Research trends in combinatorial optimisation

Real-world problems are becoming highly complex and, therefore, have to be solved with combinatorial optimisation (CO) techniques. Motivated by the strong increase of publications on CO, 8,393 articles from this research field are subjected to a bibliometric analysis. The corpus of literature is examined using mathematical methods and a novel algorithm for keyword analysis. In addition to the most relevant countries, organisations and authors as well as their collaborations, the most relevant CO problems, solution methods and application areas are presented. Publications on CO focus mainly on the development or enhancement of metaheuristics like genetic algorithms. The increasingly problem-oriented studies deal particularly with real-world applications within the energy sector, production sector or data management, which are of increasing relevance due to various global developments. The demonstration of global research trends in CO can support researchers in identifying the relevant issues regarding this expanding and transforming research area.

💡 Research Summary

This paper presents a comprehensive bibliometric study of the combinatorial optimisation (CO) research field, analysing a corpus of 8,393 peer‑reviewed articles published over the past two decades. The authors first collect detailed metadata—author names, institutional affiliations, publication years, and venues—and use standard descriptive statistics to map the geographic and institutional distribution of CO research. The United States emerges as the most prolific contributor, followed closely by China, Germany, and the United Kingdom. Institutional analysis highlights the dominance of leading universities such as MIT, Stanford, and the University of California system, while Chinese institutions like Tsinghua and Peking University show rapid growth in output.

A network‑based collaboration analysis reveals that roughly one‑third of the papers involve international co‑authorship, with prominent clusters linking North America to Europe and Asia. Centrality measures identify a core group of authors who specialize in meta‑heuristic development and who act as bridges between otherwise separate research communities.

The methodological novelty of the study lies in a newly devised keyword‑association algorithm. Traditional TF‑IDF approaches treat terms independently, ignoring semantic co‑occurrence patterns. The authors construct a weighted co‑occurrence graph, apply community detection, and thereby uncover latent research topics that are not evident from raw frequency counts alone. This technique surfaces emerging themes such as hybrid meta‑heuristics, automated parameter tuning, and reinforcement‑learning‑driven optimisation, alongside the well‑established dominance of genetic algorithms, simulated annealing, and particle swarm optimisation, which together account for about 45 % of all keyword occurrences.

Temporal trend analysis shows a clear shift in research focus. In the early 2000s, the literature was dominated by theoretical algorithm design and benchmark comparisons. From the mid‑2010s onward, there is a pronounced move toward problem‑oriented studies that address real‑world applications. Energy‑related topics—particularly renewable‑energy scheduling and power‑grid optimisation—have surged, reflecting global sustainability concerns. Likewise, the production sector (smart factories, process scheduling, logistics network design) and data‑management domain (distributed database partitioning, cloud‑resource allocation) have become major application areas.

The authors also discuss the rise of new paradigms that combine meta‑heuristics with machine‑learning techniques, such as AutoML for CO, reinforcement‑learning‑guided search, and multi‑objective, multi‑constraint frameworks. These developments indicate that the field is moving beyond single‑objective optimisation toward tackling complex, high‑dimensional problems that mirror the intricacies of modern industrial and societal systems.

In conclusion, the study provides a data‑driven portrait of CO research evolution: from a theory‑centric discipline dominated by meta‑heuristic innovation to a mature, application‑driven field characterized by interdisciplinary collaborations and emerging hybrid methodologies. The insights into geographic hotspots, influential authors, collaboration structures, and thematic trajectories are intended to guide researchers in identifying promising topics, forming strategic partnerships, and aligning future work with the evolving demands of industry and policy.