AbstractAs the quantitative hazard research, particularly stemming from the engineering fields, aims to move from component- and building-level modeling into the interdisciplinary space of community-level modeling for resilience, the need to test, verify, and validate community resilience algorithms becomes a critical challenge; virtual testbeds are an effective tool for such purposes. We define a virtual testbed as an environment with enough supporting architecture and metadata to be representative of one or more systems such that the testbed can be used to design experiments, examine model or system integration, and test theories. Testbeds enable researchers to assess multidisciplinary integrated community resilience models, thereby helping decision makers to make better community hazard mitigation plans and recovery decisions. This paper leverages the current momentum on using virtual testbeds for community resilience analysis to dissect what testbeds are in practice. To obtain consensus on the presented definition of a testbed, the paper conducted a virtual survey with testbed experts. The survey primarily explored how testbeds have been used across different disciplines, how testbeds differ from case studies, and what are the minimum requirements for a testbed. The paper, then, presents findings from a systematic literature review on 22 identified existing community resilience testbeds and 103 associated publications. According to the literature review and survey results, community resilience testbeds should have both a hazard module and a community module that ideally includes physical, social, and economic systems. The literature review concludes with a discussion on the available tools for testbed development, typical challenges testbed developers encounter, and areas for future testbed research. The availability of existing testbeds for reuse by other researchers, standardization of the development and publication process of new testbeds including obtaining, cleaning, and validating the required data, and verification of numerical algorithms are the main detected issues that need to be addressed in future research.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *