An Autonomic Problem Determination and Remediation Agent for Ambiguous Situations Based on Singular Value Decomposition Technique

Authors:
Hoi Chan;Thomas Kwok
Affiliations:
IBM Thomas J. Watson Research Center, USA;IBM Thomas J. Watson Research Center, USA
Venue:
IAT '06 Proceedings of the IEEE/WIC/ACM international conference on Intelligent Agent Technology
Year:
2006

Citing 6
Cited 2

Merging and Splitting Eigenspace Models

IEEE Transactions on Pattern Analysis and Machine Intelligence
The Vision of Autonomic Computing

Computer
Latent semantic models for collaborative filtering

ACM Transactions on Information Systems (TOIS)
A Goal-based Approach to Policy Refinement

POLICY '04 Proceedings of the Fifth IEEE International Workshop on Policies for Distributed Systems and Networks
Principle Components and Importance Ranking of Distributed Anomalies

Machine Learning
Probabilistic anomaly detection in distributed computer networks

Science of Computer Programming

Research and Implementation of Knowledge-Enhanced Information Services

ICSOC '07 Proceedings of the 5th international conference on Service-Oriented Computing
An autonomic problem determination system using adaptive multi-levels dictionaries and singular value decomposition techniques

International Journal of Autonomic Computing

Quantified Score

Hi-index	0.00

Visualization

Abstract

The growing cost of managing computer systems leads to the development of autonomic systems. Problem determination and remediation (PDR) play an important role in an autonomic computer system, especially in ambiguous and unexpected error situations due to operating environment and requirement changes. A PDR agent system that can learn and adapt to changing environments, react to existing or new error situations and predict possible problems and take actions proactively is an integral part of an autonomic system. Traditional static problem determination and remediation system becomes insufficient. In this paper, we introduce a mathematical technique called "Singular Value Decomposition (SVD)" with a feedback system to enable a PDR agent to be responsive to ambiguous and undefined situations, predict possible problems and take actions proactively, and be able to adapt itself to the changing environment. This approach treats the situations-remediation actions relationship as a statistical problem. Using SVD technique, implicit higher order structure in situations and remediation actions association is modeled and extracted, coupled with a feedback system with human interactions, the agent is able to learn from new error situations and derive the corresponding remediation actions.