Relay channel¶
An information-theoretic channel with a source, a causally acting relay, and a destination, whose joint channel law determines what cooperation can achieve.
Core Idea¶
A relay channel is an information-theoretic communication model with a source, an observing relay, and a destination. The relay may transmit based only on observations already available to it. A joint channel law specifies how source and relay inputs produce the relay's and destination's outputs; that law determines which rate claims can be proved.[^ref-8d5395711bf1]
Scope of Application¶
The model covers channel laws with different noise and dependence structures. Cover and El Gamal analyze a degraded Gaussian case; El Gamal and Aref analyze a semideterministic class in which the relay observation is a deterministic function of transmitted inputs. Both retain the source, causal relay, destination, and joint-law roles.[ref-8d5395711bf1][ref-6c41185aebf9]
Clarity¶
A drawing with three terminals says where a relay sits, but does not say what it learns or when it can act. A rate upper bound, an achievable rate, and exact capacity are also different results. Special conditions that make a bound tight cannot be silently applied to every relay channel.[^ref-8d5395711bf1]
Manages Complexity¶
The model keeps the message, encoders, observations, decoder, and channel law needed for a reliable-communication question. It leaves implementation details outside the proof while keeping the assumptions that decide whether a theorem applies.
Abstract Reasoning¶
Ask what the relay has observed before each transmission, then write how the two transmitted inputs generate both outputs. Check a degraded or semideterministic theorem only after matching those assumptions to the proposed law.[ref-8d5395711bf1][ref-6c41185aebf9]
Knowledge Transfer¶
The role map travels among formal communication models with different alphabets and laws. An ordinary device or person called a relay is not yet this model unless its causal coding role and channel law are specified.
Example¶
In the semideterministic case, the relay's output is fixed by source and relay inputs while the destination's output can still follow a more general law. The relay acts from past observations; its deterministic observation is a special property of this class, not a property of all relay channels.[^ref-6c41185aebf9]
Neighborhood in Abstraction Space¶
Relay channel sits in a sparse region of the domain-specific corpus (97th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Information Theory & Error-Correcting Codes (20 abstractions)
Nearest neighbors
- Shannon–Weaver Communication Model — 0.80
- Error-Correcting Codes with Feedback — 0.77
- Sequential decoding — 0.76
- Information Flow (Information-Flow Security) — 0.76
- Convolutional code — 0.76
Computed from structural-signature embeddings · 2026-10-08
Not to Be Confused With¶
A physical repeater is not by itself a relay-channel specification. A degraded channel is one special class. An achievable rate or upper bound is not a universal exact capacity formula.[^ref-8d5395711bf1]
[^ref-8d5395711bf1]: Thomas M. Cover and Abbas El Gamal, “Capacity Theorems for the Relay Channel,” IEEE Transactions on Information Theory 25(5), 572–584 (1979), especially pp. 573, 577–581, Section IV pp. 578–579 and Figure 3. https://isl.stanford.edu/~cover/papers/transIT/0572cove.pdf [^ref-6c41185aebf9]: Abbas El Gamal and Mohammad Aref, “The Capacity of the Semideterministic Relay Channel,” IEEE Transactions on Information Theory 28(3), 536 (1982), abstract, introduction and corollary. https://isl.stanford.edu/groups/elgamal/abbas_publications/J011.pdf