amplitudes and frequencies combined with variable phases. An exemplary application is emergency broadcasting using widely used analogue broadcasting techniques: citizens band (CB) or VHF FM commercial broadcasting. The work presented illustrates a?relatively simple problem, which, however, is characterised by large combinatorial complexity, so direct (exhaustive) search becomes completely impractical. The process of minimisation is based on genetic algorithm (GA), which proves its usability for given problem. The final result is a?significant reduction of peak-to-peak level of given multi-tone signal, demonstrated by three real-life examples.

7

Towards fully decentralized multi-objective energy scheduling

Bremer Jörg, Lehnhoff Sebastian

Annals of Computer Science and Information Systems

|

2019

|

Vol. 18

193--201

EN

Future demand for managing a huge number of individually operating small and often volatile energy resources within the smart grid is preponderantly answered by involving decentralized orchestration methods for planning and scheduling. Many planning and scheduling problems are of a multi-objective nature. For the single-objective case - e.g. predictive scheduling with the goal of jointly resembling a wanted target schedule - fully decentralized algorithms with self-organizing agents exist. We extend this paradigm towards fully decentralized agent-based multi-objective scheduling for energy resources e.g. in virtual power plants for which special local constraint-handling techniques are needed. We integrate algorithmic elements from the well-known S-metric selection evolutionary multi-objective algorithm into a gossiping-based combinatorial optimization heuristic that works with agents for the single-objective case and derive a number of challenges that have to be solved for fully decentralized multi-objective optimization. We present a first solution approach based on the combinatorial optimization heuristics for agents and demonstrate viability and applicability in several simulation scenarios.

8

Deep learning hyper-parameter tuning for sentiment analysis in twitter based on evolutionary algorithms

Rodríguez-Barroso Nuria, Moya Antonio R., Fernández José A., Romero Elena, Martínez-Cámara Eugenio, Herrera Francisco

Annals of Computer Science and Information Systems

|

2019

|

Vol. 18

255–-264

EN

The state of the art in Sentiment Analysis is defined by deep learning methods, and currently the research efforts are focused on improving the encoding of underlying contextual information in a sequence of text. However, those neural networks with a higher representation capacity are increasingly more complex, which means that they have more hyper-parameters that have to be defined by hand. We argue that the setting of hyper-parameters may be defined as an optimisation task, we thus claim that evolutionary algorithms may be used to the optimisation of the hyper-parameters of a deep learning method. We propose the use of the evolutionary algorithm SHADE for the optimisation of the configuration of a deep learning model for the task of sentiment analysis in Twitter. We evaluate our proposal in a corpus of Spanish tweets, and the results show that the hyper-parameters found by the evolutionary algorithm enhance the performance of the deep learning method.

9

Flying ant colony optimization algorithm for combinatorial optimization

Fidanova S., Atanassov K.

Studia Informatica

|

2017

|

Vol. 38, nr 4

31--40

EN

In this paper is introduce "flying" ants in Ant Colony Optimization (ACO). In traditional ACO algorithms the ants construct their solution regarding one step forward. In proposed ACO algorithm, the ants make their decision, regarding more than one step forward, but they include only one new element in their solutions.

PL

Artykuł przedstawia "latające" mrówki w problemie optymalizacji algorytmów mrówkowych. W tradycyjnych podejściach dla algorytmów mrówkowych agenci (mrówki) budują swoje rozwiązanie w kolejnych krokach. W zaproponowanym podejściu optymalizacji algorytmu mrówkowego agenci podejmują decyzję na podstawie więcej niż jednego kroku, jednakże tylko jeden element wprowadzany jest do rozwiązania.

10

Zastosowanie rozmytej mapy kognitywnej w prognozowaniu efektywności pracy wypożyczalni rowerowych

Jastriebow A., Kubuś Ł., Poczęta K.

Informatyka, Automatyka, Pomiary w Gospodarce i Ochronie Środowiska

|

2017

|

T. 7, nr 4

70--73

PL

W pracy zaproponowano zastosowanie rozmytej mapy kognitywnej wraz z ewolucyjnymi algorytmami uczenia do modelowania systemu prognozowania efektywności pracy wypożyczalni rowerowych. Na podstawie danych historycznych zbudowano rozmytą mapę kognitywną, którą następnie zastosowano do prognozowania liczby rowerzystów i klientów wypożyczalni w trzech kolejnych dniach. Proces uczenia zrealizowano z zastosowaniem indywidualnego kierunkowego algorytmu ewolucyjnego IDEA oraz algorytmu genetycznego z kodowaniem zmiennoprzecinkowym RCGA. Analizę symulacyjną systemu prognozowania efektywności pracy wypożyczalni rowerowych przeprowadzono przy pomocy oprogramowania opracowanego w technologii JAVA.

EN

This paper proposes application of fuzzy cognitive map with evolutionary learning algorithms to model a system for prediction of effectiveness of bike sharing systems. Fuzzy cognitive map was constructed based on historical data and next used to forecast the number of cyclists and customers of bike sharing systems on three consecutive days. The learning process was realized with the use of Individually Directional Evolutionary Algorithm IDEA and Real-Coded Genetic Algorithm RCGA. Simulation analysis of the system for prediction of effectiveness of bike sharing systems was carried out with the use of software developed in JAVA.

11

The On-line Evolutionary Method for Soft Fault Diagnosis in Diode-transistor Circuits

Korzybski M., Ossowski M.

International Journal of Electronics and Telecommunications

|

2015

|

Vol. 61, No. 1

109-115

EN

The paper is devoted to diagnostic method enabling us to perform all the three levels of fault investigations - detection, localization and identification. It is designed for analog diode-transistor circuits, in which the circuit’s state is defined by the DC sources’ values causing elements operating points and the harmonic components with small amplitudes being calculated in accordance with small-signal circuit analysis rules. Geneexpression programming (GEP), differential evolution (DE) and genetic algorithms (GA) are a mathematical background of the proposed algorithms. Time consumed by diagnostic process rises rapidly with the increasing number of possible faulty circuit elements in case of using any of mentioned algorithms. The conncept of using two different circuit models with partly different elements allows us to decrease a number of possibly faulty elements in each circuit because some of possibly faulty elements are absent in one of two investigated circuits.

12

Using Genetic Algorithms for Optimizing Algorithmic Control System of Biomimetic Underwater Vehicle

Praczyk T.

Computational Methods in Science and Technology

|

2015

|

Vol. 21, No. 4

251--260

EN

Autonomous underwater vehicles are vehicles that are entirely or partly independent of human decisions. In order to obtain operational independence, the vehicles have to be equipped with a specialized control system. The main task of the system is to move the vehicle along a path with collision avoidance. Regardless of the logic embedded in the system, i.e. whether it works as a neural network, fuzzy, expert, or algorithmic system or even as a hybrid of all the mentioned solutions, it is always parameterized and values of the system parameters affect its effectiveness. The paper reports the experiments whose goal was to optimize an algorithmic control system of a biomimetic autonomous underwater vehicle. To this end, three different genetic algorithms were used, i.e. a canonical genetic algorithm, a steady state genetic algorithm and a eugenic algorithm.

13

Zastosowanie grupowego algorytmu genetycznego w celu poszukiwania rozwiązania zagadnienia ekonomicznego rozdziału obciążeń pomiędzy bloki elektroenergetyczne

Handzel Z., Gajer M.

Przegląd Elektrotechniczny

|

2014

|

R. 90, nr 11

208--211

PL

Tematyka artykułu dotyczy zastosowania nowej wersji algorytmu genetycznego, określanej mianem grupowego algorytmu genetycznego, w celu rozwiązania zagadnienia ekonomicznego rozdziału obciążeń pomiędzy cieplne bloki energetyczne występujące w systemie elektroenergetycznym. Główną zaletą grupowego algorytmu genetycznego jest fakt, że wartość funkcji dopasowania wyznaczana jest jednocześnie dla większej grupy osobników, co zapobiega przedwczesnej zbieżności tego rodzaju algorytmu do jednego z licznych optimów lokalnych.

EN

The topic of the paper is about implementation of a novel version of genetic algorithm, which is called a group-based genetic algorithm. We use this kind of algorithm in order to solve an economic dispatch problem among energetic blocks in the electrical energetic system. The main merit of the group-based genetic algorithm is that the fitness function is calculated simultaneously for a larger group of individuals which disables the premature convergence to some of the numerous local optima.

14

Advanced supervision of oil wells based on soft computing techniques

Camargo E., Aguilar J.

Journal of Artificial Intelligence and Soft Computing Research

|

2014

|

Vol. 4, No. 3

215--225

EN

In this work is presented a hybrid intelligent model of supervision based on Evolutionary Computation and Fuzzy Systems to improve the performance of the Oil Industry, which is used for Operational Diagnosis in petroleum wells based on the gas lift (GL) method. The model is composed by two parts: a Multilayer Fuzzy System to identify the operational scenarios in an oil well and a genetic algorithm to maximize the production of oil and minimize the flow of gas injection, based on the restrictions of the process and the operational cost of production. Additionally, the first layers of the Multilayer Fuzzy System have specific tasks: the detection of operational failures, and the identification of the rate of gas that the well requires for production. In this way, our hybrid intelligent model implements supervision and control tasks.

15

Zastosowanie techniki obliczeń ewolucyjnych na potrzeby sterowania procesem magazynowania energii elektrycznej w elektrowniach szczytowo-pompowych. Cz. 1, Analiza zagadnienia optymalizacyjnego

Handzel Z., Gajer M.

Elektronika : konstrukcje, technologie, zastosowania

|

2014

|

Vol. 55, nr 10

88-90

PL

Tematyka artykułu dotyczy zagadnienia ekonomicznego rozdziału obciążeń pomiędzy bloki elektroenergetyczne elektrowni cieplnych w sytuacji zainstalowania znacznego poziomu mocy w elektrowniach solarnych opartych na ogniwach fotowoltaicznych. Na potrzeby rozwiązania rozważanego zagadnienia optymalizacyjnego autorzy zaproponowali wykorzystanie techniki obliczeń ewolucyjnych.

EN

The topic of the paper is about finding the solution of economic dispatch problem of thermal units in the electrical energetic system in the case of presence of solar power plants with high values of installed power. In the paper we propose to use evolutionary computations technique to solve the above mentioned optimization problem.

16

Zastosowanie techniki obliczeń ewolucyjnych na potrzeby sterowania procesem magazynowania energii elektrycznej w elektrowniach szczytowo-pompowych. Cz. 2, Implementacja algorytmu ewolucyjnego

Handzel Z., Gajer M.

Elektronika : konstrukcje, technologie, zastosowania

|

2014

|

Vol. 55, nr 11

111-113

PL

Tematyka artykułu dotyczy zastosowania obliczeń ewolucyjnych w celu optymalizacji procesu magazynowania energii wyprodukowanej w elektrowniach solarnych. W artykule zaproponowano wykorzystanie binarnego sposobu kodowania rozwiązań na materiale genetycznym podlegających ewolucji osobników. Zdefiniowano także postać funkcji celu pozwalającej na efektywne przeprowadzenie oceny uzyskiwanych na drodze ewolucyjnej rozwiązań.

EN

The topic of the paper is about implementation of evolutionary computation technique for the purpose of optimization of energy storage which was generated in solar power plants. In the paper we propose the way of coding of solutions on genetic material of evolving individuals. We also define fitness function which allows us to compare the solutions that are obtained with the use of evolutionary methods.

17

Wyznaczanie wartości parametrów czterobiegunowego filtra Butterwortha z wykorzystaniem obliczeń ewolucyjnych

Handzel Z., Gajer M.

Elektronika : konstrukcje, technologie, zastosowania

|

2013

|

Vol. 54, nr 6

60-61

PL

W artykule rozważono wykorzystanie techniki obliczeń ewolucyjnych na potrzeby wyznaczania wartości pojemności występujących w filtrach aktywnych zrealizowanych w układzie Butterwortha. Rozważono filtr stanowiący kaskadowe połączeniu dwóch filtrów dwubiegunowych Butterwortha. Zastosowanie algorytmu ewolucyjnego pozwoliło na uzyskanie w tym wypadku czterobiegunowego filtra aktywnego o stromości charakterystyki w paśmie zaporowym 80 dB na każdą dekadę częstotliwości.

EN

In the paper we discuss using the technique of evolutionary computation for the purpose of determining the values of capacitors in Butterworth active filters. We consider the filter which is a cascade of two Butterworth filters with two poles each. Using the evolutionary algorithm allowed us to obtained the active filter with four poles which have the steepness of characteristic in the amount of 80dB on every decade of frequency.

18

Analiza filtrów Butterwortha wyższego rzędu metodą obliczeń ewolucyjnych

Handzel Z., Gajer M.

Elektronika : konstrukcje, technologie, zastosowania

|

2013

|

Vol. 54, nr 5

63-65

PL

Artykuł został poświęcony zagadnieniom projektowania filtrów Butterwortha wyższego rzędu, czyli takich, które stanowią kaskadowe połączenie pewnej liczby filtrów dwubiegunowych lub trójbiegunowych. Na potrzeby wyznaczania wartości rezystancji i pojemności występujących w rozważanych układach kaskadowych w artykule zaproponowano wykorzystanie techniki obliczeń ewolucyjnych. Użycie algorytmu ewolucyjnego pozwala na dokonanie optymalizacji charakterystyki amplitudowej filtru kaskadowego poprzez maksymalne zbliżenie jej kształtu do charakterystyki filtru idealnego.

EN

The paper was devoted to the issues of designing of higher order Butter-worth filters that constitute the cascade of filters with two or three poles. In order to calculate the resistance and capacity of passive elements of the filters we proposed to use the technique of evolutionary computation. Using evolutionary algorithm allows to optimize the amplitude characteristic of the filter by approaching the shape of characteristic of the ideal filter.

19

Evolutionary multi-agent systems in non-stationary environments

Kisiel-Dorohinicki M.

Computer Science

|

2013

|

Vol. 14 (4)

563--575

EN

In this article, the performance of an evolutionary multi-agent system in dynamic optimization is evaluated in comparison to classical evolutionary algorithms. The starting point is a general introduction describing the background, structure and behavior of EMAS against classical evolutionary techniques. Then, the properties of energy-based selection are investigated to show how they may influence the diversity of the population in EMAS. The considerations are illustrated by experimental results based on the dynamic version of the well-known, high-dimensional Rastrigin function benchmark.

20

Evolutionary multi-agent computing in inverse problems

Wróbel K., Torba P., Paszyński M, Byrski A.

Computer Science

|

2013

|

Vol. 14 (3)

367--383

EN

This paper tackles the application of evolutionary multi-agent computing to solve inverse problems. High costs of fitness function call become a major difficulty when approaching these problems with population-based heuristics. However, evolutionary agent-based systems (EMAS) turn out to reduce the fitness function calls, which makes them a possible weapon of choice against them. This paper recalls the basics of EMAS and describes the considered problem (Step and Flash Imprint Lithography), and later, shows convincing results that EMAS is more effective than a classical evolutionary algorithm.