Bedir Tekinerdogan

Alireza Joonbakhsh, Alireza Rostami, AmirMohammad Kamalinia et al.

The rapid proliferation of artificial intelligence (AI) models and methods presents growing challenges for research software engineers and researchers who must select, integrate, and maintain appropriate models within complex research workflows. Model selection is often performed in an ad hoc manner, relying on fragmented metadata and individual expertise, which can undermine reproducibility, transparency, and overall research software quality. This work proposes a structured and evidence-driven approach to support AI model selection that aligns with both technical and contextual requirements. We conceptualize AI model selection as a Multi-Criteria Decision-Making (MCDM) problem and introduce an evidence-based decision-support framework that integrates automated data collection pipelines, a structured knowledge graph, and MCDM principles. Following the Design Science Research methodology, the proposed framework (ModelSelect) is empirically validated through 50 real-world case studies and comparative experiments against leading generative AI systems. The evaluation results show that ModelSelect produces reliable, interpretable, and reproducible recommendations that closely align with expert reasoning. Across the case studies, the framework achieved high coverage and strong rationale alignment in both model and library recommendation tasks, performing comparably to generative AI assistants while offering superior traceability and consistency. By framing AI model selection as an MCDM problem, this work establishes a rigorous foundation for transparent and reproducible decision support in research software engineering. The proposed framework provides a scalable and explainable pathway for integrating empirical evidence into AI model recommendation processes, ultimately improving the quality and robustness of research software decision-making.

Siamak Farshidi, Amir Saberhabibi, Behbod Eskafi et al.

Selecting third-party software packages in open-source ecosystems like Python is challenging due to the large number of alternatives and limited transparent evidence for comparison. Generative AI tools are increasingly used in development workflows, but their suggestions often overlook dependency evaluation, emphasize popularity over suitability, and lack reproducibility. This creates risks for projects that require transparency, long-term reliability, maintainability, and informed architectural decisions. This study formulates software package selection as a Multi-Criteria Decision-Making (MCDM) problem and proposes a data-driven framework for technology evaluation. Automated data pipelines continuously collect and integrate software metadata, usage trends, vulnerability information, and developer sentiment from GitHub, PyPI, and Stack Overflow. These data are structured into a decision model representing relationships among packages, domain features, and quality attributes. The framework is implemented in PySelect, a decision support system that uses large language models to interpret user intent and query the model to identify contextually appropriate packages. The approach is evaluated using 798,669 Python scripts from 16,887 GitHub repositories and a user study based on the Technology Acceptance Model. Results show high data extraction precision, improved recommendation quality over generative AI baselines, and positive user evaluations of usefulness and ease of use. This work introduces a scalable, interpretable, and reproducible framework that supports evidence-based software selection using MCDM principles, empirical data, and AI-assisted intent modeling.

Ferhat Erata, Claire Gardent, Bikash Gyawali et al. · amazon-science

The ModelWriter platform provides a generic framework for automated traceability analysis. In this paper, we demonstrate how this framework can be used to trace the consistency and completeness of technical documents that consist of a set of System Installation Design Principles used by Airbus to ensure the correctness of aircraft system installation. We show in particular, how the platform allows the integration of two types of reasoning: reasoning about the meaning of text using semantic parsing and description logic theorem proving; and reasoning about document structure using first-order relational logic and finite model finding for traceability analysis.

Bedir Tekinerdogan, Kaan Özcan, Sevil Yağız et al.

Many systems nowadays require protection against security or safety threats. A physical protection system (PPS) integrates people, procedures, and equipment to protect assets or facilities. PPSs have targeted various systems, including airports, rail transport, highways, hospitals, bridges, the electricity grid, dams, power plants, seaports, oil refineries, and water systems. Hence, PPSs are characterized by a broad set of features, from which part is common, while other features are variant and depend on the particular system to be developed. The notion of PPS has been broadly addressed in the literature, and even domain-specific PPS development methods have been proposed. However, the common and variant features are fragmented across many studies. This situation seriously impedes the identification of the required features and likewise the guidance of the systems engineering process of PPSs. To enhance the understanding and support the guidance of the development of PPS, in this paper, we provide a feature-driven survey of PPSs. The approach applies a systematic domain analysis process based on the state-of-the-art of PPSs. It presents a family feature model that defines the common and variant features and herewith the configuration space of PPSs

Bedir Tekinerdogan, Namik Aktekin

We report on our experiences in an e-government project for supporting the automatic generation of E-forms for services provided by local governments. The approach requires the integration of both the model-based user interface design (MBUID) and software product line engineering approaches. During the domain engineering activity the commonality and variability of product services is modeled using feature diagrams and the corresponding UI models are defined. To support the automation of e-forms the implemented feature models are on their turn used to generate E-forms automatically to enhance productivity, increase quality and reduce cost of development. We have developed three different approaches for e-form generation in increasing complexity: (1) offline model transformation without interaction (2) model transformation with initial interaction (3) model-transformation with run-time interaction. We discuss the lessons learned and propose a systematic approach for defining model transformations that is based on an interactive paradigm.

Murat Acar, Bedir Tekinerdogan

Context: User assistance is generally defined as the guided assistance to a user of a software system in order to help accomplish tasks and enhance user experience. Automated user assistance systems are equipped with online help system that provides information to the user in an electronic format and which can be opened directly in the application. Various different automated user assistance approaches have been proposed in the literature. However, there has been no attempt to systematically review and report the impact of automated user assistance systems. Objective: The overall objective of this systematic review is to identify the state of art in automated user assistance systems, and describe the reported evidence for automated user assistance. Method: A systematic literature review is conducted by a multiphase study selection process using the published literature since 2002. Results: We reviewed 575 papers that are discovered using a well-planned review protocol, and 31 of them were assessed as primary studies related to our research questions. Conclusions: Our study shows that user assistance systems can provide important benefits for the user but still more research is required in this domain.

Vahid Garousi, Ehsan Abbasi, Bedir Tekinerdogan

Development of scientific and engineering software is usually different and could be more challenging than the development of conventional enterprise software. The authors were involved in a technology-transfer project between academia and industry which focused on engineering, development and testing of a software for optimization of pumping energy costs for oil pipelines. Experts with different skillsets (mechanical, power and software engineers) were involved. Given the complex nature of the software (a sophisticated underlying optimization model) and having experts from different fields, there were challenges in various software engineering aspects of the software system (e.g., requirements and testing). We report our observations and experience in addressing those challenges during our technology-transfer project, and aim to add to the existing body of experience and evidence in engineering of scientific and engineering software. We believe that our observations, experience and lessons learnt could be useful for other researchers and practitioners in engineering of other scientific and engineering software systems.