A Novel Framework for Medical Web Information Foraging Using Hybrid ACO and Tabu Search.

Area of Science:

• Computer Science
• Artificial Intelligence
• Information Retrieval

Background:

• Web information foraging presents challenges due to the web's dynamic and vast nature.
• Existing methods often rely on static datasets or lack adaptability to web changes.

Purpose of the Study:

• To develop and validate a novel multi-agent system for efficient web information foraging.
• To address the challenges of web dynamism and ensure real-time performance.

Main Methods:

• Proposed a two-phase architecture: initial learning on a fixed web instance and incremental learning for dynamic adaptation.
• Implemented the system using a colony of artificial ants hybridized with tabu search.
• Validated the approach on MedlinePlus, a real-world health research website.

Main Results:

• The multi-agent system demonstrated effectiveness in localizing relevant web pages.
• The system exhibited a very short response time, meeting real-time constraints.
• Promising results were observed regarding the exploitation of web regularities.

Conclusions:

• The proposed multi-agent approach offers an effective and efficient solution for web information foraging.
• The system's adaptability and real-time performance make it suitable for dynamic web environments.
• Validation on a real-world dataset confirms the practical applicability in the health domain.