Efficient discrete firefly algorithm for Ctrie based caching of multiple sequence alignment on optimally scheduled parallel machines
| Autor: | Soniya Lalwani, Harish Sharma, Abhay Verma, Rajesh Kumar |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2019 |
| Předmět: |
query processing
parallel machines computational complexity minimisation cache storage swarm intelligence bioinformatics search problems particle swarm optimisation scheduling statistical testing makespan minimisation swarm-intelligence based implementation Ctrie based caching MSA DFFA efficient discrete firefly algorithm multiple sequence alignment optimally scheduled parallel machines two-level strategy complex heterogeneous sequences pairwise alignment multiple queries multiclient problem parallel connected machines BAliBASE 4 dataset MUSCLE dataset statistical significance testing one-way ANOVA Bonferroni posthoc analysis Computational linguistics. Natural language processing P98-98.5 Computer software QA76.75-76.765 |
| Zdroj: | CAAI Transactions on Intelligence Technology (2019) |
| Druh dokumentu: | article |
| ISSN: | 2468-2322 |
| DOI: | 10.1049/trit.2018.1040 |
| Popis: | This study introduces a two-level strategy for efficient execution of multiple sequence alignment (MSA) of complex heterogeneous sequences. The two levels of the proposed technique are comprised of: designing the discrete firefly algorithm (DFFA) for the formation and implementation of makespan minimisation on parallel machines, followed by performing Ctrie-based caching for pairwise alignment to reduce the load on the data servers for handling multiple queries. The proposed strategy addresses a multi-client problem that aims to acquire the full advantage of the computational power of parallel connected machines. Further, it is shown that the inclusion of Ctrie as caching mechanism successively improves the performance of the system with accretion in several sequences. Performance of proposed DFFA is also compared with discrete versions of four swarm intelligence based algorithms at the criteria of makespan minimisation and the rate of convergence on two kinds of complex and diverse datasets. The work is unique in this sense: it is the first swarm-intelligence-based implementation for the addressed problem; it is so far the first approach for Ctrie based caching of the MSA on the scheduled parallel machines; hybridisation of DFFA with Ctrie for caching the MSA results is also a novel implementation. |
| Databáze: | Directory of Open Access Journals |
| Externí odkaz: |
|
Plný text