Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Distributed Multi-Head Learning Systems for Power Consumption Prediction

Jia-Hao Syu, Jerry Chun-Wei Lin, Philip S. Yu

TL;DR

Power consumption prediction for AGVs in smart factories faces noise and bandwidth constraints. The paper introduces Distributed Multi-Head Learning (DMH), a split-learning framework that groups features by Pearson correlation and deploys multiple head networks (DMH-T and DMH-E) to improve accuracy while minimizing data sharing; the server aggregates predictions from client-held heads without access to raw data. DMH-T predicts next-step features before the final power forecast, whereas DMH-E ensembles head predictions for the final output, with a loss-balancing mechanism to stabilize multi-task training. Evaluated on AIUT, BMW, and Husky datasets, DMH-T and DMH-E achieve top-2 performance, with DMH-E delivering up to a 24% MAE reduction over state-of-the-art methods and robust performance across 5- and 10-step horizons, while preserving privacy and reducing transmission bandwidth to $1/W$. The approach offers practical, privacy-preserving, and bandwidth-efficient power management for real-time smart factory operations.

Abstract

As more and more automatic vehicles, power consumption prediction becomes a vital issue for task scheduling and energy management. Most research focuses on automatic vehicles in transportation, but few focus on automatic ground vehicles (AGVs) in smart factories, which face complex environments and generate large amounts of data. There is an inevitable trade-off between feature diversity and interference. In this paper, we propose Distributed Multi-Head learning (DMH) systems for power consumption prediction in smart factories. Multi-head learning mechanisms are proposed in DMH to reduce noise interference and improve accuracy. Additionally, DMH systems are designed as distributed and split learning, reducing the client-to-server transmission cost, sharing knowledge without sharing local data and models, and enhancing the privacy and security levels. Experimental results show that the proposed DMH systems rank in the top-2 on most datasets and scenarios. DMH-E system reduces the error of the state-of-the-art systems by 14.5% to 24.0%. Effectiveness studies demonstrate the effectiveness of Pearson correlation-based feature engineering, and feature grouping with the proposed multi-head learning further enhances prediction performance.

Distributed Multi-Head Learning Systems for Power Consumption Prediction

TL;DR

Power consumption prediction for AGVs in smart factories faces noise and bandwidth constraints. The paper introduces Distributed Multi-Head Learning (DMH), a split-learning framework that groups features by Pearson correlation and deploys multiple head networks (DMH-T and DMH-E) to improve accuracy while minimizing data sharing; the server aggregates predictions from client-held heads without access to raw data. DMH-T predicts next-step features before the final power forecast, whereas DMH-E ensembles head predictions for the final output, with a loss-balancing mechanism to stabilize multi-task training. Evaluated on AIUT, BMW, and Husky datasets, DMH-T and DMH-E achieve top-2 performance, with DMH-E delivering up to a 24% MAE reduction over state-of-the-art methods and robust performance across 5- and 10-step horizons, while preserving privacy and reducing transmission bandwidth to . The approach offers practical, privacy-preserving, and bandwidth-efficient power management for real-time smart factory operations.

Abstract

As more and more automatic vehicles, power consumption prediction becomes a vital issue for task scheduling and energy management. Most research focuses on automatic vehicles in transportation, but few focus on automatic ground vehicles (AGVs) in smart factories, which face complex environments and generate large amounts of data. There is an inevitable trade-off between feature diversity and interference. In this paper, we propose Distributed Multi-Head learning (DMH) systems for power consumption prediction in smart factories. Multi-head learning mechanisms are proposed in DMH to reduce noise interference and improve accuracy. Additionally, DMH systems are designed as distributed and split learning, reducing the client-to-server transmission cost, sharing knowledge without sharing local data and models, and enhancing the privacy and security levels. Experimental results show that the proposed DMH systems rank in the top-2 on most datasets and scenarios. DMH-E system reduces the error of the state-of-the-art systems by 14.5% to 24.0%. Effectiveness studies demonstrate the effectiveness of Pearson correlation-based feature engineering, and feature grouping with the proposed multi-head learning further enhances prediction performance.
Paper Structure (22 sections, 3 equations, 3 figures, 10 tables, 1 algorithm)

This paper contains 22 sections, 3 equations, 3 figures, 10 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Literature Review
  3. Power Consumption Prediction
  4. Feature Engineering
  5. Distributed Learning
  6. Proposed DMH Systems
  7. Feature Engineering
  8. Distributed Multi-Head Learning
  9. Distributed Learning
  10. Head and Prediction Networks Design
  11. Loss Balancing
  12. Pseudo-Code
  13. Environment Setting
  14. Dataset
  15. Feature Groups
  16. ...and 7 more sections

Figures (3)

  • Figure 1: Flow chart of the proposed DMH
  • Figure 2: Designed distributed learning mechanism
  • Figure 3: Effectiveness of modules (blue, orange, and green bars are for 1, 5, and 10 time steps, respectively)