Primary Broadcast Channels

• Primary broadcast channels are a class of two-receiver discrete-memoryless channels defined by weighted-sum comparisons that equate Marton’s inner bound with an auxiliary-receiver outer bound.
• They enable exact capacity characterizations for sums of degraded, less-noisy, more-capable, deterministic, and semi-deterministic components by decomposing the channel into tractable cases.
• The approach unifies classical receiver orderings and employs strengthened conditions (e.g., split subclasses) to address limitations of standard outer bounds in broadcast channel capacity analysis.

Primary broadcast channels are a context-dependent notion in broadcast-channel theory. In the most specific and recent information-theoretic sense, they are a class of two-receiver discrete-memoryless broadcast channels introduced so that an auxiliary-receiver outer bound matches Marton’s inner bound for sum broadcast channels, thereby yielding exact capacity characterizations for sums of degraded, less-noisy, more-capable, deterministic, and semi-deterministic components (Gohari et al., 11 Jun 2026). In earlier literature, the same phrase also appears in special-case degraded-message-set formulations and in operational models for fading and wireless broadcast systems, so the term does not denote a single universal object across subfields (0902.1591).

1. Broadcast-channel background and the problem primary classes address

A two-receiver discrete-memoryless broadcast channel is written as $T_{YZ|X}(y,z|x)$, with input alphabet $\mathcal X$ and outputs $\mathcal Y,\mathcal Z$. The sender communicates a common message $M_0$ to both receivers and private messages $M_1$ to $Y$ and $M_2$ to $Z$. In the formulation used for sum broadcast channels, Marton’s achievable region $\mathcal M(T)$ is described by auxiliaries $(U,V,W)$ satisfying $\mathcal X$0, with rate constraints

$\mathcal X$1

$\mathcal X$2

$\mathcal X$3

$\mathcal X$4

The corresponding UVW outer bound $\mathcal X$5 has the same common-message constraint, but replaces Marton’s final sum-rate term by two larger outer-bound expressions involving $\mathcal X$6 and $\mathcal X$7. The key issue is that for many channels Marton and UVW coincide, but not always (Gohari et al., 11 Jun 2026).

This matters because the general two-receiver discrete-memoryless broadcast-channel capacity region is unknown. Even in the restricted class of broadcast channels with binary inputs and symmetric outputs, the literature emphasizes that the general capacity region remains unresolved, although tractable subclasses can be identified through more-capable and essentially-less-noisy orderings (Geng et al., 2010). Primary broadcast channels were introduced precisely to enlarge the collection of tractable cases for sum constructions.

2. Formal definition of the primary class

The paper "The Capacity Region for Classes of Sum-Broadcast Channels" (Gohari et al., 11 Jun 2026) defines primary broadcast channels through weighted-sum comparisons between inner and outer bounds. For $\mathcal X$8, a broadcast channel $\mathcal X$9 is in

$\mathcal Y,\mathcal Z$0

if, for every $\mathcal Y,\mathcal Z$1,

$\mathcal Y,\mathcal Z$2

For $\mathcal Y,\mathcal Z$3, the corresponding condition is

$\mathcal Y,\mathcal Z$4

The same work also introduces stronger split subclasses $\mathcal Y,\mathcal Z$5 and $\mathcal Y,\mathcal Z$6, in which the inner bound dominates one specific outer expression for all $\mathcal Y,\mathcal Z$7. These subclasses are used when the two components of a sum channel satisfy different one-sided comparisons.

The conceptual content of the definition is explicit in the source: primary channels identify when the outer bound and inner bound coincide at the level of weighted sum-rates, and this is the right notion because the capacity region is convex and is determined by supporting hyperplanes (Gohari et al., 11 Jun 2026). A plausible implication is that “primary” is not a physical degradation property in the usual sense; it is a converse-compatible structural property of the weighted-sum geometry.

3. Sum broadcast channels and the exact characterization they enable

The primary-class definition is tailored to sum broadcast channels. A sum channel has the disjoint-union form

$\mathcal Y,\mathcal Z$8

with

$\mathcal Y,\mathcal Z$9

and

$M_0$0

For this disjoint-union structure, the weighted sum-rate of Marton’s region tensorizes cleanly over the two components. The paper proves max-$M_0$1 formulas of the form

$M_0$2

followed by an equivalent logarithmic representation after minimizing over $M_0$3. This decomposition is the analytical basis for the primary-channel theorem (Gohari et al., 11 Jun 2026).

The main theorem states that if $M_0$4 and, for a given weight triple $M_0$5, at least one of the following holds—

1. $M_0$6,
2. one component is in $M_0$7 and the other in $M_0$8,
3. one component is in $M_0$9 and the other in $M_1$0 (or vice versa)—

then

$M_1$1

If the condition holds for all admissible weights, the full capacity region is characterized: $M_1$2 This generalizes El Gamal’s earlier result for the sum of two reversely degraded broadcast channels (Gohari et al., 11 Jun 2026).

4. Relation to classical tractable broadcast-channel classes

The primary class was designed to absorb several previously tractable broadcast-channel families into a single sum-capacity framework. The 2026 analysis proves that less-noisy channels belong to $M_1$3, semi-deterministic channels belong to $M_1$4, more-capable channels belong to $M_1$5 or $M_1$6 depending on which receiver is more capable, and MIMO Gaussian broadcast channels are also in $M_1$7. Consequently, the theorem covers sums of degraded, less-noisy, more-capable, deterministic, and semi-deterministic channels (Gohari et al., 11 Jun 2026).

This placement is easier to interpret against older receiver-order taxonomies. In the binary-input symmetric-output setting, the literature defines $M_1$8 by

$M_1$9

and $Y$0 through a sufficient class of inputs for which

$Y$1

Within the BISO class, the relations reverse: $Y$2 The same work shows that if either receiver in a BISO broadcast channel is a BSC or a BEC, then the superposition coding region is the capacity region (Geng et al., 2010).

Primary broadcast channels therefore do not replace degraded, more-capable, or less-noisy orderings; they sit above them as a weighted-sum criterion tailored to sum channels. This suggests that “primary” is best understood as a unifying operational class for exact converse matching, rather than as a rival dominance notion.

5. Earlier and alternate uses of the term

Across the literature, “primary broadcast channel” is not used uniformly. The following usages appear in the cited sources.

Paper	Setting	Use of the phrase
"Correlated Sources over Broadcast Channels" (0902.1591)	Two-receiver DM-BC with correlated sources	The theorem covers the classical degraded-message-set situation as a special case, and the paper states that this is the sense in which the result connects directly to primary broadcast channels.
"The Broadcast Approach in Communication Networks" (Tajer et al., 2021)	State-dependent fading channels without instantaneous CSIT	The broadcast approach for primary broadcast channels treats each possible fading gain as a virtual user in a degraded broadcast channel and uses layered or superposition coding.
"Throughput-Optimal Broadcast in Wireless Networks with Point-to-Multipoint Transmissions" (Sinha et al., 2017)	Multi-hop wireless network broadcast	Primary broadcast channels are treated as a transmission model in which a single node transmission is heard by all of its out-neighbors simultaneously, under interference-constrained scheduling.
"Broadcast Strategies with Probabilistic Delivery Guarantee in Multi-Channel Multi-Interface Wireless Mesh Networks" (Oliveira et al., 2011)	Wireless mesh networks with channel/interface assignment	The paper explicitly states that it does not define a special universal “primary broadcast channel,” although in some strategies a common control or broadcast channel effectively plays that role.

These usages are related but not identical. In source coding and channel coding, the phrase is tied to degraded-message-set structure. In fading systems, it names the virtual-receiver interpretation of state uncertainty. In wireless networking, it often refers to point-to-multipoint transmission or to a designated common broadcast resource. The modern 2026 definition is more formal and is anchored in weighted-sum converse matching for sum broadcast channels.

6. Significance, limitations, and nearby developments

The main significance of the primary class is methodological. The 2026 paper emphasizes that UVW is a standard outer bound, but for some sum channels it is too loose; by contrast, the auxiliary-receiver outer bound can be tailored to the sum structure and can match Marton even when UVW does not. It gives an explicit example of a reversely semi-deterministic sum broadcast channel for which

$Y$3

while the auxiliary-receiver bound still captures the capacity (Gohari et al., 11 Jun 2026).

The limitations are equally clear. Primary broadcast channels do not solve the general broadcast-channel capacity problem. The broader literature continues to describe that problem as notoriously hard, and in some settings the best known inner and outer bounds still differ (Geng et al., 2010). An instructive contrast appears in broadcast channel simulation: under common-randomness assistance, the asymptotic simulation region of a discrete-memoryless broadcast channel admits an exact single-letter characterization in terms of multipartite mutual information, even though the ordinary broadcast-channel capacity problem remains unresolved (Cao et al., 2022).

A plausible implication is that the importance of primary broadcast channels lies less in introducing a new physical channel model than in isolating a class for which exact capacity survives the sum operation. In that sense, the term marks a shift from receiver ordering alone to a more refined criterion: compatibility between Marton’s weighted-sum inner bound and an auxiliary-receiver converse.