<![CDATA[This study maps the development, collaboration patterns, and conceptual structure of research connecting Problem-Based Learning (PBL), computational methods, and artificial intelligence (AI) in physics-related education. A descriptive bibliometric design was applied to 251 English-language journal articles and conference papers indexed in Scopus between 2016 and 2025. The dataset was retrieved on January 16, 2026, and analyzed using Microsoft Excel and VOSviewer 1.6.20 through performance analysis, country-level collaboration mapping, keyword co-occurrence analysis, and temporal overlay visualization. Publication output showed an overall upward trend with annual fluctuations, increasing from 10 documents in 2016 and reaching a peak of 36 documents in 2023. The reported citation counts also increased across the study period. The United States led in publication output, citation count, and collaboration strength, followed by China in productivity, while Indonesia ranked third in publication output. Of 2,330 identified keywords, 96 met the minimum occurrence threshold and formed four clusters. Problem-based learning was the most prominent keyword, with 143 occurrences and a total link strength of 999, and was strongly connected with students, simulations, curriculum, and active learning. Artificial intelligence and machine learning each recorded 30 occurrences and comparatively recent average publication years, indicating their emerging position within the network. However, their connections with PBL remained weaker than those of established simulation-based themes. The presence of medical, engineering, and technology-enhanced education terms further indicates that the corpus was interdisciplinary rather than restricted exclusively to physics education. These findings provide a structured basis for future qualitative and empirical research on the effectiveness, design, and responsible implementation of AI-supported PBL environments.]]>
Copyrights © 2026
