Research on the construction and dynamic adaptation algorithm of cognitive graph multimodal knowledge network for...

Area of Science:

• Enterprise Management
• Data Science
• Supply Chain Management

Background:

• Modern enterprises manage diverse, fragmented data, hindering decision speed and collaboration, especially during supply chain disruptions.
• Existing methods like fuzzy cognitive maps (FCMs) are static and struggle with multimodal data integration.

Purpose of the Study:

• To develop a dynamic multimodal management knowledge network using cognitive maps and an event-driven update algorithm.
• To address the limitations of static models in handling real-time data and multimodal information.

Main Methods:

• Feature alignment of multimodal data using an enhanced ELMo model.
• Construction of a quantum embedding model for cross-domain generalization.
• Implementation of an event-driven weight update for dynamic network adaptation.

Main Results:

• A 74.4% increase in decision-making efficiency in a supply chain test.
• 92.1% accuracy in predicting chip-shortage risk paths.
• A 77.1% rise in supplier turnover efficiency.

Conclusions:

• The proposed method effectively builds dynamic multimodal management knowledge networks.
• The approach significantly enhances decision-making efficiency and predictive accuracy in dynamic environments like supply chains.