Table 4 Comparison of DFBA performance when different prediction approaches are used
Scenario conf. and sample set size | 4/12, 2 MHz, 10 % | 4/8, 5 MHz, 5 % | ||||||
|---|---|---|---|---|---|---|---|---|
Transport protocol | TCP | UDP | TCP | UDP | ||||
MAC scheduler | PF | RR | PF | RR | PF | RR | PF | RR |
SINR | 85 % | 83 % | 83 % | 86 % | 53 % | 54 % | 53 % | 42 % |
Min SINR | 91 % | 82 % | 89 % | 89 % | 71 % | 77 % | 58 % | 49 % |
Sum SINR/MAC THR Mapping | 72 % | 72 % | 70 % | 81 % | 61 % | 61 % | 35 % | 43 % |
Min SINR/MAC THR Mapping | 89 % | 81 % | 88 % | 89 % | 55 % | 64 % | 55 % | 50 % |
Learning based (TBAG) | 100 % | 85 % | 100 % | 95 % | 96 % | 95 % | 97 % | 92 % |
Exhaustive search | 100 % | 100 % | 100 % | 100 % | 100 % | 100 % | 100 % | 100 % |
Exhaustive search (Mbps) | 9 | 8 | 9 | 7 | 12 | 10 | 26 | 25 |