2018 International Conference on Advanced Science and Engineering (ICOASE) 2018
DOI: 10.1109/icoase.2018.8548816
|View full text |Cite
|
Sign up to set email alerts
|

Performance Evaluation of Parallel Particle Swarm Optimization for Multicore Environment

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2

Citation Types

0
8
0

Year Published

2019
2019
2023
2023

Publication Types

Select...
3
2
2

Relationship

1
6

Authors

Journals

citations
Cited by 9 publications
(9 citation statements)
references
References 8 publications
0
8
0
Order By: Relevance
“…Xicheng Fu et al [17] proposed a local PSOA based on medical image registration GPU. Their method has obvious advantages in the optimization of a high dimensional objective function, with the maximum acceleration ratio reaching 95 times.…”
Section: Traditional Gpu-based Psomentioning
confidence: 99%
See 1 more Smart Citation
“…Xicheng Fu et al [17] proposed a local PSOA based on medical image registration GPU. Their method has obvious advantages in the optimization of a high dimensional objective function, with the maximum acceleration ratio reaching 95 times.…”
Section: Traditional Gpu-based Psomentioning
confidence: 99%
“…So researchers consider that the parallelism of particles can be exploited to effectively accelerate the PSO.At present, there are three main strategies of PSO by parallel computation: 1) hardware environment-based parallel PSO algorithms, which are generally implemented through using hardware architectures [10][11] such as controllers [12] , Field Programmable Gate Array (FPGA) [13][14] , etc. to effectively improve the optimization speed of PSO algorithm, 2) CPUbased parallel PSO algorithms, which usually adopt multi-threading techniques [15][16] or multi-core processors [17] to express the independence of particles, making full use of parallelization to enhance the e ciency of PSO algorithm, 3) GPU-based parallel PSO algorithms, which exibly use the architecture of GPU to synchronize the parallel optimization process of particles. However, each strategy has its advantages and disadvantages like the parallel hardware environment are accessible to deploy and implement but depends on the cluster node scale.…”
mentioning
confidence: 99%
“…Through experiments with three optimization functions, they demonstrated that, compared to traditional CPU-PSO, their algorithm can achieve a computational speedup of up to 90 times while maintaining function convergence. Xicheng Fu et al [17] proposed a local PSOA based on medical image registration GPU. Their method has distinct advantages in optimizing high-dimensional objective functions, with the maximum acceleration ratio reaching 95 times.…”
Section: Traditional Gpu-based Psomentioning
confidence: 99%
“…Nevertheless, when encountering large-scale and highly complex cyclic optimization problems, PSO convergence speed decreases signi cantly as the number of particles increases, resulting in unsatisfactory optimization e ciency.Considering the interdependence and cooperation of particles in the PSO algorithm, researchers have explored parallelism to accelerate PSO effectively. There are three main parallel computation strategies for PSO: 1) hardware environmentbased parallel PSO algorithms, typically implemented using hardware architectures [10][11] such as controllers [12] and Field Programmable Gate Array (FPGA) [13][14] ; 2) CPU-based parallel PSO algorithms, utilizing multi-threading techniques [15][16] or multi-core processors [17] to express particle independence and exploit parallelism; 3) GPU-based parallel PSO algorithms, leveraging GPU architecture to synchronize parallel optimization processes for particles. Each strategy has its pros and cons, with parallel hardware environments being easy to deploy and implement but dependent on cluster node scale.…”
mentioning
confidence: 99%
“…Apart from scenarios in which the number of nodes in the network and the speed of movement changed, the rest of conditions were same [26], [27]. The following metrics are used to evaluate the performance of routing protocols [28]: -End-to-End delay: sum of delays from moment first transmitted byte untel the last byte reached to recivers.…”
mentioning
confidence: 99%