20 years of computational science: Selected papers from 2020 International Conference on Computational Science

Computational science [1] is currently developing as a large multidisciplinary scientific direction where multiple areas are joined together. Initially appeared at the intersection of computer science, information technologies, and mathematical modeling, it is providing new methods and tools for researchers in many fields, from traditional natural sciences to new applications in medicine, social sciences, and humanities. Moreover, computational science is aimed not only toward building solutions to existing domain problems. It provides new approaches and opens new directions in the research, industry, regulation, etc. (see, e.g., Simulation-Based Engineering Science [2]). Computational science combines two principal scientific approaches organized around a computational experiment. First, the computational experiment is designed and developed in accordance with the domain knowledge and problem definition, forming a deductive or knowledgebased inference of models and computational solutions. Second, the data-driven approaches and methods can be considered as inductive (empirical) inference. The combination of these approaches with wide availability of computational resources gives computational science a boost to develop new and adapt existing general-purpose concepts and technologies. Initially, the research field was based on the intersection of mathematical modeling algorithms (including numerical simulation) and computational-intensive solutions (including high-performance, distributed, hybrid, etc. computing). The development of technologies provides new ways to resolve the existing issues in the area. For example, there is a large amount of data collected through observation, measurement, or as a result of earlier modeling and simulation. These data provide a significant source for scientific discovery, and some authors consider this as a new paradigm (data-intensive scientific discovery) [3]. Together with BigData concepts as a technological backbone, these approaches took their place in the computational science area [4]. Artificial intelligence and machine learning methods recently attracted significant attention within the scientific society and can be considered as another example of such an extension. Within the domain of computational science, these methods may be used for a) management of complex models; b) substitute of computationally-intensive models; c) exploration or interpolation of model parameters and data; d) prediction of model characteristics (including performance, uncertainty, etc.). An example of a powerful such method is surrogate modeling [5, 6]. One of the important application areas of computational science is multidisciplinary studies, where a combination of models may provide a comprehensive view of the nature of systems and phenomena. For example, the concept of system-level science [7] considers a holistic description of a system to provide the ability for analysis and computational experiments with different goals using the same solution. Moreover, with the help of computational science, many classes of systems could be described through modeling and simulation, including complex systems [8,9] and global systems [10]. The International Conference on Computational Science (ICCS)1 brings together researchers and scientists working in fundamental computer science disciplines and in various application areas, who are pioneering computational methods in sciences such as physics, chemistry, life sciences, and engineering, as well as in arts and humanities. Since its inception in 2001, ICCS forms a space where the problem domains, IT, and modeling join together to discuss the present and future research directions. With the 20th ICCS, we were celebrating the 20 years of highly successful conferences and an active society built around the conference. During the past years, the conference was hosted by a variety of institutions and cities in 12 countries across the globe: Australia, China, Iceland, Poland, Portugal, Russia, the Netherlands, Singapore, Spain, Switzerland, UK, USA. The conference was always focused on the recent advances in computational science. Analysis of ICCS topics evolved through its history [11] shows that a significant amount of the works presented in the conference are concentrated around key sub-areas of computational science, including modeling and simulation, high-performance and distributed computing, and numerical methods. Moreover, ICCS reacts to emergent technologies and approaches like the development of GPGPU or IPv6 technologies, which was followed by the growing presence of works in these respective areas. The ICCS society always attracts both well-known scientists and young researchers. During the previous years, more than 120 (!) leading scientists in the area gave invited talks, tutorials, and lectures as a valuable contribution to the conference. Twelve of them contributed to the conference more than once, bringing the evolution of their ideas to empower ICCS content. It was our pleasure to announce a special issue of the Journal of Computational Science with 12 selected papers prepared by the leading scientists in the area (acting as keynote speakers during ICCS history) and their colleagues, reflecting the vision of issues, recent advances, challenges, and solutions in various sub-areas [12].