Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Improved Coded Caching Scheme for Multi-User Information Retrieval System

Junyi Wang, Quan Zang, Jinyu Wang, Minquan Cheng

TL;DR

The paper addresses coded caching for multi-user information retrieval (MIR) with an $(L,K,M,N)$ configuration, focusing on minimizing normalized delivery time (NDT) while reducing subpacketization and computational complexity. It leverages multi-antenna placement delivery arrays (MAPDA) to design uplink precoding, and shows that when $t=\frac{KM}{N}\ge L$, the BS downlink can use identity precoding, simplifying the scheme. The authors propose three MAPDA-based schemes that achieve the same NDT as the existing ASMST scheme but with substantially lower subpacketization ($F_1$, $F_2$, $F_3$ vs $F_{ASMST}$) and reduced computational complexity, with explicit Big-O improvements and asymptotic reductions. The findings demonstrate that MAPDA is a powerful tool for practical MIR coded caching, yielding scalable performance as the user count grows, and they provide a pathway to extend to scenarios where $L>t$ by antenna silencing. Overall, the work advances efficient, implementable MIR caching by combining combinatorial arrays with uplink-downlink precoding design under the NDT framework.

Abstract

In this paper, we study the coded caching scheme for the $(L, K, M, N)$ multi-user information retrieval (MIR) system, which consists of a content library containing $N$ files, a base station (BS) with $L$ antennas that cannot access the library, and $K$ single-antenna users, each of which can cache at most $M$ files from the library. The users communicate with the others assisted by the BS to decode their required files. In this paper, we focus on designing a coded caching scheme with low communication latency measured by normalized delivery time (NDT), computational complexity, and subpacketizations. When $\frac{KM}{N}\geq L$ we first simply the precoding matrix in the downlink step to an identity matrix and use the multiple-antenna placement delivery array (MAPDA), which was originally proposed for the multiple-input single-output networks, to generate several new schemes for MIR system. Compared to the existing schemes, both the theoretical and numerical analyses show that our new schemes achieve much lower computational complexity and smaller subpacketizations with the same NDT.

Improved Coded Caching Scheme for Multi-User Information Retrieval System

TL;DR

The paper addresses coded caching for multi-user information retrieval (MIR) with an configuration, focusing on minimizing normalized delivery time (NDT) while reducing subpacketization and computational complexity. It leverages multi-antenna placement delivery arrays (MAPDA) to design uplink precoding, and shows that when , the BS downlink can use identity precoding, simplifying the scheme. The authors propose three MAPDA-based schemes that achieve the same NDT as the existing ASMST scheme but with substantially lower subpacketization (, , vs ) and reduced computational complexity, with explicit Big-O improvements and asymptotic reductions. The findings demonstrate that MAPDA is a powerful tool for practical MIR coded caching, yielding scalable performance as the user count grows, and they provide a pathway to extend to scenarios where by antenna silencing. Overall, the work advances efficient, implementable MIR caching by combining combinatorial arrays with uplink-downlink precoding design under the NDT framework.

Abstract

In this paper, we study the coded caching scheme for the multi-user information retrieval (MIR) system, which consists of a content library containing files, a base station (BS) with antennas that cannot access the library, and single-antenna users, each of which can cache at most files from the library. The users communicate with the others assisted by the BS to decode their required files. In this paper, we focus on designing a coded caching scheme with low communication latency measured by normalized delivery time (NDT), computational complexity, and subpacketizations. When we first simply the precoding matrix in the downlink step to an identity matrix and use the multiple-antenna placement delivery array (MAPDA), which was originally proposed for the multiple-input single-output networks, to generate several new schemes for MIR system. Compared to the existing schemes, both the theoretical and numerical analyses show that our new schemes achieve much lower computational complexity and smaller subpacketizations with the same NDT.
Paper Structure (15 sections, 5 theorems, 34 equations, 2 figures, 3 tables)

This paper contains 15 sections, 5 theorems, 34 equations, 2 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Contributions and organization
  3. Notations
  4. Preliminaries
  5. Model system
  6. Multi-antenna Placement Delivery Array
  7. Main results
  8. Performance Analyses
  9. Sketch of designing the scheme in Theorem \ref{['th-1']}
  10. The proof of Theorem \ref{['th-1']}
  11. The Scheme Realized by $\mathbf{P}$
  12. Decodability for Each User
  13. The Existence of $\mathbf{V}^{(s)}$ that Satisfies \ref{['eq-coding-vector2']}
  14. Proof of The Existence of $\mathbf{V}^{(s)}$
  15. Conclusion

Key Result

Lemma 1

For any positive integers $L$, $K$, $M$ and $N$ satisfying that $t=\frac{KM}{N}$ is an integer and $t+L\leq K$, there exists an $(L,K,M,N)$ MIR coded caching scheme with subpacketization $F={K\choose t}{K-t-1\choose L-1}$, the NDTs $\tau_{\text{UL}}= \tau_{\text{DL}}=\frac{{K - t}}{{t + L}}$.

Figures (2)

  • Figure 1: Multi-user information retrieval (MIR) system.
  • Figure 2: Subpacketization comparison between the schmes in Table \ref{['tab-Theorem-2']} and in ASMST, where $K$=$150$ and $L$=$10$.

Theorems & Definitions (7)

  • Lemma 1: ASMST
  • Definition 1: YWCC
  • Example 1
  • Lemma 2: Existing MAPDAs
  • Theorem 1
  • Theorem 2
  • Theorem 3