Abstract
본 논문에서는 종래의 PSO 알고리즘 성능저하의 주요 원인들 중 하나인 입자들의 조기수렴 현상을 개선한 DPSO-QI (Distributed PSO with quantum-infusion mechanism) 기법을 제안한다. DPSO-QI 알고리즘은 다음과 같은 두 가지 특징을 지닌다. 첫째, 분산형 구조의 PSO 기법을 도입한다. 이는 먼저 적절한 수의 입자들로 소그룹을 형성하고, 최적해 탐색에 필요한 다양한 정보의 교환이 각 소그룹 내에서만 이루어지도록 한 기법이다. 이러한 기법을 바탕으로 입자들의 탐색 다양성을 증대시킴으로서 조기수렴 현상을 감소시키는 효과를 달성할 수 있다. 둘째, 상기의 입자 소그룹에 Quantum-infusion (QI) 메커니즘에 기반 한 기법을 도입시킨다. 이를 통해 입자들의 전역 최적해 탐색 정밀도를 보다 향상시킬 수 있다. 끝으로 다양한 수치예제를 통하여 제안하는 새로운 PSO 기법이 종래의 방식들에 비해 매우 뛰어난 성능을 구현할 수 있음을 입증하고자 한다. In this paper, a novel DPSO-QI (Distributed PSO with quantum-infusion mechanism) algorithm improving one of the fatal defect, the so-called premature convergence, that degrades the performance of the conventional PSO algorithms is proposed. The proposed scheme has the following two distinguished features. First, a concept of neighborhood of each particle is introduced, which divides the whole swarm into several small groups with an appropriate size. Such a strategy restricts the information exchange between particles to be done only in each small group. It thus results in the improvement of particles' diversity and further minimization of a probability of occurring the premature convergence phenomena. Second, a quantum-infusion (QI) mechanism based on the quantum mechanics is introduced to generate a meaningful offspring in each small group. This offspring in our PSO mechanism improves the ability to explore a wider area precisely compared to the conventional one, so that the degree of precision of the algorithm is improved. Finally, some numerical results are compared with those of the conventional researches, which clearly demonstrates the effectiveness and reliability of the proposed DPSO-QI algorithm.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
More From: Journal of Korean Institute of Intelligent Systems
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.