Wyniki wyszukiwania - Biblioteka Nauki

1

Zastosowanie przykładowego algorytmu stadnego w optymalizacji kombinatorycznej

100%

Filipowicz B. , Kwiecień J.

Automatyka / Akademia Górniczo-Hutnicza im. Stanisława Staszica w Krakowie

|

2010

|

tom T. 14, z. 3/2

885-891

PL

Algorytmy metaheurystyczne inspirowane naturą znajdują szerokie zastosowanie w problemach optymalizacji kombinatorycznej. Do tej klasy metod należy algorytm optymalizacji rojem cząstek oparty na zachowaniach stada ptaków. W artykule przedstawiono zastosowanie binarnego algorytmu optymalizacji rojem cząstek do rozwiązania wielowymiarowego problemu plecakowego. Zaprezentowano również wyniki eksperymentów dla wybranych instancji testowych.

EN

Nature-inspired metaheuristic algorithms are successfully applied to combinatorial optimization problems. They incorporate particle swarm optimization inspired by the behaviors of bird flocks. This paper presents the applying of binary particle swarm optimization to the multidimensional knapsack problem. The results of computational experiments for standard test problems have been also presented.

2

Minimalizacja błędów nadążania w wieloosiowych maszynach CNC poprzez zastosowanie sterowania predykcyjnego z neuronowym modelem obiektu regulacji

88%

Paprocki M. , Erwiński K. , Grzesiak L.

Przegląd Elektrotechniczny

|

2012

|

tom R. 88, nr 4b

169-174

PL

W układach sterowania maszyn numerycznych CNC istnieje możliwość zmniejszenia błędów odzwierciedlenia zadanej trajektorii ruchu oraz poprawienia dokładności wykonania elementów obrabianych poprzez wykorzystanie sterowania optymalnego. W proponowanym algorytmie sterowania użyto neuronowego modelu obiektu regulacji oraz sterowania predykcyjnego. Błędy obróbki są kompensowane poprzez modyfikację zadanej trajektorii ruchu.

EN

In CNC multi axis machine control systems it is possible to decrease motion trajectory errors and increase manufacturing precision by using optimal control. The proposed algorithm uses a neural network plant model and predictive control. Machining errors are compensated by modification of the reference trajectory.

3

Prediction of Standard Time of the Sewing Process using a Support Vector Machine with Particle Swarm Optimization

75%

Shao Y. , Ji X. , Zheng M. , Chen C.

Autex Research Journal

|

2022

|

tom Vol. 22, no. 3

290--297

EN

Standard time is a key indicator to measure the production efficiency of the sewing department, and it plays a vital role in the production forecast for the apparel industry. In this article, the grey correlation analysis was adopted to identify seven sources as the main influencing factors for determination of the standard time in the sewing process, which are sewing length, stitch density, bending stiffness, fabric weight, production quantity, drape coefficient, and length of service. A novel forecasting model based on support-vector machine (SVM) with particle swarm optimization (PSO) is then proposed to predict the standard time of the sewing process. On the ground of real data from a clothing company, the proposed forecasting model is verified by evaluating the performance with the squared correlation coefficient (R2) and mean square error (MSE). Using the PSO-SVM method, the R2 and MSE are found to be 0.917 and 0.0211, respectively. In conclusion, the high accuracy of the PSO-SVM method presented in this experiment states that the proposed model is a reliable forecasting tool for determination of standard time and can achieve good predicted results in the sewing process.

4

A modified particle swarm optimization procedure for triggering fuzzy flip-flop neural networks

75%

Kowalski P. A. , Słoczyński T.

International Journal of Applied Mathematics and Computer Science

|

2021

|

tom Vol. 31, no. 4

577--586

EN

The aim of the presented study is to investigate the application of an optimization algorithm based on swarm intelligence to the configuration of a fuzzy flip-flop neural network. Research on solving this problem consists of the following stages. The first one is to analyze the impact of the basic internal parameters of the neural network and the particle swarm optimization (PSO) algorithm. Subsequently, some modifications to the PSO algorithm are investigated. Approximations of trigonometric functions are then adopted as the main task to be performed by the neural network. As a result of the numerical verification of the problem, a set of rules are developed that can be helpful in constructing a fuzzy flip-flop type neural network. The obtained results of the computations significantly simplify the structure of the neural network in relation to similar conditions known from the literature.

5

Coordination of phase shifting transformers by means of the swarm algorithm

75%

Korab R. , Owczarek R. , Połomski M.

Prace Naukowe Politechniki Śląskiej. Elektryka

|

2017

|

tom z. 3-4

37--47

EN

The use of several phase shifting transformers (PSTs) in an interconnected power system must be coordinated in order to take full advantage of these devices and to avoid adverse interactions. This paper presents an optimization method of PST settings based on the particle swarm optimization (PSO) algorithm. The minimization of an unscheduled flow through a given system was used as the optimization criterion. Simulation results for an IEEE 118-bus test system are given.

PL

Zastosowanie kilku przesuwników fazowych w połączonym systemie elektroenergetycznym musi być skoordynowane w celu pełnego wykorzystania tych urządzeń i uniknięcia ich niekorzystnych interakcji. W artykule przedstawiono metodę optymalizacji nastaw przesuwników fazowych, opartą na algorytmie roju cząstek (PSO). Jako kryterium optymalizacji zastosowano minimalizację przepływu nieplanowego przez dany system. Pokazano wyniki obliczeń dla sieci testowej zawierającej 118 węzłów.

6

Optymalizacja wzmocnień obserwatora Luenbergera silnika indukcyjnego metodą roju cząstek

75%

Białoń T.

Przegląd Elektrotechniczny

|

2022

|

tom R. 98, nr 11

277--283

PL

W artykule opisano eksperymenty z doborem wzmocnień obserwatora zmiennych stanu silnika indukcyjnego przy wykorzystaniu metody optymalizacyjnej opartej na roju cząstek (PSO). W badaniach skupiono się na porównaniu różnych wersji algorytmu PSO, poddając optymalizacji funkcję celu o zawsze takich samych parametrach. Przeanalizowano trzy różne metody uczenia, GB (Global Best), LB (Local Best) oraz FIPS (Fully Informed Particle Swarm). Dwie ostatnie metody działają w oparciu o zadaną topologię roju, do wyboru spośród kraty pierścieniowej, kraty Von Neumanna oraz FDR (Fitness Distance Ratio). Przeanalizowano zagadnienia zbieżności i stabilności algorytmu, zależne od parametrów takich jak współczynnik uczenia.

EN

The paper describes experiments with the gain selection of an induction motor state observer, using particle swarm optimization (PSO) method. The research focused on comparing different versions of the PSO algorithm, optimizing the fitness function elaborated during preceding research. Three different learning methods were analyzed, GB (Global Best), LB (Local Best) and FIPS (Fully Informed Particle Swarm). The last two methods operate on the basis of a given swarm topology, to be selected from a ring lattice, Von Neumann lattice and FDR (Fitness Distance Ratio). The problems of convergence and stability of the algorithm, depending on parameters such as a cognition factor, were analyzed. The results for 1000 runs of PSO with 20 different sets of parameters were presented and compared.

7

Particle swarm intelligence based optimisation of high speed end-milling

75%

Cus F. , Zuperl U.

Archives of Computational Materials Science and Surface Engineering

|

2009

|

tom Vol. 1, nr 3

148-154

EN

Purpose: Selection of machining parameters is an important step in process planning therefore a new evolutionary computation technique is developed to optimize machining process. This study has presented multi-objective optimization of milling process by using neural network modelling and Particle swarm optimization. Particle Swarm Optimization (PSO) is used to efficiently optimize machining parameters simultaneously in high-speed milling processes where multiple conflicting objectives are present. The goal of optimization is to determine the objective function maximum (predicted cutting force surface) by consideration of cutting constraints. Design/methodology/approach: First, an Artificial Neural Network (ANN) predictive model is used to predict cutting forces during machining and then PSO algorithm is used to obtain optimum cutting speed and feed rates. Findings: During optimization the particles 'fly' intelligently in the solution space and search for optimal cutting conditions according to the strategies of the PSO algorithm. The simulation results show that compared with genetic algorithms (GA) and simulated annealing (SA), the proposed algorithm can improve the quality of the solution while speeding up the convergence process. Research limitations/implications: The experimental results show that the MRR is improved by 28%. Machining time reductions of up to 20% are observed. Practical implications: While a lot of evolutionary computation techniques have been developed for combinatorial optimization problems, PSO has been basically developed for continuous optimization problem. PSO can be an efficient optimization tool for solving nonlinear continuous optimization problems, combinatorial optimization problems, and mixed-integer nonlinear optimization problem. Originality/value: An algorithm for PSO is developed and used to robustly and efficiently find the optimum machining conditions in end-milling. This paper opens the door for a new class of EC based optimization techniques in the area of machining. This paper also presents fundamentals of PSO optimization techniques.

8

Analiza skuteczności algorytmów PSO w szacowaniu jakości prognoz godzinowego zapotrzebowania na energię elektryczną wybranych grup odbiorców

75%

Baczyński D. , Piotrowski P.

Przegląd Elektrotechniczny

|

2012

|

tom R. 88, nr 11a

25-29

PL

W tekście przedstawiono opis opracowanej metody szacowania jakości prognoz godzinowego zapotrzebowania na energię elektryczną wybranych grup odbiorców. Metoda pozwala oszacować poziom błędu prognoz dla różnych horyzontów prognoz (1, 2, 7, 14 oraz 21 dni naprzód) na podstawie informacji o danych statystycznych odbiorców. Optymalizacja parametrów metody szacowania wykonana została kilkoma metodami, w tym przy wykorzystaniu algorytmu optymalizacji rojem cząstek (PSO). Wykonano analizę porównawczą uzyskanych wyników. Sformułowano wnioski końcowe.

EN

The paper presents description of worked out method of estimating hourly demand electric energy forecasts quality for chosen consumer groups. The method allows to estimate the error level for various horizons of forecasts (1, 2, 7, 14 and 21 days ahead) based on information about consumers statistical data. Parameters optimization was performed using different methods including particle swarm optimization algorithm (PSO). The final conclusions have been presented.

9

Scheduling electric power generators using particle swarm optimization combined with the Lagrangian relaxation method

75%

Balci H. H. , Valenzuela J. F.

International Journal of Applied Mathematics and Computer Science

|

2004

|

tom Vol. 14, no 3

411-421

EN

This paper describes a procedure that uses particle swarm optimization (PSO) combined with the Lagrangian Relaxation (LR) framework to solve a power-generator scheduling problem known as the unit commitment problem (UCP). The UCP consists of determining the schedule and production amount of generating units within a power system subject to operating constraints. The LR framework is applied to relax coupling constraints of the optimization problem. Thus, the UCP is separated into independent optimization functions for each generating unit. Each of these sub-problems is solved using Dynamic Programming (DP). PSO is used to evolve the Lagrangian multipliers. PSO is a population based search technique, which belongs to the swarm intelligence paradigm that is motivated by the simulation of social behavior to manipulate individuals towards better solution areas. The performance of the PSO-LR procedure is compared with results of other algorithms in the literature used to solve the UCP. The comparison shows that the PSO-LR approach is efficient in terms of computational time while providing good solutions.

10

Optimization of Neural Network Model Design: An Electoral Cooperative Particle Swarm Optimization Approach

75%

Li D. , Xu H.

Przegląd Elektrotechniczny

|

2013

|

tom R. 89, nr 1b

165--167

EN

This paper proposes an electoral cooperative particle swarm optimization approach to optimize the model of neural network from both structure and linked weights. Different with other related research work, a new encoding method is adopted to divide the neural network into several modules, each of them corresponding to a sub-swarm. Based on the experiments on typical problems and classic dataset, the results show that the proposed algorithm outperforms all the compared ones in perspective of test error, correctness, connection number, and the CPU time of the training phase.

PL

W przedstawionym artykule opisano zastosowanie metod optymalizacji roju cząstek do optymalizacji struktury i współczynników wagowych sieci neuronowej. Zaimplementowano nową metodę analizy, do dzielenia podzielenia sieci na moduły, reprezentujące mniejsze roje. Weryfikacja eksperymentalna i porównanie z metodami klasycznymi wykazały wysoką sprawność i skuteczność analizy.

11

Adaptive protection scheme for optimally coordinated relay setting using modified PSO algorithm

75%

Atteya A. I. , El Zonkoly A. M. , Ashour H. A.

Journal of Power Technologies

|

2017

|

tom Vol. 97, nr 5

463--469

EN

With growing power demand and heightened concern about the use of fossil fuels in conventional power plants, the integration of distributed energy resources into power networks is gaining attention due to their ability to cater for localized energy needs, putting the concept of the Smart grid center stage. Network protection systems, faced with a gradual increase in complexity, will have to develop responses to the changes brought about by ever greater penetration by distributed generation and sophisticated network topologies. The main goal of this paper is to provide optimal relay coordination of an adaptive protection scheme. Designed software based on a Modified Particle Swarm Optimization (MPSO) algorithm is implemented to solve the relay coordination problem. In this study, the 14 IEEE bus system is tested across a range of power system scenarios to validate the suggested technique. The results obtained show that optimal relay settings are achieved by the proposed algorithm regardless of the prevailing network topology.

12

Unsupervised classification and particle swarm optimization

75%

Truszkowski A. , Topczewska M.

Zeszyty Naukowe Politechniki Białostockiej. Informatyka

|

2012

|

tom Z. 9

119-132

EN

This article considers three algorithms of unsupervised classification -K-means, Gbest and the Hybrid method, the last two have been proposed in [14]. All three algorithms belong to the class of non-hierarchical methods. At first, the initial split of objects into known in advance number of classes is performed. If it is necessary, some objects are then moved into other clusters to achieve better split - between cluster variation should be much larger than within cluster variation. The first algorithm described in this paper (K-means) is wellknown classical method. The second one (Gbest) is based on the particle swarm intelligence idea. While the third is a hybrid of two mentioned algorithms. Several indices assessing the quality of obtained clusters are calculated.

PL

W niniejszym artykule porównywane są trzy algorytmy analizy skupień - metoda k-średnich, algorytm gbest oraz metoda hybrydowa. Algorytmy gbest oraz hybrydowy zostały zaproponowane w publikacji [14]. Wszystkie trzy metody nalezą a do rodziny metod niehierarchicznych, w których na początku tworzony jest podział obiektów na znaną z góry liczbę klastrów. Następnie, niektóre obiekty przenoszone są pomiędzy klastrami, by uzyskać jak najlepszy podział - wariancja pomiędzy skupieniami powinna być znacznie większa niż wariancja wewnątrz skupień. Pierwszy algorytm (k-means) jest znaną, klasyczną metodą. Drugi oparty jest na idei inteligencji roju cząstek. Natomiast trzeci jest metodą hybrydową łączącą dwa wymienione wcześniej algorytmy. Do porównania uzyskanych skupień wykorzystano kilka różnych indeksów szacujących jakość otrzymanych skupień.

13

Application of the PSO Algorithm with Sub-domain Approach for the Optimization of a Radio Telescope Array

75%

Borowska B. , Nadolski S.

Journal of Applied Computer Science

|

2008

|

tom Vol. 16, nr 1

7-14

EN

In this research a modified PSO (Particle Swarm Optimization) algorithm was applied to solve an optimization problem of a radio telescope array distribution. The objective of the proposed algorithm was to determine locations of the telescopes on a circle so that the entire array gained its maximal resolving power. A modification of the algorithm relied on the division of the domain into sub-domains. Furthermore, while the initial population was generated the particles were divided into groups. In the initial iterations, each group was moving only inside a subdomain previously assigned to this group thus making the distribution of the whole swarm more homogenous. This paper presents an evaluation of the efficiency of the proposed algorithm.

14

An electoral quantum-behaved PSO with Lévy flights for permutation flow shop scheduling problem

75%

Li D. , Deng N.

Przegląd Elektrotechniczny

|

2013

|

tom R. 89, nr 1b

138--142

EN

Permutation flow shop scheduling problem (PFSSP), a NP-hard combinatorial optimization problem, has strong engineering background of finding the optimal processing sequence and time of jobs on machines under the constraints of resources. Recently, several approaches based on Particle Swarm Optimization (PSO) have been developed to solve the PFSSP, and the experimental results show that they are efficient. To solve this issue, a novel variant of quantum-behaved particle swarm optimization algorithm for permutation flow shop scheduling is proposed in this paper. This algorithm is a combination of quantum-behaved PSO, electoral mechanism, and a disturbance generated by Lévy flights. Inspired by the election behavior in society, an electoral and cooperative mechanism is imported to get the elite particles from the primitive sub-swarms respectively. Moreover, the character unequal hop length of Lévy flights provides a method to escape the local optima efficiently. The numerical results on the Taillard's benchmark also show it outperforms other related algorithms.

PL

Problem szeregowania zmiany przepływów magazynowych (PFSSP) jest silnie nie–wielomianowym (NP) problemem optymalizacji kombinatorycznej. Ma ważny inżynierski aspekt w wyznaczaniu optymalnej kolejności procesu i czasu pracy maszyn, wymuszonej zmianą zasobów. Ostatnio, do rozwiązania PFSSP, zastosowano szereg przybliżeń opartych o algorytm optymalizacji rojem cząstek (PSO) a wyniki praktyczne pokazują, że są to rozwiązania efektywne. W prezentowanym opracowaniu, do szeregowania przepływów magazynowych, zaproponowano nowy wariant algorytmu optymalizacji rojem cząstek z zachowaniem kwantowym (QPSO). Algorytm jest kombinacją QPSO, mechanizmu wyborczego i zakłóceń generowanych rozkładem lotów Levy’ego. Do wyłonienia cząstek elitarnych z prymitywnego pod-roju wykorzystano, inspirowany zachowaniami wyborczymi w społeczeństwie, mechanizm wyborczy i współpracy. Ponadto, unikalny charakter długości skoków lotów Levy’ego pozwala skutecznie uniknąć optimów lokalnych. Wyniki numeryczne, przeprowadzone na danych testowych Taillard’a, także wskazują na przewagę nad innymi porównywalnymi algorytmami.

15

Zastosowanie algorytmu optymalizacji rojem cząstek do znajdowania ekstremów globalnych wybranych funkcji testowych

75%

Figielska E. , Wiatrak M.

Zeszyty Naukowe Warszawskiej Wyższej Szkoły Informatyki

|

2015

|

tom nr 13

7--19

PL

Praca dotyczy zastosowania algorytmu optymalizacji rojem cząstek do znajdowania ekstremów globalnych dla wybranych funkcji jedno i wielomodalnych. Na podstawie wyników eksperymentu obliczeniowego wyłoniono warianty ustawień parametrów algorytmu zapewniające jego największą skuteczność.

EN

In this paper, we present the particle swarm optimization algorithm for finding the global extrema of several single and multimodal functions. The values of the algorithm parameters which ensure its best performance are determined on the basis of the computational results.

16

High speed end-milling optimisation using Particle Swarm Intelligence

75%

Cus F. , Zuperl U. , Gecevska V.

Journal of Achievements in Materials and Manufacturing Engineering

|

2007

|

tom Vol. 22, nr 2

75-78

EN

Purpose: In this paper, Particle Swarm Optimization (PSO), which is a recently developed evolutionary technique, is used to efficiently optimize machining parameters simultaneously in high-speed milling processes where multiple conflicting objectives are present. Design/methodology/approach: selection of machining parameters is an important step in process planning therefore a new methodology based on PSO is developed to optimize machining conditions. Artificial neural network simulation model (ANN) for milling operation is established with respect to maximum production rate, subject to a set of practical machining constraints. An ANN predictive model is used to predict cutting forces during machining and PSO algorithm is used to obtain optimum cutting speed and feed rate. Findings: The simulation results show that compared with genetic algorithms (GA) and simulated annealing (SA), the proposed algorithm can improve the quality of the solution while speeding up the convergence process. PSO is proved to be an efficient optimization algorithm. Research limitations/implications: Machining time reductions of up to 30% are observed. In addition, the new technique is found to be efficient and robust. Practical implications: The results showed that integrated system of neural networks and swarm intelligence is an effective method for solving multi-objective optimization problems. The high accuracy of results within a wide range of machining parameters indicates that the system can be practically applied in industry. Originality/value: An algorithm for PSO is developed and used to robustly and efficiently find the optimum machining conditions in end-milling. The new computational technique has several advantages and benefits is suitable for use combined with ANN based models where no explicit relation between imputs and outputs is available. This research opens the door for a new class of optimization techniques which are based on Evolution Computation in the area of machining.

17

Feature selection using particle swarm optimization in text categorization

63%

Aghdam M. H. , Heidari S.

Journal of Artificial Intelligence and Soft Computing Research

|

2015

|

tom Vol. 5, No. 4

231--238

EN

Feature selection is the main step in classification systems, a procedure that selects a subset from original features. Feature selection is one of major challenges in text categorization. The high dimensionality of feature space increases the complexity of text categorization process, because it plays a key role in this process. This paper presents a novel feature selection method based on particle swarm optimization to improve the performance of text categorization. Particle swarm optimization inspired by social behavior of fish schooling or bird flocking. The complexity of the proposed method is very low due to application of a simple classifier. The performance of the proposed method is compared with performance of other methods on the Reuters-21578 data set. Experimental results display the superiority of the proposed method.

18

Estimating Parameters in Production Function based on Quadratic Synchronous Artificial Fish- Particle Swarm Optimization

63%

Mo Y. , Zhao X. , Xiang S.

Przegląd Elektrotechniczny

|

2013

|

tom R. 89, nr 3b

104--106

EN

Investment analysis is an important element of the process of economic activities, and the rational analysis and selection of production function as well as the parameter estimation, are important part of the investment analysis. Common production functions are characterized by too strong non-linearity for the use of traditional method to estimate parameters; therefore a fast, simple and robust algorithm becomes a hot research interest to optimize the production function. To this end this paper presents QSAFPSO for solving this problem. The algorithm enhances the genelevel exchange of information between individuals, creates a genetic template, employs genetic template evolution, mutation and other operations to improve the convergence speed, solution accuracy, and better helps algorithm out of local optimum. The typical function tests show that QSAFPSO, compared with like algorithms, features fast convergence and higher solution precision. A simulation based on the annual output value from 1820 to 1926, capital investment and labor input in Massachusetts shows that the algorithm is characterized by fast optimization of production function parameter estimation and by small residual sum of squares.

PL

W artykule opisano opracowany algorytm decyzyjny QSA-FPSO służący do analizy inwestycji. Jego działanie opiera się na budowie szablonu genetycznego i między genowej wymianie informacji. Wykonane testy funkcjonalności algorytmu pokazują jego większą, w porównaniu z innymi algorytmami, precyzję i szybkość osiągnięcia rozwiązania. Przedstawiono także wyniki badań symulacyjnych, dokonanych na rzeczywistych danych, potwierdzające wysoką skuteczność działania.

19

Intelligent financial time series forecasting: A complex neuro-fuzzy approach with multi-swarm intelligence

63%

Li C. , Chiang T. W.

International Journal of Applied Mathematics and Computer Science

|

2012

|

tom Vol. 22, no. 4

787-800

EN

Financial investors often face an urgent need to predict the future. Accurate forecasting may allow investors to be aware of changes in financial markets in the future, so that they can reduce the risk of investment. In this paper, we present an intelligent computing paradigm, called the Complex Neuro-Fuzzy System (CNFS), applied to the problem of financial time series forecasting. The CNFS is an adaptive system, which is designed using Complex Fuzzy Sets (CFSs) whose membership functions are complex-valued and characterized within the unit disc of the complex plane. The application of CFSs to the CNFS can augment the adaptive capability of nonlinear functional mapping, which is valuable for nonlinear forecasting. Moreover, to optimize the CNFS for accurate forecasting, we devised a new hybrid learning method, called the HMSPSO-RLSE, which integrates in a hybrid way the so-called Hierarchical Multi-Swarm PSO (HMSPSO) and the well known Recursive Least Squares Estimator (RLSE). Three examples of financial time series are used to test the proposed approach, whose experimental results outperform those of other methods.

20

Optymalizacja rojowa nastaw przesuwników fazowych w systemie elektroenergetycznym

63%

Korab R. , Owczarek R. , Połomski M.

Zeszyty Naukowe Wydziału Elektrotechniki i Automatyki Politechniki Gdańskiej

|

2017

|

tom Nr 53

99--102

PL

W artykule przedstawiono zastosowanie algorytmu rojowego PSO do optymalizacji nastaw grupy przesuwników fazowych w systemie elektroenergetycznym. Jako kryterium optymalizacji zastosowano minimalizację strat mocy czynnej w sieci testowej IEEE 118. Przeanalizowano wpływ maksymalnej dozwolonej prędkości cząstek na efektywność algorytmu optymalizacji. Wyniki badań pokazują ważność tego parametru.

EN

In response to the growing problem of unscheduled flows, a larger and larger number of transmission system operators in Europe equip their systems with phase shifting transformers (PSTs). PSTs are special transformers which installed in a transmission line enable regulation of the voltage phase angle and thereby change of the active power flow in the line. However, the use of several PSTs installed geographically close to each other must be coordinated in order to efficiently use those devices and avoid their adverse interactions. The coordination of a group of such devices leads to a multidimensional optimization problem. In this paper, the coordination problem was solved by optimization of settings of all analyzed PSTs, based on the swarm algorithm. This approach was examined and tested on an IEEE 118-bus test system. The minimization of active power losses in this system was used as the optimization criterion. The impact of maximum allowed velocity of particles on the effectiveness of the optimization algorithm was analyzed. The result shows that the improved effectiveness of the proposed approach can be obtained by careful selection of this parameter.