From: Task admission control for application service operators in mobile cloud computing
Name | Description |
---|---|
L | The number of application service levels |
M | The number of application service classes |
cl | The computing resources allocated for the l-level application service |
bl | The radio resources allocated for the l-level application service |
\(P_{l}^{r}\) | The rejecting probability guaranteed for the l-level application service |
BL | The lower radio resource bound |
BU | The upper radio resource bound |
\(A_{l}^{m}\) | The arrival event of the l-level and m-class task request |
\(D_{l}^{m}\) | The departure event of the l-level and m-class task |
O | The radio resource variation event |
C | The base computing resources |
CU | The upper bound of the extendable computing resources |
aa | The action to accept the task request |
ar | The action to reject the task request |
ad | The action taken when a departure event occurs |
ao | The action taken when a radio resource variation event occurs |
\(R_{l}^{m}\) | The income from accepting and executing a l-level and m-class task |
F0 | The penalty coefficient |
fc(·) | The cost coefficient of occupying the computing resources |
fb | The cost coefficient of occupying the radio resources |
α, β | Two learning rates |
N, V | Two maximum iterations in the RL-based policy algorithm |
agreedy | The greedy action that results in the highest action-value function |
arandom | The random action selected from the action space |
ρ | The average reward |
T | The current temperature |
T0 | The initial temperature |
Tγ | The temperature dropping coefficient |
rn | The accumulated reward until the nth iteration |
τn | The accumulated time until the nth iteration |
ωl | The Lagrange multiplier |
\(P_{\omega _{l}}^{r}\) | The rejecting probability with ωl |
δl | The coefficient to update ωl |
εl | The accuracy requirement |
\(\lambda _{l}^{m}\) | The arrival rate of the l-level and m-class task request |
\(\mu _{l}^{m}\) | The departure rate of the l-level and m-class task |
λo | The occurrence rate of the radio resource variation event |