An Improved Entropy-Weighted Topsis Method for Decision-Level Fusion Evaluation System of Multi-Source Data

Area of Science:

• Industrial Internet of Things (IIoT)
• Data Science
• Manufacturing Engineering

Background:

• Traditional manufacturing faces challenges in digital transformation due to rapid IIoT advancements.
• Multi-source data fusion is crucial for integrating diverse data streams in industrial settings.
• Inconsistent data scales and low fusion accuracy hinder effective model integration.

Purpose of the Study:

• To propose an improved entropy-weighted TOPSIS method for a multi-source data fusion evaluation system.
• To address challenges of data scale inconsistency and low fusion accuracy in multi-source data.
• To develop a dynamic fusion strategy for optimal fusion results.

Main Methods:

• An improved entropy-weighted TOPSIS method was developed.
• A fusion evaluation system based on a decision-level fusion algorithm was integrated.
• A dynamic fusion strategy was proposed to manage data inconsistencies.

Main Results:

• The proposed system effectively solves data scale inconsistency issues among multi-source data.
• Optimal fusion results were achieved, overcoming difficulties in model fusion and accuracy.
• Experimental validation demonstrated effectiveness in multilayer feature fusion and decision-level fusion.

Conclusions:

• The developed fusion evaluation system provides a robust solution for multi-source data fusion in manufacturing.
• The findings offer practical value for intelligent production and assembly plants in the discrete industry.
• The system supports management and decision-making in digitally transforming manufacturing environments.