Method, system and storage medium for relay selection and load allocation based on rateless coding distributed computation

By optimizing relay selection and load distribution using LT codes and greedy algorithms, the problems of uneven load and communication latency in coded distributed computing are solved, thereby minimizing task execution latency and improving system efficiency.

CN117915430BActive Publication Date: 2026-06-16HARBIN INSTITUTE OF TECHNOLOGY (SHENZHEN) (INSTITUTE OF SCIENCE AND TECHNOLOGY INNOVATION HARBIN INSTITUTE OF TECHNOLOGY SHENZHEN)
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202410062986.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-01-15
Publication Date
2026-06-16
Estimated Expiration
2044-01-15

AI Technical Summary

Technical Problem

Existing distributed computing systems suffer from uneven load distribution and significant communication latency, leading to increased computation time. This is especially true in multi-hop heterogeneous networks where relay selection and load distribution cannot be effectively optimized.

Method used

A distributed computing method based on LT codes is adopted, and a relay selection and load distribution strategy is designed in combination with a greedy algorithm. The relay selection and load distribution are optimized by the minimum penalty value algorithm. An optimization problem of minimizing task completion time is established, the formulas for computation and communication delay are derived, and the load distribution is controlled to minimize task execution delay.

Benefits of technology

It effectively reduces the execution latency of coding computation tasks, improves system efficiency, has better robustness than existing solutions, and has lower computational complexity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117915430B_ABST
    Figure CN117915430B_ABST
Patent Text Reader

Abstract

The application provides a relay selection and load distribution method, system and storage medium based on code-free coding distributed computing, and the relay selection and load distribution method comprises the following steps: step one: combining the coding characteristics of the LT code and the task result calculation and transmission mechanism, deducing the task execution time; step two: taking the relay selection strategy and the load distribution scheme as the decision variable, establishing a minimum task completion time optimization problem; step three: solving the optimization problem through a minimum penalty value relay selection and load distribution algorithm based on the greedy algorithm, obtaining the optimal working node selection strategy and the optimal task load distribution scheme, and achieving the minimum task execution total delay. The beneficial effects of the application are: 1. minimizing the execution delay of the calculation task, and improving the system efficiency; 2. the relay selection strategy of the working node and the task load distribution scheme are designed by using the greedy algorithm, and the robustness is better than each scheme in the current coding distributed computing research field.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Relay selection method based on network encoding and space division multiplexing

    CN103826292A

  • Cloud computing cluster load scheduling method based on GA algorithm

    CN108881432A