Kriging-based simulation optimization: An emergency medical system application

Metamodel-Based Simulation Optimization: A Systematic Literature Review

“…Several authors have used SED for problems related to the secondary and tertiary economic sectors, like inventory controls (e.g., Dengiz et al, 2009;Quan et al, 2015), shop floor planning (e.g., Dunke and Nickel, 2020;Greinacher et al, 2020), allocating and sizing resources (Aydin et al, 2018;Coelho and Pinto 2018;Cuckler et al, 2017), and production scheduling (Chang, 2016;Hassannayebi et al, 2019). No SED metamodeling application was conducted for the primary economic sector, perhaps because this type of simulation is commonly taught in specific engineering courses, limiting this tool's dissemination.…”

“…This algorithm, which is responsible for recursive and adaptive learning, is referred to as Efficient Global Optimization (EGO) in most studies. The choice for the point(s) that should be added to each interaction is usually made using the Expected Improvement (EI) algorithm (Coelho and Pinto 2018;Pedrielli et al, 2020;Quan et al, 2013;Rojas-Gonzalez et al, 2020;Wang et al 2018). Table 6 shows the main software programs used, and performance metrics found in literature.…”

“…However, cross-validation techniques were used in most studies, and 20 (16.4%), of all authors used leave-one-out cross-validation (e.g., Azizi et al, 2019;Xiqun et al, 2019), and 15 (12.3%), used k-fold cross-validation, (e.g., Coelho and Pinto, 2018;Hou et al, 2019;Wade, 2019). For a more detailed explanation of cross-validation techniques see Bergmeir and Benitez (2012).…”

“…Examples include applications related to the automotive (Fu and Sahin, 2004;Ivanova and Kuhnt, 2014;Stork et al, 2008), energy (Sharif and Hammad 2019;Storti et al, 2019), hydropower (Bananmah et al, 2020;Mooselu et al, 2019), and transit infrastructure (Yin et al, 2016) sectors. Many authors have also studied MBSO applied to resource allocation problems (e.g., (Coelho and Pinto 2018;Song et al, 2005;Yousefi and Yousefi 2019;Zeinali et al, 2015), to synthetic mathematical functions (e.g., Baquela and Olivera, 2019;Kim and Boukouvala, 2020;Gonzalez et al, 2020;Wang et al, 2020), and to urban traffic issues Chong and Osorio 2018;Osorio and Bierlaire 2013;Osorio and Chong 2015).…”

“…The global optimization method is another technique that has shown good results for building metamodels, which uses EGO and EI techniques to build adaptive and interactive metamodels. Kriging was mostly used as a metamodeling algorithm for applications involving this strategy, e.g., (Coelho and Pinto, 2018;Ivanova and Kuhnt, 2014;Wang et al, 2020). We recommend that future research explore this adaptive strategy for building metamodels based on other algorithms, like ANNs, SVM, Extreme Learning Machines, decision trees, etc.…”