Advances, Systems and Applications
From: MRLCC: an adaptive cloud task scheduling method based on meta reinforcement learning
Dataset | MRLCC | PG-based DRL | Heuristic Algorithm | ||||
---|---|---|---|---|---|---|---|
 | 20 steps | 20 steps | FirstFit | Random | Shortest | Mincost | Round Robin |
Scenario 1 | 0.7973 | 0.7891 | 0.7752 | 0.7820 | 0.7793 | 0.7692 | 0.7483 |
Scenario 2 | 0.8031 | 0.7997 | 0.7955 | 0.7981 | 0.7945 | 0.7925 | 0.7732 |
Scenario 3 | 0.8234 | 0.8175 | 0.8031 | 0.8093 | 0.8035 | 0.7972 | 0.7634 |
CPU=12 | 0.7719 | 0.7682 | 0.7632 | 0.7592 | 0.7521 | 0.7491 | 0.7094 |
CPU=15 | 0.8479 | 0.8456 | 0.8415 | 0.8439 | 0.8394 | 0.8408 | 0.8103 |