Capacity of Finite State Channels With Feedback: Algorithmic and Optimization Theoretic Properties

被引:0
作者
Grigorescu, Andrea [1 ]
Boche, Holger [1 ,2 ,3 ,4 ,5 ]
Schaefer, Rafael F. [6 ,7 ]
Vincent Poor, H. [8 ]
机构
[1] Tech Univ Munich, Theoret Informat Technol, D-80333 Munich, Germany
[2] Tech Univ Munich, BMBF Res Hub 6G life, D-80333 Munich, Germany
[3] Ruhr Univ, Excellence Cluster Cyber Secur Age Large Scale Adv, D-44801 Bochum, Germany
[4] Munich Ctr Quantum Sci & Technol MCQST, D-80799 Munich, Germany
[5] Munich Quantum Valley MQV, D-80807 Munich, Germany
[6] Tech Univ Dresden, Cluster Excellence Ctr Tactile Internet Human Inth, Informat Theory & Machine Learning, BMBF Res Hub 6G life, D-01062 Dresden, Germany
[7] Tech Univ Dresden, 5G Lab Germany, D-01062 Dresden, Germany
[8] Princeton Univ, Dept Elect & Comp Engn, Princeton, NJ 08544 USA
关键词
Power capacitors; Computational modeling; 6G mobile communication; Reliability; Turing machines; Monte Carlo methods; Digital computers; channel capacity; channels with feedback; finite-state channel (FSC); COMMUNICATION;
D O I
10.1109/TIT.2024.3411919
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
The capacity of finite state channels (FSCs) with feedback has been expressed by a limit of a sequence of multi-letter expressions. Despite many efforts, a closed-form single-letter capacity characterization remains unknown to date. In this paper, the feedback capacity is studied from a fundamental algorithmic point of view by addressing the question of whether or not the capacity can be algorithmically computed. To this aim, the concept of Turing machines is used, which provides fundamental performance limits of digital computers. It is shown that the feedback capacity of FSCs is not Banach-Mazur computable and therefore also not Borel-Turing computable. It is further shown that it is even impossible to approximate the feedback capacity function of FSCs by a computable function. As a consequence, it is shown that computable achievability and converse can never be tight, which means that there are FSCs for which it is impossible to find computable tight upper and lower bounds. Furthermore, it is shown that the feedback capacity cannot be characterized as the maximization of a finite-letter formula of entropic quantities.
引用
收藏
页码:5413 / 5426
页数:14
相关论文
共 44 条
  • [1] Aharoni Z, 2019, IEEE INT SYMP INFO, P837, DOI [10.1109/ISIT.2019.8849364, 10.1109/isit.2019.8849364]
  • [2] [Anonymous], 1930, Monatshefte f. Mathematik u. Physik, DOI [DOI 10.1007/BF01696781, 10.1007/BF01696781]
  • [4] Avigad V., Computability and Analysis: TheLegacy of Alan Turing
  • [6] Algorithmic Computability and Approximability of Capacity-Achieving Input Distributions
    Boche, Holger
    Schaefer, Rafael F.
    Poor, H. Vincent
    [J]. IEEE TRANSACTIONS ON INFORMATION THEORY, 2023, 69 (09) : 5449 - 5462
  • [7] Boche H, 2020, COMMUN INF SYST, V20, P81
  • [8] Communication Under Channel Uncertainty: An Algorithmic Perspective and Effective Construction
    Boche, Holger
    Schaefer, Rafael F.
    Poor, H. Vincent
    [J]. IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2020, 68 : 6224 - 6239
  • [9] Denial-of-Service Attacks on Communication Systems: Detectability and Jammer Knowledge
    Boche, Holger
    Schaefer, Rafael F.
    Poor, H. Vincent
    [J]. IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2020, 68 : 3754 - 3768
  • [10] Boche H, 2019, IEEE INT SYMP INFO, P470, DOI [10.1109/ISIT.2019.8849851, 10.1109/isit.2019.8849851]