A discrete memoryless channel has the following channel matrix :
[P(Y|X)] = [1 − p p 0 ; 0 p 1 − p]
(i) Draw the channel diagram.
(ii) If the source has equally likely outputs, compute the probabilities associated with the channel output for p = 0·2.
10 marks · directive: Compute
Answer
The core of a good answer
The matrix describes a binary erasure channel with two inputs and three outputs; with P(x1) = P(x2) = 0.5 and p = 0.2, P(y1) = 0.4, P(y2) = 0.2 and P(y3) = 0.4.
An outline written by upsc.help, not an official UPSC answer; every fact checked by a separate review. How we work
What the question asks
The candidate must read a 2 × 3 channel matrix, recognise the binary erasure channel, describe its transition diagram, and use the total probability theorem [P(Y)] = [P(X)][P(Y|X)] to obtain the output probabilities.
Answer outline
Introduction
Inputs x1, x2 (rows); outputs y1, y2, y3 (columns). P(x1) = P(x2) = 0.5 and p = 0.2.
Step 1: Channel diagram (part i)
- Two input nodes x1 and x2 on the left, three output nodes y1, y2, y3 on the right
- x1 → y1 with probability 1 − p; x1 → y2 with probability p
- x2 → y3 with probability 1 − p; x2 → y2 with probability p
- No branch from x1 to y3 or from x2 to y1; y2 is the erasure output
- Each row sums to 1: (1 − p) + p + 0 = 1
Step 2: Channel matrix for p = 0.2
- [P(Y|X)] = [0.8 0.2 0 ; 0 0.2 0.8]
Step 3: Output probabilities (part ii)
- [P(Y)] = [P(X)][P(Y|X)] = [0.5 0.5][0.8 0.2 0 ; 0 0.2 0.8]
- P(y1) = 0.5 × 0.8 + 0.5 × 0 = 0.4
- P(y2) = 0.5 × 0.2 + 0.5 × 0.2 = 0.2
- P(y3) = 0.5 × 0 + 0.5 × 0.8 = 0.4
Step 4: Joint probability matrix
- [P(X, Y)] = [0.4 0.1 0 ; 0 0.1 0.4]
- Column sums give the output probabilities above
Conclusion
P(y1) = 0.4, P(y2) = 0.2, P(y3) = 0.4; check: the three add up to 1.
Examples and data to cite
- The capacity of a binary erasure channel is C = 1 − p bits per symbol, so for p = 0.2, C = 0.8 bit/symbol, achieved with equiprobable inputs.
- An erasure channel never confuses 0 with 1; it only reports that a symbol was lost.
Facts
Facts to remember
- A discrete memoryless channel is fully described by its transition (channel) matrix [P(Y|X)], each row summing to 1.
- [P(Y)] = [P(X)][P(Y|X)] (total probability theorem).
- [P(X, Y)] = [P(X)]d [P(Y|X)], with [P(X)]d the diagonal matrix of input probabilities.
- Binary erasure channel: binary input, ternary output (0, 1 or an erasure symbol), erasure probability p; its capacity 1 − p is attained with a uniform input distribution.