•  
  •  
 

Journal of Undergraduate Research

Keywords

error discovery, model checking, computer systems

College

Physical and Mathematical Sciences

Department

Computer Science

Abstract

As computer systems grow in complexity it is becoming harder to completely test them. The cost of finding and correcting errors after deployment is staggering. Current validation methods are not well suited to finding subtle errors laden in the concurrency of modern systems. This project searches the behavior in such systems, guiding the search to focus its time in areas that seem most likely to contain errors. This work decreases the amount of time and memory needed for the errors that are found and corrected before deployment of the system. In this project we have developed a heuristic that takes a guided search algorithm to find defects in formal models of computer systems, a communication protocol, or circuits. The guided search explores a model using a heuristic function to estimate error locations as opposed to an exhaustive search which enumerates all the possible states to find the error in a brute force manner.

Download

Share

COinS