Article
Computer Science, Artificial Intelligence
Mehdi Esnaashari, Amir Hossein Damia
Summary: Software testing is crucial for software quality assurance, and automating the test data generation process can help reduce time and cost. The study focuses on using structural methods to cover all finite paths effectively. The proposed method, a memetic algorithm combining reinforcement learning with genetic algorithms, outperforms traditional evolutionary and meta-heuristic algorithms in terms of speed and coverage.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Information Systems
M. Nosrati, H. Haghighi, M. Vahidi Asl
Summary: This paper discusses the application of search-based methods in finding underlying path constraints of program input parameters, aiming to improve search efficiency and avoid problems with symbolic execution by constructing approximate constraints. Experimental results demonstrate that the proposed method achieves outstanding performance on most benchmark programs.
INFORMATION AND SOFTWARE TECHNOLOGY
(2021)
Article
Chemistry, Multidisciplinary
Tatiana Avdeenko, Konstantin Serdyukov
Summary: This paper investigates an approach to intelligent support of software white-box testing process using an evolutionary paradigm, solving the problem of automated generation of optimal test data set. By formulating a fitness function with two terms and implementing genetic algorithms, it is possible to achieve maximum statement coverage and population diversity in one launch of the GA. The optimal relation between the two terms of fitness function was obtained for two different programs under testing.
APPLIED SCIENCES-BASEL
(2021)
Article
Computer Science, Information Systems
Rahul Dubey, Simon Hickinbotham, Mark Price, Andy Tyrrell
Summary: This paper presents an approach for generating offspring based on a local fitness landscape exploration to increase the speed of the search for optimal/sub-optimal solutions and to evolve better fitness solutions. The proposed algorithm, FLEX-GA, is compared with other algorithms on single and multi-objective benchmark problems, and the results show improved search speed and solution quality.
Article
Mathematics
Rong Wang, Yuji Sato, Shaoying Liu
Summary: The study introduces a new method that combines formal specifications with genetic algorithm to generate test cases effectively to kill program mutants, contributing to the further maintenance of software.
Article
Computer Science, Hardware & Architecture
Shunhui Ji, Shaoqing Zhu, Pengcheng Zhang, Hai Dong, Jianan Yu
Summary: This study proposes an improved genetic algorithm-based approach for generating test cases for smart contract data flow testing. By introducing the theory of particle swarm optimization into the genetic algorithm, the approach enhances the coverage and efficiency of test case generation. Experimental results demonstrate that the proposed approach outperforms three baseline methods in terms of coverage, iteration numbers, and execution time.
IEEE TRANSACTIONS ON RELIABILITY
(2023)
Article
Computer Science, Artificial Intelligence
Feng Zou, Debao Chen, Hui Liu, Siyu Cao, Xuying Ji, Yan Zhang
Summary: Fitness landscape analysis (FLA), as a powerful analytical tool, has been widely applied in various optimization areas. It helps to gain a deep understanding of the characteristics of complex optimization problems and improve algorithm performance on specific problems.
Article
Computer Science, Artificial Intelligence
Mohammad Hassan Tayarani Najaran
Summary: This paper analyzes the fitness landscape of timetabling problems and proposes a new operator for Quantum Evolutionary Algorithms to guide the search process towards better regions in the search space by clustering good solutions. The algorithm consists of two phases, utilizing a tabu mechanism to collect information and reinitializing individuals with data from previous search processes.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Management
Beatriz Brito Oliveira, Maria Antonia Carravilla, Jose Fernando Oliveira
Summary: Tackling uncertainty is crucial for decision-support, and scenarios can be used to model different outcomes. Besides probability-based methods, alternative approaches can be used to deal with decision-making under uncertainty. A scenario generation methodology based on genetic algorithms is proposed, which aims to obtain a diverse set of scenarios for decision-makers. This method does not require prior knowledge of probability distributions and can be applied to different problems.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2022)
Article
Biology
Gonzalez Louisa Somermeyer, Aubin Fleiss, Alexander S. Mishin, Nina G. Bozhanova, Anna A. Igolkina, Jens Meiler, Maria-Elisenda Alaball Pujol, Ekaterina Putintseva, Karen S. Sarkisyan, Fyodor A. Kondrashov
Summary: Studies on protein fitness landscapes help predict functional proteins, but the lack of systematic data on proteins with defined evolutionary relationships limits the generalization of these findings. We studied the fitness peaks of four orthologous fluorescent proteins and found that some peaks were sharp while others were flat.
Article
Computer Science, Artificial Intelligence
Bai Yan, Qi Zhao, Mengke Li, Jin Zhang, J. Andrew Zhang, Xin Yao
Summary: This paper investigates the phase-shift optimization problem between RIS and the base station, and proposes a niching genetic algorithm to solve the problem. The results show that the current methods perform poorly in cases with a large-sized RIS, while the proposed niching genetic algorithm achieves significant capacity gains.
APPLIED SOFT COMPUTING
(2022)
Article
Automation & Control Systems
Ahmed S. Ghiduk, Abdullah Alharbi
Summary: This study investigates and compares the performance of genetic algorithms and the harmony search algorithm in the test data generation process. The results show that the harmony search algorithm is significantly faster than genetic algorithms, but there is no significant difference between the two algorithms in generating adequate test data.
INTELLIGENT AUTOMATION AND SOFT COMPUTING
(2023)
Article
Computer Science, Artificial Intelligence
Rebeka Coric, Mateja Dumic, Domagoj Jakobovic
Summary: Fitness landscape analysis is used to automatically calculate good parameters for solving scheduling problems using tree-based genetic programming, resulting in better performance compared to manual parameter selection.
APPLIED INTELLIGENCE
(2021)
Article
Mathematics
Zheng Xu
Summary: This paper extends methods for testing the association between a continuous response variable and a group of common or rare genetic variants without the need for genotype calling. The proposed NGS data-based methods, derived from a linear model framework, show better statistical power compared to genotype-based methods, with improved performance as sequencing depth increases.
Article
Computer Science, Hardware & Architecture
Vincent A. Cicirello
Summary: In this paper, the theory and practice of fitness landscape analysis for optimization problems over the space of permutations are explored. A survey of distance metrics for permutations is conducted, and a classification of these metrics is provided using principal component analysis. The classification is shown to assist in selecting appropriate metrics for fitness landscape analysis based on problem characteristics, as well as guide the choice of mutation operators in evolutionary algorithms.
MOBILE NETWORKS & APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Thomas Vogel, Chinh Tran, Lars Grunske
Summary: By conducting a fitness landscape analysis, researchers were able to avoid trial-and-error experiments and find a more suitable configuration for SAPIENZ. The new approach, SAPIENZ(DIV), achieves better or at least similar test results in terms of faults and coverage compared to SAPIENZ, but typically produces longer test sequences and requires more execution time.
INFORMATION AND SOFTWARE TECHNOLOGY
(2021)
Article
Computer Science, Software Engineering
Manuel Ohrndorf, Christopher Pietsch, Udo Kelter, Lars Grunske, Timo Kehrer
Summary: In Model-driven Engineering, models are primary development artifacts that are heavily edited and may become temporarily inconsistent. There are multiple methods to resolve inconsistencies, with the suitable one depending on various factors. By analyzing inconsistencies in the version history and generating repair recommendations, past incomplete edits can be automatically fixed.
ACM TRANSACTIONS ON SOFTWARE ENGINEERING AND METHODOLOGY
(2021)
Article
Computer Science, Software Engineering
Sinem Getir Yaman, Esteban Pavese, Lars Grunske
Summary: This article introduces a probabilistic verification algorithm for stochastic regular expressions over a probabilistic extension of ACTL*, including a model checking algorithm and semantics on the probabilistic action logic for SREs. The focus is on defining SREs through local probabilistic functions, allowing for local verification of properties and reuse of results for global verification. The article demonstrates how to model a system with SREs and verify it using probabilistic action logic, along with a preliminary performance evaluation based on the reachability algorithm's execution time.
FUNDAMENTA INFORMATICAE
(2021)
Article
Computer Science, Software Engineering
Anjana Perera, Aldeida Aleti, Burak Turhan, Marcel Bohme
Summary: This paper proposes a new SBST technique called PreMOSA, which combines coverage information with defect prediction information to determine where to increase test coverage in the CUT. Experimental results show that PreMOSA is more effective and efficient than DynaMOSA in detecting bugs, detecting up to 8.3% more bugs on average.
IEEE TRANSACTIONS ON SOFTWARE ENGINEERING
(2023)
Article
Computer Science, Software Engineering
Ezekiel Soremekun, Esteban Pavese, Nikolas Havrikov, Lars Grunske, Andreas Zeller
Summary: Grammars can generate structured test inputs that are syntactically correct. By assigning probabilities to productions, the distribution of input elements can be controlled. Through learning from common inputs, uncommon inputs, and failure-inducing inputs, inputs similar or dissimilar to the sample can be generated for different testing purposes. The evaluation shows the effectiveness of these methods on different input formats.
IEEE TRANSACTIONS ON SOFTWARE ENGINEERING
(2022)
Article
Computer Science, Information Systems
Laura Wartschinski, Yannic Noller, Thomas Vogel, Timo Kehrer, Lars Grunske
Summary: This article introduces VUDENC, a deep learning-based vulnerability detection tool that automatically learns vulnerable code features with high accuracy. By applying the word2vec model and LSTM network, VUDENC can classify vulnerable code fragments finely and provide confidence levels for its predictions.
INFORMATION AND SOFTWARE TECHNOLOGY
(2022)
Review
Computer Science, Information Systems
Arut Prakash Kaleeswaran, Arne Nordmann, Thomas Vogel, Lars Grunske
Summary: This article provides an overview of the current state of counterexample explanation techniques, including the contexts, methods, and evaluation. Most studies provide graphical or trace explanations and localize errors in model checkers. However, there is a lack of research on probabilistic and real-time systems, domain-specific models, and user studies for evaluation.
INFORMATION AND SOFTWARE TECHNOLOGY
(2022)
Article
Computer Science, Software Engineering
Najam Nazar, Aldeida Aleti, Yaokun Zheng
Summary: Software design patterns are standard solutions to common problems in software design and architecture, and automatic detection of design patterns can improve efficiency. The new design pattern detection method DPDF, using machine learning classifiers and code features, achieves precision over 80% and recall over 79%, outperforming existing methods.
JOURNAL OF SYSTEMS AND SOFTWARE
(2022)
Article
Computer Science, Information Systems
Thomas Vogel, Marc Carwehl, Genaina Nunes Rodrigues, Lars Grunske
Summary: This work presents a comprehensive specification pattern catalog for UPPAAL, which supports qualitative and real-time requirements and covers all corresponding patterns of existing catalogs. The catalog is integrated with UPPAAL, allowing for the specification of requirements using patterns and providing an automated generator for translating these requirements into observer automata and TCTL formulas.
INFORMATION AND SOFTWARE TECHNOLOGY
(2023)
Article
Computer Science, Software Engineering
Alexander Schultheiss, Paul Maximilian Bittner, Alexander Boll, Lars Grunske, Thomas Thuem, Timo Kehrer
Summary: Model matching algorithms are crucial for identifying common elements in input models. Existing n-way matching algorithms often suffer from scalability issues, but our proposed algorithm utilizes multi-dimensional search trees to efficiently find suitable match candidates, leading to significant improvements in both performance and matching quality.
SOFTWARE AND SYSTEMS MODELING
(2023)
Article
Computer Science, Software Engineering
Neelofar Neelofar, Kate Smith-Miles, Mario Andres Munoz, Aldeida Aleti
Summary: Search-based software testing (SBST) is a mature area with techniques developed to tackle the challenging task of software testing. SBST techniques have been successfully applied in the industry to generate test cases for large and complex software systems. However, their effectiveness depends on the problem being addressed. This paper revisits the evaluation of SBST techniques using Instance Space Analysis (ISA) to visualize and assess their strengths and weaknesses across a broad range of problem instances from common benchmark datasets. The paper also examines the diversity and quality of benchmark datasets used in experimental evaluations.
IEEE TRANSACTIONS ON SOFTWARE ENGINEERING
(2023)
Proceedings Paper
Computer Science, Interdisciplinary Applications
Sebastian Mueller, Valentin Gogoll, Anh Duc Vu, Timo Kehrer, Lars Grunske
Summary: Software testing is crucial but often faces difficulties in determining the expected outcome. This study explores the feasibility of using metamorphic testing to automatically identify relations between input and output pairs in the exciting-NOMAD parser, a widely used software package in computational material science. The analysis of the discovered metamorphic relations focuses on their quantity and quality, and the developed tool and used data are made available through a replication package.
2022 IEEE 18TH INTERNATIONAL CONFERENCE ON E-SCIENCE (ESCIENCE 2022)
(2022)
Proceedings Paper
Computer Science, Software Engineering
Daniel Truebenbach, Sebastian Mueller, Lars Grunske
Summary: Writing software tests is crucial for ensuring software quality, and automated test case generation is particularly helpful for researchers in the scientific field. In this case study, the efficacy of automatic test case generation approaches for the ASE Python project in material sciences was investigated. The results showed that while automated methods improved the original test suite, none were able to achieve the coverage reached by the manually created test suite.
15TH SEARCH-BASED SOFTWARE TESTING WORKSHOP (SBST 2022)
(2022)
Proceedings Paper
Computer Science, Software Engineering
Hoang Lam Nguyen, Lars Grunske
Summary: This paper discusses the performance evaluation metric of fuzzers and emphasizes the importance of both coverage and behavioral diversity. Introducing BeDivFuzz, a feedback-driven fuzzing technique that achieves better behavioral diversity by employing mutation strategies based on validity and behavioral diversity.
2022 ACM/IEEE 44TH INTERNATIONAL CONFERENCE ON SOFTWARE ENGINEERING (ICSE 2022)
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Martin Kabierski, Hoang Lam Nguyen, Lars Grunske, Matthias Weidlich
Summary: This paper introduces a relevance-guided sampling approach for event logs, which learns the characteristics of event data to determine its relevance for conformance checking, resulting in improved quality of samples and conformance checking results compared to baseline strategies.
2021 3RD INTERNATIONAL CONFERENCE ON PROCESS MINING (ICPM 2021)
(2021)