Heterogeneous Parallel Island Models

Silveira, Lucas Ângelo; Soncco-Álvarez, José Luis; Lima, Thaynara Arielly de; Ayala-Rincón, Maurício

doi:10.1109/ssci50451.2021.9659938

Cited by 4 publications

(12 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We select a static and a dynamic topology that successfully addressed URD in [10] for homogeneous PIMs and in [1] for non reconfigurable heterogeneous PIMs. We choose asynchronous models since the dynamic asynchronous HePIM was the one that provided the best results.…”

Section: Communication Topologiesmentioning

confidence: 99%

“…When all islands run the same BA, such models are called homogeneous PIMs (HoPIMs). This work improves heterogeneous PIMs (HePIMs) [1], in which islands may run different BAs, allowing algorithmic reconfiguration on their islands, i.e., islands may dynamically update their BAs. In addition to an adequate and well-calibrated migration policy required by HePIMs, reconfigurable HePIMs exchange information between islands to decide how islands should reconfigure their BAs.…”

Section: Introductionmentioning

confidence: 99%

“…Silveira et al [1] introduced reconfigurable HePIMs running four different BAs in their islands, namely, Genetic Algorithm (GA), double-point crossover GA (GAD), Differential Evolution (DE), and self-adjusting Particle Swarm Optimization (PSO) (see e.g., [2], [3], and [4]). PIMs performance depends on a good calibration of the breeding cycle parameters (related to the involved BA) and, in addition to vital aspects of the parallel island architecture as island communication synchronism, island migration frequency, communication topology, and migration policy.…”

Section: Introductionmentioning

confidence: 99%

“…PIMs performance depends on a good calibration of the breeding cycle parameters (related to the involved BA) and, in addition to vital aspects of the parallel island architecture as island communication synchronism, island migration frequency, communication topology, and migration policy. We select two successful asynchronous HePIMs from [1] maintaining their parameters and adding the reconfiguration frequency.…”

Section: Introductionmentioning

confidence: 99%

“…Main contributions. We design new reconfigurable HeP-IMs that run GA, GAD, DE, and PSO in their islands using two successful asynchronous topologies from [1]. Nonreconfigurable HePIMs computed competitive solutions regarding most HoPIMs.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Reconfigurable Heterogeneous Parallel Island Models

Silveira¹,

Lima²,

Ayala-Rincón³

2022

Preprint

View full text Add to dashboard Cite

Heterogeneous Parallel Island Models (HePIMs) run different bio-inspired algorithms (BAs) in their islands. From a variety of communication topologies and migration policies finetuned for homogeneous PIMs (HoPIMs), which run the same BA in all their islands, previous work introduced HePIMs that provided competitive quality solutions regarding the best-adapted BA in HoPIMs. This work goes a step forward, maintaining the population diversity provided by HePIMs, and increasing their flexibility, allowing BA reconfiguration on islands during execution: according to their performance, islands may substitute their BAs dynamically during the evolutionary process. Experiments with the introduced architectures (RecHePIMs) were applied to the NP-hard problem of sorting permutations by reversals, using four different BAs, namely, simple Genetic Algorithm (GA), Double-point crossover Genetic Algorithm (GAD), Differential Evolution (DE), and self-adjusting Particle Swarm Optimization (PSO). The results showed that the new reconfigurable heterogeneous models compute better quality solutions than the HePIMs closing the gap with the HoPIM running the best-adapted BA.

show abstract

Section: Communication Topologiesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%