Cellular Offloading via Downlink Cache Placement

In this paper, the downlink file transmission within a finite lifetime is optimized with the assistance of wireless cache nodes. Specifically, the number of requests within the lifetime of one file is modeled as a Poisson point process. The base station multicasts files to downlink users and the selected the cache nodes, so that the cache nodes can help to forward the files in the next file request. Thus we formulate the downlink transmission as a Markov decision process with random number of stages, where transmission power and time on each transmission are the control policy. Due to random number of file transmissions, we first proposed a revised Bellman’s equation, where the optimal control policy can be derived. In order to address the prohibitively huge state space, we also introduce a low-complexity sub-optimal solution based on an linear approximation of the value function. The approximated value function can be calculated analytically, so that conventional numerical value iteration can be eliminated. Moreover, the gap between the approximated value function and the real value function is bounded analytically. It is shown by simulation that, with the approximated MDP approach, the proposed algorithm can significantly reduce the resource consumption at the base station.

💡 Research Summary

The paper addresses the problem of optimizing downlink file delivery in a cellular network when files have a finite lifetime and are requested repeatedly by users. Requests for a given file during its lifetime are modeled as a Poisson point process, which leads to a random number of transmission stages. The authors propose to use wireless cache nodes that receive the file from the base station (BS) via downlink multicast and later serve users through device‑to‑device (D2D) links on a separate spectrum (e.g., Wi‑Fi). The key idea is that by investing more power or time in the early multicast transmissions, more cache nodes can decode the file (or its segments), thereby reducing the BS’s resource consumption for subsequent requests.

The system consists of a multi‑antenna BS, N_C single‑antenna cache nodes, and randomly arriving users. Each file is divided into N_S equal‑size segments. The downlink physical layer employs space‑time block coding (STBC) to enable multicast without requiring instantaneous CSI, and the ergodic capacity is assumed for each segment transmission. A transmission is successful for a user or a cache node if the achieved rate meets the per‑segment rate requirement.

The optimization objective is the weighted sum of BS transmit energy and transmission time over the entire file lifetime. Since the number of requests is random, the problem is formulated as a Markov decision process (MDP) with a random horizon. Standard finite‑horizon MDP techniques cannot be applied directly, so the authors derive a revised Bellman equation that accommodates the random number of stages. The system state includes a binary vector indicating which cache nodes have successfully decoded each segment, together with large‑scale channel parameters (path loss and shadowing).

Because the state space grows exponentially with the number of cache nodes and segments (2^{N_C·N_S}), solving the Bellman equation exactly is computationally infeasible. To overcome this, the authors propose a linear approximation of the value function: V̂(S)=∑{i,s}α{i,s}·B_{i,s}+β, where B_{i,s} denotes the binary cache‑segment indicator. The coefficients α_{i,s} and β are derived analytically based on the expected reduction in future BS cost when a cache node holds a segment. An analytical bound on the approximation error is also derived, showing that the gap depends on the request intensity λ and the file lifetime T.

The resulting low‑complexity policy selects, for each request and remaining lifetime, the transmit power P_{f,n,s} and the number of symbols N_{f,n,s} that minimize the immediate weighted cost plus the approximated future cost. Since the approximated value function is pre‑computed, the online decision making requires only simple arithmetic, yielding O(N_C·N_S) complexity per request.

Simulation experiments evaluate the proposed scheme against two baselines: (i) a naive “always transmit from BS” policy and (ii) a fixed‑power allocation policy. Parameters such as the number of cache nodes, segment count, request rate, and file lifetime are varied. Results demonstrate that the proposed method reduces the average BS energy consumption by 20–35 % and also shortens the total transmission time, especially when the request rate λ is high (i.e., many users request the same file). The performance of the linear‑approximation policy is shown to be virtually indistinguishable from the exact optimal policy obtained via exhaustive value iteration, confirming the tightness of the analytical error bound.

In summary, the contributions of the paper are: (1) a novel system model that integrates wireless cache nodes into downlink multicast for finite‑lifetime files; (2) a reformulation of the random‑horizon MDP with a revised Bellman equation; (3) a tractable linear value‑function approximation with provable error bounds; and (4) extensive simulations validating substantial resource savings at the BS. The work provides a solid theoretical foundation for cache‑assisted offloading in future cellular networks and suggests several extensions, including cooperative caching among nodes, simultaneous multi‑file delivery, and adaptation to time‑varying channels.