From: A reinforcement learning-based computing offloading and resource allocation scheme in F-RAN
Parameters | Value |
---|---|
M | The number of FAPs |
N | The number of UEs |
K | The number of DCNs |
\({{t}_{n}}\) | The task of UE n |
\({{B}_{n}}\) | The number of CPU cycles |
\({{D}_{n}}\) | The size of the task data |
\({{d}_{n}}\) | Offloading decision vector |
\(\varvec{P}\) | Network topology matrix |
\({{d}_{FAP}}\) | The maximal distance of FAP |
\(\varvec{Y}\) | The willingness matrix of DCN |
\(f_{n}^{l}\) | The computational capacity of UE n |
\(z{}_{n}\) | The energy consumption in per CPU cycle of UE n |
\(\rho _{n}^{t}\) | The weight factors of latency |
\(\rho _{n}^{e}\) | The weight factors of energy |
\({{f}_{n,m}}\) | The allocated computational resource to UE n in FAP m |
\({{f}_{k}}\) | The computational capacity of DCN k |
\({{T}_{c}}\) | The round-trip transmission delay |
\(f_{n}^{Cloud}\) | The allocated computational resource to UE n at the cloud server |
\({{f}^{FAP}}\) | The computational resource of FAP |
\({{f}^{Cloud}}\) | The computational resource of cloud server |
\({{C}_{m}}\) | The maximum accessible number of FAP |
T | The steps in each training epoch |
\(\varvec{S}\) | The state matrix |
\(\varvec{C}\) | The optimal caching matrix |
\({{N}_{m}}\) | The number of UEs who offload their tasks to FAP m |