In the source coding problem with cost constraint, a cost function is defined over the code alphabet. This can be regarded as a noiseless channel coding problem with cost constraint. In this case, we will not distinguish between the input alphabet and the output alphabet of the channel. However, we must distinguish them for a noisy channel. In the channel coding problem with cost constraint so far, the cost function is defined over the input alphabet of the noisy channel. In this paper, we define the cost function over the output alphabet of the channel. And, the cost is paid only after the received word is observed. Note that the cost is a random variable even if the codeword is fixed. We show the channel capacity with cost constraint defined over the output alphabet. Moreover, we generalize it to tolerate some decoding error and some cost overrun. Finally, we show that the cost constraint can be described on a subset of arbitrary set which may have no structure.
Mikihiko NISHIARA
Shinshu University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Mikihiko NISHIARA, "Channel Coding with Cost Paid on Delivery" in IEICE TRANSACTIONS on Fundamentals,
vol. E105-A, no. 3, pp. 345-352, March 2022, doi: 10.1587/transfun.2021TAP0002.
Abstract: In the source coding problem with cost constraint, a cost function is defined over the code alphabet. This can be regarded as a noiseless channel coding problem with cost constraint. In this case, we will not distinguish between the input alphabet and the output alphabet of the channel. However, we must distinguish them for a noisy channel. In the channel coding problem with cost constraint so far, the cost function is defined over the input alphabet of the noisy channel. In this paper, we define the cost function over the output alphabet of the channel. And, the cost is paid only after the received word is observed. Note that the cost is a random variable even if the codeword is fixed. We show the channel capacity with cost constraint defined over the output alphabet. Moreover, we generalize it to tolerate some decoding error and some cost overrun. Finally, we show that the cost constraint can be described on a subset of arbitrary set which may have no structure.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2021TAP0002/_p
Copy
@ARTICLE{e105-a_3_345,
author={Mikihiko NISHIARA, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Channel Coding with Cost Paid on Delivery},
year={2022},
volume={E105-A},
number={3},
pages={345-352},
abstract={In the source coding problem with cost constraint, a cost function is defined over the code alphabet. This can be regarded as a noiseless channel coding problem with cost constraint. In this case, we will not distinguish between the input alphabet and the output alphabet of the channel. However, we must distinguish them for a noisy channel. In the channel coding problem with cost constraint so far, the cost function is defined over the input alphabet of the noisy channel. In this paper, we define the cost function over the output alphabet of the channel. And, the cost is paid only after the received word is observed. Note that the cost is a random variable even if the codeword is fixed. We show the channel capacity with cost constraint defined over the output alphabet. Moreover, we generalize it to tolerate some decoding error and some cost overrun. Finally, we show that the cost constraint can be described on a subset of arbitrary set which may have no structure.},
keywords={},
doi={10.1587/transfun.2021TAP0002},
ISSN={1745-1337},
month={March},}
Copy
TY - JOUR
TI - Channel Coding with Cost Paid on Delivery
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 345
EP - 352
AU - Mikihiko NISHIARA
PY - 2022
DO - 10.1587/transfun.2021TAP0002
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E105-A
IS - 3
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - March 2022
AB - In the source coding problem with cost constraint, a cost function is defined over the code alphabet. This can be regarded as a noiseless channel coding problem with cost constraint. In this case, we will not distinguish between the input alphabet and the output alphabet of the channel. However, we must distinguish them for a noisy channel. In the channel coding problem with cost constraint so far, the cost function is defined over the input alphabet of the noisy channel. In this paper, we define the cost function over the output alphabet of the channel. And, the cost is paid only after the received word is observed. Note that the cost is a random variable even if the codeword is fixed. We show the channel capacity with cost constraint defined over the output alphabet. Moreover, we generalize it to tolerate some decoding error and some cost overrun. Finally, we show that the cost constraint can be described on a subset of arbitrary set which may have no structure.
ER -