Opportunistic Communications in Fading Multiaccess Relay Channels

The problem of optimal resource allocation is studied for ergodic fading orthogonal multiaccess relay channels (MARCs) in which the users (sources) communicate with a destination with the aid of a half-duplex relay that transmits on a channel orthogonal to that used by the transmitting sources. Under the assumption that the instantaneous fading state information is available at all nodes, the maximum sum-rate and the optimal user and relay power allocations (policies) are developed for a decode-and-forward (DF) relay. With the observation that a DF relay results in two multiaccess channels, one at the relay and the other at the destination, a single known lemma on the sum-rate of two intersecting polymatroids is used to determine the DF sum-rate and the optimal user and relay policies. The lemma also enables a broad topological classification of fading MARCs into one of three types. The first type is the set of partially clustered MARCs where a user is clustered either with the relay or with the destination such that the users waterfill on their bottle-neck links to the distant receiver. The second type is the set of clustered MARCs where all users are either proximal to the relay or to the destination such that opportunistic multiuser scheduling to one of the receivers is optimal. The third type consists of arbitrarily clustered MARCs which are a combination of the first two types, and for this type it is shown that the optimal policies are opportunistic non-waterfilling solutions. The analysis is extended to develop the rate region of a K-user orthogonal half-duplex MARC. Finally, cutset outer bounds are used to show that DF achieves the capacity region for a class of clustered orthogonal half-duplex MARCs.

💡 Research Summary

The paper investigates optimal resource allocation for ergodic fading orthogonal multi‑access relay channels (MARC) in which a half‑duplex relay operates on a channel orthogonal to those used by the sources. All nodes are assumed to have perfect instantaneous channel state information (CSI). The authors focus on a decode‑and‑forward (DF) relaying strategy, which naturally decomposes the network into two multiple‑access channels (MACs): one formed by the sources and the relay, and another formed by the sources and the destination. The sum‑rate of the overall system is therefore the intersection of the two MAC rate regions, each of which is a polymatroid. By invoking a known lemma on the sum‑rate of intersecting polymatroids, the authors derive a compact expression for the DF sum‑rate and obtain the optimal power‑allocation policies for the sources and the relay.

A central contribution is the classification of fading MARCs into three topological types based on how users are “clustered’’ with respect to the relay and the destination:

1. Partially clustered MARCs – each user is close to either the relay or the destination, but not both. In this case the bottleneck link for each user is uniquely identified, and the optimal policy reduces to classic water‑filling over that bottleneck link. The relay’s power acts as a common constraint; when it is abundant, only the users’ water‑filling matters.

2. Clustered MARCs – all users are jointly close to the same receiver (either all near the relay or all near the destination). The two MACs collapse into a single MAC at that receiver, and the optimal strategy becomes opportunistic multi‑user scheduling: at any fading state the user with the strongest instantaneous channel to the common receiver is selected, and the entire power budget is allocated to that user. This yields a non‑convex but analytically tractable solution.

3. Arbitrarily clustered MARCs – a mixture of the first two cases, where some users are relay‑centric and others are destination‑centric. Here the optimal policies are no longer simple water‑filling; instead they are opportunistic non‑water‑filling solutions obtained by solving the KKT conditions of the Lagrangian formed from the intersecting polymatroid constraints. The resulting power allocation dynamically balances the competing bottleneck links and the relay’s power limitation.

The authors formalize the above intuition by constructing the Lagrangian for the DF sum‑rate maximization problem, deriving the Karush‑Kuhn‑Tucker (KKT) conditions, and solving them analytically for each topological class. The analysis reveals that the relay power appears as a coupling variable that can either dominate the sum‑rate (when the relay‑to‑destination link is weak) or become negligible (when that link is strong), thereby dictating whether the users’ power allocation is purely water‑filling or requires joint optimization with the relay.

The framework is then extended to a general K‑user orthogonal half‑duplex MARC. By indexing each user i with channel gains (h_{iR}) (source‑to‑relay) and (h_{iD}) (source‑to‑destination), the same polymatroid intersection lemma applies, and the three‑class classification remains valid irrespective of K. The optimal policies scale linearly with the number of users, and the computational complexity is dominated by sorting the instantaneous channel gains, which can be performed in (O(K\log K)) time.

To assess optimality, the paper derives cut‑set outer bounds for the orthogonal MARC and compares them with the achievable DF region. For a subclass of clustered MARCs—specifically when all users are close to the relay and the relay‑to‑destination link is sufficiently strong—the DF sum‑rate meets the cut‑set bound, establishing that DF achieves the capacity region for this class. This result underscores the practical relevance of the proposed policies for relay‑enhanced cellular systems where relays are deployed near the base station.

In summary, the paper makes several key contributions: (i) it provides a unified polymatroid‑based method to compute the DF sum‑rate for fading MARCs; (ii) it classifies MARCs into three topological regimes and derives the corresponding optimal power‑allocation policies (water‑filling, opportunistic scheduling, and opportunistic non‑water‑filling); (iii) it generalizes the results to an arbitrary number of users; and (iv) it identifies conditions under which DF is capacity‑achieving. These insights bridge information‑theoretic optimality and practical resource‑allocation algorithms, offering a solid theoretical foundation for the design of future relay‑assisted wireless networks.