Adaptive Multi-scale Quantum Harmonic Oscillator Algorithm Based on Evolutionary Strategy

This paper proposes a novel adaptive multi-scale quantum harmonic oscillator algorithm based on evolutionary strategies (AMQHOA-ES) for global numerical optimization. Since the original Multi-scale Quantum Harmonic Oscillator Algorithm (MQHOA) utilizes a fixed contraction factor to narrow the search...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2020 IEEE Congress on Evolutionary Computation (CEC) S. 1 - 8
Hauptverfasser: Ye, Xinggui, Wang, Peng
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 01.07.2020
Schlagworte:
adaptive mechanism
Benchmark testing
Convergence
differential evolution
Evolutionary strategy
Harmonic analysis
multi-scale quantum harmonic oscillator algorithm
Oscillators
population-based optimization
Sociology
Statistics
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:This paper proposes a novel adaptive multi-scale quantum harmonic oscillator algorithm based on evolutionary strategies (AMQHOA-ES) for global numerical optimization. Since the original Multi-scale Quantum Harmonic Oscillator Algorithm (MQHOA) utilizes a fixed contraction factor to narrow the search scale, the searching step decreases too fast at the later stage of the evolution and is more likely to suffer premature convergence and stagnation. To improve the convergence performance, an adaptive attenuation mechanism of scaling is proposed to dynamically adjust the exploration and exploitation properties. Evolutionary strategies such as selection, crossover and DE/rand/1 mutation are implemented in the proposed algorithm to enhance the exploration and exploitation abilities. Experimental results evaluated on several unimodal and multimodal benchmark functions indicate the significant improvement of the proposed algorithm to the original MQHOA. Meanwhile, the experimental results compared with several state-of-the-art optimizers show the superiority or competitiveness of the proposed algorithm.
DOI:10.1109/CEC48606.2020.9185738