Alexandria Engineering Journal (Aug 2020)
Adoption of fuzzy fractional order theory under Reco-Mo algorithm in data analysis of optical circuit switches
Abstract
In order to analyze the optimization value of Reco-Mo based on fuzzy fractional order theory for data analysis of optical circuit switches, first, the co-flow scheduling strategy is expounded. According to the Burkhofvon-Neumann (BvN) matrix decomposition algorithm, the co-flow scheduling algorithm is described. Then, the fuzzy fractional order theory is combined with the co-flow scheduling algorithm, and the Reco-Mo algorithm is proposed. Finally, E-FWO algorithm and E-SCFE algorithm are introduced to carry out comparative analysis in real production environment. The results show that when the number of co-flows is 30, 50, 70, 90 and 110, the co-flow completion time of Reco-Mo algorithm is less than that of E-FWO algorithm and E-SCFE algorithm. At 70, 90, and 110, the accuracy of Reco-Mo algorithm is significantly higher than that of E-FWO and E-SCFE algorithms. When the density of Reco-Mo algorithm is 0.1, the completion time of co-flow is 38.46% and 37.25% lower than that of E-FWO algorithm and E-SCFE algorithm. At 1.0 and 1.3, the accuracy of Reco-Mo algorithm is significantly higher than that of E-FWO and E-SCFE. The completion time of Reco-Mo algorithm under different data transmission volumes (0.5, 1.0, 1.5, 2.0, 2.5 GB) is less than that of E-FWO and E-SCFE. At 1.5, 2.0, and 2.5 GB, the accuracy of Reco-Mo algorithm is significantly higher than that of E-FWO and E-SCFE. When the co-flow arrival time interval is 100 ms, the Reco-Mo algorithm reduces the completion time by 50.34% and 51.97% respectively compared with E-FWO and E-SCFE algorithms. At 100, 200, and 300 ms, the accuracy of Reco-Mo algorithm is significantly higher than that of E-FWO and E-SCFE. In conclusion, compared with E-FWO and E-SCFE algorithms, Reco-Mo algorithm has always maintained its dominant position in terms of co-flow completion time and accuracy under different parameters of co-flow quantity, co-flow density, co-flow transmission data volume, and co-flow arrival interval, which provides a theoretical basis for the application of mathematical algorithm in data processing of optical circuit switches in data centers in the future.
