Dynamic Power Control for NOMA Transmissions in Wireless Caching Networks

Joint Optimization of Caching and Computation in Multi-Server NOMA-MEC System via Reinforcement Learning

“…(15) makes sure that the sum of computational resource allocated to the offloading users can not exceed the computing capacity of MEC server. (16) guarantees that the cached computation results do not exceed the caching capacity of the MEC server. In formula (17), it represents the energy consumption of N users in the same BS is limited to maximum energy consumption.…”

“…According to the above method, GRU algorithm is used to predict the popularity of different tasks in the near future periods. In the case of not exceeding the cache capacity of the MEC server, that is, in order to satisfy the constraint of formula (16), it is important to determine what computation results are stored on the server. Considering the cache capacity of the server and the size of computation results, the computing results of the task are cached on the server in order of popularity from large to small, so as to cache the computing results of the task as much as possible.…”

“…In the recent work [15], the authors proposed the method of deep reinforcement learning to solve power allocation in cache-aided NOMA systems. In terms of NOMA combined with caching, dynamic power control for NOMA transmissions in wireless caching networks was studied in [16], which designed a deep neural network (DNN)-based method to keep a balance between the performance and the computational complexity. However, caching in NOMA systems is generally about content caching.…”

Enhancing Physical Layer Security of Random Caching in Large-Scale Multi-Antenna Heterogeneous Wireless Networks

“…5 Let r b k ,ℓx,0 (d j,ℓx,0 ) denote the distance between BS b k ∈ Φ k (jammer j ∈ Φ J ) and ℓ x,0 . Based on (3), the received signal at ℓ x,0 is given by (4), as shown at the top of this page, where b n,k,0 ∈ Φ n,k,0 denotes the index of the serving BS of the typical user ℓ ,0 in tier k ,0 ∈ K, 6 b n,ke,0 ∈ Φ n,ke,0 denotes the index of an arbitrary BS storing file n in tier k e,0 ∈ K, and (·) T denotes the conjugate transpose operation. Then, the SIR at ℓ x,0 can be expressed as (5), as shown at the top of this page, where I x b n,k x,0…”

“…However, the majority of such traffic is asynchronously but repeatedly requested by many users at different times and thus a tremendous amount of mobile data traffic has actually been redundantly generated over the mobile networks [2]. Against this backdrop, caching popular contents at the base stations (BSs) has been proposed as a promising approach for shifting the huge mobile data traffic from remote clouds to the edges in mobile networks [3], [4], thereby significantly improving the users' quality of service (QoS). Motivated by this, cacheenabled wireless networks has recently been attracting great research attention.…”