SEARCH RESULT

This article attempts to solve the problem of efficient design of proportional+integral+derivative (PID) controller applied to automatic voltage regulator (AVR) by employing recently introduced symbiotic organism search (SOS) algorithm for the first time. SOS is a metaheuristic proved recently to be promising by benefitting from the idea of imitating natural phenomena of interactive behavior seen among organisms living together in a similar environment. PID controller design needs proper determination of three control parameters. Such a design problem can be taken as an optimization task and SOS is invoked to find out better controller parameters through the new cost function defined in the paper, which allows to evaluate the control behavior in both time-domain and frequency-domain. For the performance analysis, distinct analysis techniques are deployed such as transient response analysis, root locus analysis and bode analysis. The efficacy of the presented technique is widely illustrated by comparing the obtained results with those reported in some prestigious journals and it is shown that our proposal leads to a more satisfactory control performance from the perspective of both time-domain and frequency-domain specifications.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Emre Çelik

In this paper, a particle swarm optimization (PSO) algorithm is introduced to determine learning parameters required for the backpropagation (BP) learning algorithm, which is used for training of a feed-forward neural network (FFNN). PSO algorithm utilized within the paper works slightly different compared to conventional PSO (CPSO) algorithm in such a way that each particle adjusts its position based on the best midposition of all particles and its group’s previous best. The major reason of such a change is to enhance the performance of CPSO algorithm, which is explained in detail in the study suggested by Tamer, S and et.al. To test the proposed method, a FFNN with three layers is designed for function interpolation. Learning parameters of the designed NN are determined by both conventional error and trial method and the proposed method. Afterwards, using these two groups of learning parameters, the NN is trained and tested under the same conditions. According to the test results, learning parameters determined by the PSO provide a better performance and interpolating capability for the NN than those determined by the conventional method.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Emre Çelik Nihat Ozturk Adem Dalcalı