Dynamic computing resource provisioning on WattTime May 10th to May 16th, 2024 (test)
159.75Hitting CostHITMIN
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| HITMINu1=10, u2=1000, u3=3.52025.12 | 159.75 | 43.75 | 140.12 | 343.62 | |
| ROBDSelection=Optimally selected based on Theorem 42025.12 | 163.63 | 23.16 | 111.85 | 298.64 | |
| FairOBDLearning rate (η)=10^-42025.12 | 167.47 | 28.17 | 52.68 | 248.31 | |
| FairOBDLearning rate (η)=10^-32025.12 | 167.8 | 27.52 | 51.05 | 246.36 | |
| DMDLearning rate (η)=10^-32025.12 | 169.46 | 93.53 | 54.16 | 317.15 | |
| FairOBDLearning rate (η)=10^-22025.12 | 170.06 | 25.65 | 41.36 | 237.07 | |
| OPTType=Offline2025.12 | 171.3 | 23.27 | 36.36 | 230.93 | |
| FairOPTType=Offline, Objective=Minimize long-term fairness cost2025.12 | 177.2 | 351.48 | 33.33 | 562 |