Genetic Programming in Automatic Discovery of Relationships in Computer System Monitoring Data

Modern computer systems have become very complex. Analyzing and modifying them requires substantial experience and knowledge. To make administrative tasks easier, automated methods for discovering relationships between system components are required. In this paper we discuss the use of genetic programming as a method for identification of meaningful relationships between computer system components. We present our implementation of evolutionary computations environment and compare it with an already existing solution. Next we analyze results of a sample experiment and share our conclusions. The final section provides directions for future work.