Advanced International Journal of Multidisciplinary Research

E-ISSN: 2584-0487 •   Impact Factor: 9.11

An Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

Call for Paper Volume 4 Issue 4 July-August 2026 Submit your research before last 3 days of August to publish your research paper in the issue of July-August.
Submit Research Paper Join as a Reviewer

Plagiarism is checked by the leading plagiarism checker TurnItIn

Call for Paper

Volume 4 Issue 4
July-August 2026

Submit your research paper

Indexing Partners
Academia Advanced Sciences Index Bielefeld Academic Search Engine CiteSeer DRJI Google Scholar Independent Search Engine & Directory Network ISI (International Scientific Indexing) Issuu Mendeley Research Networks
RefSeek ResearcherId - Thomson Reuters ResearchGate Scribd Semantic Scholar UTeM - Universiti Teknikal Malaysia Melaka Wiki for Call for Papers WorldCat Zenodo

A Unified Framework for Self-Evolving and Self-Healing Artificial Intelligence Agents

Author(s) Mr. Prudvi Saisaran Ponduru
Country India
Abstract Artificial intelligence agents are increasingly expected to operate for long periods, use external tools, adapt to changing environments and recover from failures without continuous human supervision. Current work, however, is divided between self evolving agents that improve through memory, reflection, continual learning, reinforcement learning or evolutionary search, and self healing systems that monitor, diagnose, contain, repair and verify operational faults. This paper integrates these traditions into a unified Self Evolving and Self Healing Agent framework. A four axis taxonomy is introduced across the object of change, timescale of adaptation, recovery boundary and assurance mechanism. A dual loop architecture is then proposed in which a fast runtime healing loop handles detection, diagnosis, containment, repair and verification, while a slower evidence gated evolution loop updates durable artifacts such as structured memory, prompts, workflows, tool skills, code patches and model adapters. Formal objectives are specified for utility, risk, recovery cost, retention and transfer, and a selective persistence rule is defined to prevent emergency behavior from becoming an unsafe permanent update. A multi domain, fault injected evaluation protocol is presented with baselines, ablations, benchmarks and metrics for task success, adaptation gain, retention, transfer, detection delay, time to recovery, degradation area, false recovery and safety violations. The analysis shows that self evolution and self healing should be evaluated as complementary capabilities under runtime assurance, auditability, interruptibility and staged governance. The paper is a conceptual and methodological contribution; it does not claim new benchmark results for the proposed architecture.
Keywords Agentic Artificial Intelligence, Self-Evolution, Self-Healing, Continual Learning, Runtime Assurance, Autonomous Recovery
Discipline Computer > AI / ML
Published In Volume 4, Issue 4, July-August 2026
Published On 2026-07-26
DOI https://doi.org/10.62127/aijmr.2026.v04i04.1435
View / Download Full-text Research PDF File

Share this



E-ISSN 2584-0487
Downloads
Cover Page Vol 2 Isu 5 Cover Page Vol 2 Isu 4 Cover Page Vol 2 Isu 3 Research Paper Format

All research papers published on this website are licensed under Creative Commons Attribution-ShareAlike 4.0 International License, and all rights belong to their respective authors/researchers.

CC-BY-SA Open Access

About AIJMR
Fees & Payment
Current Issue
Publication Archive 		Submit Research Paper
Track Submission Status
Publication Guidelines
Publication Ethics
Peer Review & Plagiarism 		Join as a Reviewer
Editors & Reviewers
Reviewer Referral Program
Get Reviewer Membership Certi. 		Website/Journal Policies
Usage Policy
Content Policies
Privacy Policy
Contact Us +91-9687-828-838 editor@aijmr.com