adaptive penalty method
Recently Published Documents


TOTAL DOCUMENTS

19
(FIVE YEARS 8)

H-INDEX

4
(FIVE YEARS 1)

2020 ◽  
Vol 18 (2) ◽  
pp. 14
Author(s):  
Afonso Celso de Castro Lemonge ◽  
Patrícia Habib Hallak ◽  
José Pedro Gonçalves Carvalho

Este artigo apresenta um Algoritmo Evolucionário (AE) baseado no comportamento de enxame de partículas (Particle Swarm Optimization - PSO) adaptado para a obtenção de soluções de problemas de otimização estrutural com restrições. O PSO é um algoritmo de fácil implementação e competitivo perante os demais algoritmos populacionais inspirados na natureza. Neste artigo, são analisados problemas de otimização estrutural de treliças submetidas a restrições de frequências naturais de vibração. Para o tratamento destas restrições, incorpora-se ao PSO uma técnica de penalização adaptativa (Adaptive Penalty Method - APM), que tem demonstrado robustez e eficiência quando aplicada no tratamento de problemas de otimização com restrições. O algoritmo proposto é validado através de experimentos computacionais em problemas de otimização estrutural amplamente discutidos na literatura.


2018 ◽  
Vol 1 (1) ◽  
pp. 15-26
Author(s):  
D B Fatemeh ◽  
C K Loo ◽  
G Kanagaraj ◽  
S G Ponnambalam

Most real-life optimization problems involve constraints which require a specialized mechanism to deal with them. The presence of constraints imposes additional challenges to the researchers motivated towards the development of new algorithm with efficient constraint handling mechanism. This paper attempts the suitability of newly developed hybrid algorithm, Shuffled Complex Evolution with Quantum Particle Swarm Optimization abbreviated as SP-QPSO, extended specifically designed for solving constrained optimization problems. The incorporation of adaptive penalty method guides the solutions to the feasible regions of the search space by computing the violation of each one. Further, the algorithm’s performance is improved by Centroidal Voronoi Tessellations method of point initialization promise to visit the entire search space. The effectiveness and the performance of SP-QPSO are examined by solving a broad set of ten benchmark functions and four engineering case study problems taken from the literature. The experimental results show that the hybrid version of SP-QPSO algorithm is not only overcome the shortcomings of the original algorithms but also outperformed most state-of-the-art algorithms, in terms of searching efficiency and computational time.


2018 ◽  
Vol 20 (1) ◽  
pp. 65-88 ◽  
Author(s):  
Dênis E. C. Vargas ◽  
Afonso C. C. Lemonge ◽  
Helio J. C. Barbosa ◽  
Heder S. Bernardino

Sign in / Sign up

Export Citation Format

Share Document