Deep Imperative Mutations have Less Impact

Created by W.Langdon from gp-bibliography.bib Revision:1.8051

@Article{langdon:2024:ASE,

• author = "W. B. Langdon and David Clark",
• title = "Deep Imperative Mutations have Less Impact",
• journal = "Automated Software Engineering",
• note = "Accepted 26 Oct 2024",
• keywords = "genetic algorithms, genetic programming, genetic improvement, Software testing, robust software, fault masking, fault localization, resilience, repair, automatic code optimisation, failed disruption propagation, FDP, Voas PIE (propagation, infection, and execution), fitness landscape, information theory, local search, SBSE, image processing, RNAfold",
• ISSN = "0928-8910",
• URL = "http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/langdon_2024_ASE.pdf",
• DOI = "doi:10.1007/s10515-024-00475-4",
• size = "39 pages",
• abstract = "Information theory and entropy loss predict deeper more hierarchical software will be more robust. Using the genetic improvement (GI) tool MAGPIE we measure the impact of source code mutations and how this varies with execution depth in two diverse multi-level nested software. gem5 is a million line single threaded state-of-the-art C++ discrete time VLSI circuit simulator, whilst PARSEC VIPS is a non-deterministic parallel computing multi-threaded image processing benchmark written in C. More than 53 percent (VIPS) 28 percent (gem5, omitting obviously equivalent) of mutants compile and generate identical results to the original program. We observe 12 percent and 16 percent Failed Disruption Propagation (FDP). Excluding internal errors, exceptions and asserts, most faults below 30 nested function levels which are Executed and Infect data or divert control flow are not Propagated to the output, i.e. these deep PIE changes have no visible external effect. Suggesting (where it relies on runtime analysis or testing) automatic software engineering on highly structured code will be hard.",

}

Genetic Programming entries for William B Langdon David Clark

Citations