Using Model-Based Diagnosis to Improve Software Testing

Authors: Tom Zamir, Roni Stern, Meir Kalech

AAAI 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results show the benefits of using an MDP-based planning algorithms over greedy test planning in three benchmarks. To evaluate TDP and the proposed heuristics we performed experiments on three benchmarks.
Researcher Affiliation Academia Tom Zamir and Roni Stern and Meir Kalech tom.zamir.i@gmail.com roni.stern@gmail.com kalech@bgu.ac.il Department of Information Systems Engineering Ben Gurion University of the Negev Be er Sheva, Israel
Pseudocode Yes Algorithm 1: An Algorithmic View of TDP
Open Source Code No The paper does not contain any explicit statements or links indicating that the source code for the described methodology is publicly available.
Open Datasets Yes The first benchmark we used, inspired by the vending machine benchmark (Campos et al. 2013; Burger and Zeller 2011; Orso et al. 2006), is a simple implementation of a vending machine logic having 19 components and 240 lines of code.
Dataset Splits No The paper does not provide specific training, validation, and test dataset splits with percentages or sample counts. It describes an iterative testing process and initial test selection, but not formal data splits.
Hardware Specification No The paper does not provide any specific hardware details such as GPU or CPU models, memory, or cloud instance types used for running the experiments.
Software Dependencies No The paper mentions software like HP Quality Center, Bugzilla, IBM Rational Clear Quest, Java’s JVMTI, and MS Visual Studio, but it does not provide specific version numbers for any of these or other key software components used in their methodology or experiments.
Experiment Setup Yes In our experiments we set B to be 0.9. C and H are parameters of the algorithm called width and depth, respectively. We experimented with two methods for setting the value of leaf states so as to estimate the amount of further testing required to reach a terminal state. In our experiments we used all the proposed heuristics (HP, BD, and Entropy) as well as simple random test selection as default heuristics.