key: cord-0794485-f0mkvfzu authors: Tordecilla, Rafael D.; Juan, Angel A.; Montoya-Torres, Jairo R.; Quintero-Araujo, Carlos L.; Panadero, Javier title: Simulation-Optimization Methods for Designing and Assessing Resilient Supply Chain Networks under Uncertainty Scenarios: A Review date: 2020-08-11 journal: Simul Model Pract Theory DOI: 10.1016/j.simpat.2020.102166 sha: 62a7bfb2667dceb92ecf64df98f5ac8f0151a1de doc_id: 794485 cord_uid: f0mkvfzu The design of supply chain networks (SCNs) aims at determining the number, location, and capacity of production facilities, as well as the allocation of markets (customers) and suppliers to one or more of these facilities. This paper reviews the existing literature on the use of simulation-optimization methods in the design of resilient SCNs. From this review, we classify some of the many works in the topic according to factors such as their methodology, the approach they use to deal with uncertainty and risk, etc. The paper also identifies several research opportunities, such as the inclusion of multiple criteria (e.g., monetary, environmental, and social dimensions) during the design-optimization process and the convenience of considering hybrid approaches combining metaheuristic algorithms, simulation, and machine learning methods to account for uncertainty and dynamic conditions, respectively. A supply chain network (SCN) is a typical example of a complex and largescale system. Bidhandi et al. [11] define it as a network of suppliers, manufacturing plants, warehouses, and distribution channels organized to acquire raw materials, convert these raw materials into finished products, and distribute 5 these products among customers. Many decisions must be made in such a complex system in order to guarantee a good performance. However, the more complex a system is, the more imprecise or inexact is the information available to characterize it and, therefore, the greater the uncertainty level [15] . Supply chain network design (SCND) is a concept broadly studied during 10 the last decades, both from a qualitative and a quantitative perspective. Authors have referred to it by using the terms supply chain design and supply chain network design. Carvalho et al. [23] state that a SCND problem "comprises the decisions regarding the number and location of production facilities, the amount of capacity at each facility, the assignment of each market region to one or more 15 locations, and supplier selection for sub-assemblies, components and materials". These decisions are related to a strategic level, and must be optimized considering a long-term (usually several years) efficient operation of the supply chain as a whole [6] . One of the more challenging responsibilities in SCND is addressing uncertainty. Anticipating the future is crucial in planning and design processes. 20 However, the future conditions of the business environment is generally difficult to predict. Blackhurst et al. [13] state that one of the causes of SCNs complexity is their dynamic nature and the uncertainty in variables such as demand, capacities, transportation times, or manufacturing times. In recent years, a trend in the literature has been the consideration of re- 25 silience for designing and assessing SCNs in order to face uncertainty. Christopher & Peck [27] define resilience as "the ability of a system to return to its original state or move to a new, more desirable state after being disturbed". Similar definitions can be found in fields different to SCND, such as ecology, psychology and economy [122] , or natural disasters risks mitigation and adap- 30 tation in urban systems [63] . For instance, a concept from earthquake studies is given by [17] , who state that "seismic resilience is the ability of both physical and social systems to withstand earthquake-generated forces and demands and to cope with earthquake impacts through situation assessment, rapid response, and effective recovery strategies." 35 Addressing resilience from the civil infrastructure point of view is very usual in engineering. For instance, in order to design and assess this type of systems, [14] propose a unified framework integrating resilience and sustainability concepts. [12] present a probabilistic approach to assess the lifetime of concrete structures seismic resilience. The joint effects of seismic and environmental -40 e.g., corrosion -hazards are studied. Applications to a concrete frame building and a continuous bridge are considered. Bridges and bridge networks are also considered by [3] , who assess the effects of earthquakes and other independent and interacting hazards in the resilience of this type of structures. These authors highlight that the most recent research has focused on studying the civil 45 infrastructure as a connected system instead of individual components. It is relevant to highlight that although the civil infrastructure is a very important part of a supply chain, it is not the only one subject to uncertainty and risks, as we will study in this paper. For instance, after a disruptive event, the recovery of a supply chain takes more time than the infrastructure restoration [114] , 50 given the multiple and different components of a SCN. Hence, the position of our review is more holistic. Resilient SCND has been a topic able to attract the attention of researchers, specially when trends such as leanness and globalization have increased the risks that supply chains must face. Regarding leanness, it makes SCNs more 55 vulnerable due to the reduction or even removal of redundancies [10] . Regarding globalization, the increasing complexity of SCNs in a globalized world causes higher uncertainty [66] . Moreover, globalization increases supply chain vulnerabilities [38] . Expanding globally a supply chain raises the likelihood of facing new risks that might not exist in a local range. For instance, a natural disaster 60 such as the 2011 earthquake in Japan, which triggered a tsunami and a nuclear crisis, affected many global companies like those in the silicon wafers industry. Since 60% of silicon wafers world demand were supplied by Japan [118] , this product availability decreased considerably. The same disaster affected also all Toyota factories. Although most of them were not directly affected, a two-week 65 shutdown was caused by disruptions in the components supply, given the Toyota's lean production planning [53] . Human-induced disasters are also a source of disturbances for supply chains, either they are deliberate (e.g., terrorist attacks) or caused by involuntary mistakes or negligence (e.g., the 2010 oil spill in the Gulf of Mexico), as described in [128] . These examples show the relevance of 70 considering resilience aspects when designing and assessing supply chains, since they need to recover successfully after the occurrence of such disruptive events. The terms risk and vulnerability are closely related to resilience. Carvalho et al. [23] relate supply chain vulnerability to the incapacity of a SCN to react to disturbances. More exactly, Heckmann et al. [65] define supply chain 75 vulnerability as "the extent to which a supply chain is susceptible to a specific or unspecific risk event". Here, the disturbance concept is similar to the risk concept, being this a primary term previous to vulnerability. Peck [120] defines supply chain risk as "anything that disrupts or impedes the information, material or product flows from original suppliers to the delivery of the final product 80 to the ultimate end-user". Therefore, the more resilient a SCN is, the lower its vulnerability to risks [127] . A review about the use of quantitative approaches in supply chain risk management is carried out by [116] . They perform a Systematic Literature Review (SLR) to analyze and synthesize the contribution of simulation and optimization methods in this field. Moreover, when risks cause 85 a disruption in a few nodes, their effects can easily spread to other parts of the supply chain. This phenomenon is known as the ripple effect [98] . According to [39] , the ripple effect causes lower revenues, delivery delays, loss of market share and reputation, as well as stock return decreases, hence affecting the global performance of the supply chain. 90 Epidemic outbreaks are a very special case of SCN risks characterized by a long-term disruption, disruption propagation (i.e., the ripple effect), and high uncertainty due to simultaneous disruptions in supply, demand, and logistics infrastructure [71] . Particularly in 2020, the global pandemic caused by the COVID-19 disease has largely affected all areas of the economy and society 95 worldwide. Some supply chains have experienced an increase of demand that they are not able to satisfy (facial masks, hand sanitizer, ventilators, etc.), while others are suffering long-time production stops like the ones of non-essential products. These companies are in danger of bankruptcies and needing help from governments. As pointed out by Ivanov & Dolgui [72] , supply availability 100 in global supply chains has been largely decreased and imbalanced with the demands. Thus, this pandemic is an unprecedented and extraordinary situation that clearly shows the need for advancing in research and practices of SCN resilience. In addition, new concepts related to resilience, such as supply chain survivability, are emerging in the literature. 105 In logistics and supply chain management, quantitative approaches are mainly classified into two groups: optimization and simulation, which are mostly used independently to address uncertainty -e.g., see [56] and [145] for each group, respectively. However, given the growth in computational power, the use of hybrid simulation-optimization (sim-opt) methods has increased in recent years 110 [82] in order to combine the most important advantages of both worlds, mainly because of its suitability to address uncertainty [26] . Nevertheless, in the more specific topic of SCND, applications of hybrid sim-opt methods are still scarce and, to the best of our knowledge, it is almost nonexistent in SCND resilience. In regard to existing review articles about this topic, most of them still address 115 conceptual papers, which shows the relevance of carrying out a review analyzing papers following a quantitative approach. Accordingly, this work provides a review that synthesizes the main studies related to quantitative SCND resilience, as well as to the sim-opt methods employed for that. Moreover, the paper also highlights some open challenges that need to be addressed by the 120 sim-opt community. The remaining of this document is organized as follows: Section 2 extends the motivation of this paper by explaining the extent of previous literature reviews addressing SCND resilience and sim-opt methods in this field. Section 3 explains the research methodology employed to carry out this review. The findings of this 125 paper are presented in Section 4, where discussions on relevant works are also presented in Subsections 4.1 and 4.2, respectively. Section 5 provides insights and future research directions by analyzing how emerging hybrid methods can be useful for designing resilient SCNs under uncertainty or dynamic conditions, as well as the concept of 'agile' SCND. The paper ends in Section 6 by outlining 130 some concluding remarks. To the best of our knowledge, there are no published literature reviews that combine sim-opt methods with SCN resilience. A review by Pourhejazy & Kwon [123] highlights the use of sim-opt frameworks as a growing research area. In-135 tegrated problems such as location-routing, inventory-routing, and locationinventory are analyzed, and sim-opt applications are studied. Finally, the authors analyze papers addressing sustainability issues, concluding that this is a relevant trend along with resilience. Therefore, they state that sim-opt frameworks are the main tool to design and manage SCNs. However, despite the 140 relevance these authors give to resilience, they do not analyze papers considering this dimension. A total of 19 review papers discussing the concept of "supply chain resilience" were found since 2015. Only one review previous to 2015 was found [122] . Here, the authors relate the SCN resilience concept to traditional re-145 silience concepts from the ecological, social, psychological, and economic fields. Most of these works are conceptual, i.e., they discuss about resilience and some related concepts. These conceptual papers help to clarify terms in order to have a better understanding on the topic. For instance, Zhao et al. [167] carry out a systematic review in which risk sources and resilience factors in agri-food 150 supply chains are identified. In addition, particular parameters that can affect this type of SCNs are presented. A systematic review by Stone & Rahimifard [146] includes 137 articles from different fields such as engineering, operations management, ecology, and social sciences in order to identify definitions, elements, and strategies that can be relevant for resilience in agri-food SCNs. 155 More recently, [52] perform also a review from a multidisciplinary point of view (including supply chain management, information systems, psychology, among others) to establish differences and similarities between agility and resilience concepts. Tukamuhabwa et al. [153] reviewed 91 papers related to SCN resilience, 160 showing that most studies (43%) are conceptual or theoretical, and 36% of them adopt modeling approaches. Several definitions of "resilience" are provided, as well as proactive and reactive strategies for building resilient SCNs. Associated concepts like flexibility, redundancy, collaboration, and agility are identified. Little attention is paid by these authors to modeling articles. Such 165 concepts and others related to SCN resilience are also identified by [139] , who call them "formative elements". These are compared with characteristics of high reliability organizations. Alternatively, [126] call these concepts as "key capabilities". They identify 4 of them: flexibility, velocity, visibility and collaboration, as well as 13 attributes related to SCN resilience. [140] expand such 170 concepts by identifying resilience-related barriers, metrics and strategies. These are presented into a unified framework after analyzing 125 papers. Additional topics are analyzed by [5] , who combine SLR with VOSviewer Co-occurrence Analysis to identify a set of drivers, barriers, theories, moderators, mediators and research methods in SCN resilience. A systematic review of 67 papers is carried out by Hohenstein et al. [66] . Here, many SCN resilience definitions are presented. The quantitative approach is only addressed by analyzing some papers regarding how to measure resilient designs. Kochan recover, and to learn). Nevertheless, its most important contribution is the conceptual synthesis, which is performed via a holistic model. Wang et al. [156] state the importance of analyzing SCN definitions besides the usual resiliencerelated ones. Then, a review of studies that apply resilience to supply chain management is provided. The authors conclude that the SCN resilience con- and [67] . These authors' remarks establish that most of their reviewed papers are conceptual and, therefore, there is a lack regarding the use of quantitative models for addressing SCN resilience. For instance, Ivanov et al. [74] provide a review about both disruptions and recovery in supply chain design and planning. Their perspective is to show papers that use quantitative tools regarding disruptions risks (natural or human-induced disasters, strikes, etc.), by differentiating 220 these from operational risks (produced by uncertainty in demand, lead-time, or any other business-related variables). The authors dedicate a few paragraphs to papers that address sustainability. Dolgui et al. [39] review papers addressing the ripple effect in the supply chain through a quantitative approach. Quantitative tools such as mathemati- Notice that both conceptual and quantitative reviews are quite recent. This 235 shows a growing interest in the SCN resilience topic. Moreover, although the papers by [67] , Dolgui et al. [39] and Ivanov et al. [74] address the resilience concept along them, the focus is different than the one employed in our review. For instance, the main topic of the review by Dolgui et al. [39] is not resilience but the ripple effect. Furthermore, we define a taxonomy more exhaustive and 240 explicit than that used by [67] and Ivanov et al. [74] . For example, the latter shows a clear focus in disruption risks, which we extend by including operational risks in our analysis. Finally, we address some aspects that these authors do not consider explicitly, such as the uncertain parameters analysis or our exposition of papers tackling real-world cases. Regarding the article by Povoa [132], they analyze all decisions levels (strategic, tactical, and operational) from only 39 papers, which shows the need of studying more thoroughly each of these levels. This is an additional contribution of our work. An important branch of resilience studies is the metrics used to assess the supply chain performance. Measuring resilience becomes relevant not only to 250 design new supply chains, but also to evaluate an already operational SCN. In this case, resilience metrics are useful for decision-makers to implement strategies that increase resilience at minimal cost. Both quantitative and qualitative indicators can be found. For instance, [22] propose a mixed-integer linear programming model to design a resilient supply chain. The network performance 255 is assessed through 11 quantitative indicators, which are classified into three types: network design, network centralization and operational indicators. A real-world European supply chain is considered as a case study. [147] perform a review about resilience metrics from the transportation infrastructure point of view. The functionality of these networks is taken as a core to define two types 260 of resilience metrics: functionality-related and socioeconomic metrics. From a qualitative point of view, [143] present a study in which 10 resilience enablers are identified to design a unique supply chain resilience index through a graph theory model. Data from both a literature review and surveys answered by Indian firms are used to identify the enablers. Agility, collaboration, supply 265 chain structure, among others, are identified as enablers. [142] perform an SLR where 17 performance indicators are identified. The supply chain network design is presented as one of these indicators, as well as agility, collaboration and others. Authors divide the SCN resilience into 3 phases: anticipation, resistance, and response and recovery. Notice in all these cited papers that authors highlight 270 the supply chain network design as an important factor to assess resilience, i.e., resilience can not be studied properly without considering design and long-term decisions. This fact shows the relevance of our review. A key requirement in a literature review is that each stage of the process has 275 to be defined in a protocol "intended to guide the whole review, thereby reducing the possible sources of bias which arise from authors making idiosyncratic decisions at different stages of the review process" [9] . Our review methodology is based on the systematic literature review (SLR) approach introduced by Denyer & Tranfield [36] , which is characterized by "its distinct and exact-280 ing principles", its replicability, transparency, and robustness to produce solid and reliable evidence. In addition, reviews related to resilience in SCNs (see Section 2) have been carried out mostly using an SLR methodology. The SLR steps are: (i) question formulation; (ii) location of studies; (iii) study selection and evaluation; (iv) analysis and synthesis; and (v) reporting 285 and use of results. Firstly, it is important to formulate suitable questions to delimit the research and avoid to search topics that do not fit into the reviews' objectives. Therefore, a general question was formulated as: How sim-opt methods have been used to design resilient SCNs? To answer this general question, the following specific queries were formulated: Web of Science, as well as papers found in the ScienceDirect database, and 300 in the Google Scholar search engine were collected. The search was conducted from a more general combination of terms to a more specific one, namely: • resilien* AND supply AND chain • resilien* AND supply AND chain AND network AND design • resilien* AND supply AND chain AND network AND design AND math-305 ematical AND model • resilien* AND supply AND chain AND network AND design AND optimi?ation • resilien* AND supply AND chain AND network AND design AND simulation 310 • resilien* AND supply AND chain AND network AND design AND optimi?ation AND simulation Notice that the search term "optimi?ation" was employed in order to capture both "optimisation" in British English and "optimization" in American English. The search was carried out in the fields title, abstract, and keywords. The term 315 network was then suppressed to make additional search, since some papers do not use such a word. Mainly qualitative studies were found, even after using the terms mathematical model, optimi?ation, and simulation. These key terms were mainly used to search for review papers already analyzed in Section 2. In this regard, an additional search was conducted by limiting the document 320 types to reviews. In addition, papers previous to the year 2000 were excluded, as well as those studies that do not tackle strategic decisions, according to the definition by Ghiani et al. [49] . However, when using the last combination of terms just a few papers were found. Some of them had already been collected, and others were not coherent with our research questions. For this reason, a 325 sixth research question without including the term resilien* was formulated, namely: How sim-opt methods have been used to design SCNs? Then, a new search was conducted by using the next additional terms: • supply AND chain AND design AND simulation AND optimi?ation • supply AND chain AND network AND design AND simulation AND op-330 timi?ation As a result of this last search, papers not including the combination of optimization and simulation techniques were excluded. In this way, we increased our scope and it was possible to establish a more complete framework. A total of 163 articles were short-listed by considering the aforementioned criteria. From these, 93 papers are related to SCND resilience, 49 to sim-opt methods in SCND, and 21 articles are reviews (all of them already analyzed in Section 2). The short-listed papers were organized in a spreadsheet, where basic information about the papers was registered: title, authors, year, and journal. Also, after an initial review, a taxonomy was built to analyze and synthesize SCND 340 resilience and sim-opt papers, namely: • Mathematical approach: this refers to the method used to model the problem, e.g., robust optimization, stochastic programming, etc. • Solving approach: this refers to the method employed by the authors to solve the proposed model, e.g., exact methods, metaheuristics, etc. • Uncertain parameters: in a model, it is usual to consider some parameters as uncertain and others as known. Common uncertain parameters are: demand, cost, capacity, etc. • Uncertainty approach: this refers to the way the authors model the uncertain parameters, e.g., by means of probability distributions, fuzzy sets, 350 etc. • Objective criterion: variables like cost, income, or profit are usually minimized or maximized in mathematical models when designing SCNs. In addition, it is possible that several conflicting objectives are considered, which leads to adapt the model and its solving approach. • Supply chain design special case: some special cases regarding SCND can be found in the literature. These are conceptual models based on real-world characteristics that influence the design of a supply chain considering criteria that are specific of that model. These criteria can be: objectives, constraints, variables, or parameters. For instance, the sustainable 360 SCND is a special case that not only considers the traditional economic goal but also environmental and social objectives. Other special cases found in the literature are: green SCND, closed-loop SCND, etc. • Application to a real-world case: this refers to the fact that the problem and its solution have been applied to a real-life case; otherwise the paper 365 is classified as a theoretical contribution. Besides, when analyzing SCND resilience papers, the additional criterion type of risk is considered to build the taxonomy, which refers to operational risks, disruption risks, or both [149] . Then, a deeper review of the full text of papers was carried out, after which 42 papers (25 related to SCND resilience and The taxonomy previously defined has been used to analyze and synthesize 68 short-listed papers. Table 1 shows that most papers use stochastic programming as a modeling tool for the supply chain, followed by mixed-integer linear pro- Heuristic and metaheuristic methods are used very scarcely in SCND resilience ( and flexibility strategies are assessed to mitigate the effects of "supply delay" disturbance. The use of scenarios is the most frequent uncertainty approach (Table 3) Table 5 shows the model objective criterion. Both optimization (minimization and maximization) and non-optimization criteria are considered. The latter one refers especially to simulation models in which the objectives are the model output variables. In this case, measuring resilience is the most used objective. In general, the most utilized criterion is cost minimization (53% of papers), Some researchers also address environmental aspects, such as CO 2 emissions or 545 environmental impact, as well as social impact. Also, 25 out of 68 references optimize multiple objectives, which are identified with an asterisk in Table 5 . Notice that neither resilience nor de-resiliency are never a unique objective, i.e., they are always optimized along with cost. This shows that, despite resilience may have a relative relevance in an explicit 550 objective, it should not be considered alone due to the high cost that a resilient design may lead to. Papers that employ multiple objectives must use some procedure in order to address such multiplicity. Khalili et al. [87] are the only authors who employ the Reservation Level-driven Tchebycheff Procedure (RLTP), and explain why both the weighted sum method and the ε-constraint method 555 are not appropriate. However, evidence shows that these methods are preferred when considering multiple objectives, as well as the use of the LP-metric method or Compromise programming (Table 6 ). For example, Jabbarzadeh et al. [76] propose a bi-objective model that seeks to minimize expected total cost and to maximize expected sustainability performance. A set of scenarios with a proba-560 bility of occurrence is defined and the model is applied to a plastic pipe industry. In addition, authors never consider more than 4 objectives. Besides, there is not a clear relation between the number of objectives and the solving method, which shows that apparently such number does not have any influence in the chosen method. Anyway, since 2018 there is a clear trend in employing either 565 the ε-constraint or the LP-metric method instead of weighted sums. Deeper research is necessary to establish pros and cons of using one or another method to solve multi-objective models oriented to design and assess resilient SCNs. SCND special cases are also studied. Ghomi-Avili et al. [50] Hamdan & Diabat [61] Jabbarzadeh et al. [76] Margolis et al. [102] Zahiri et al. [164] Zhao & You [168] Jabbarzadeh et al. [75] Mohammed et al. [110] Mohammed et al. [ Haeri et al. [60] these special cases with SCND resilience ( Table 7 ). Notice that sustainability 575 and green practices may impact negatively on the supply chain resilience [70] . For instance, a paper by Fahimnia & Jabbarzadeh [44] couples the sustainability concept with the resilience concept. This is achieved by designing a SCN through stochastic fuzzy goal programming. The sustainability approach implies that the model is multi-objective. These authors define SCND resilience 580 as "the capacity of a supply chain to absorb disturbances and retain its basic function and structure in the face of disruptions". This definition is similar to the robust supply chain definition given by Behzadi et al. [10] . Agricultural SCN is also a studied special case in our taxonomy. We classify a paper as agricultural if at least one modeled uncertain parameter is directly 585 related to agricultural aspects, e.g., harvest time and yield affected by diseases in crops [10] , or supply impacted by floods or droughts [101] . Other specific characteristics that increase vulnerability of agricultural SCNs are seasonality in supply and demand [155] , as well as perishability of products [150] . Competitive supply chain is another identified special case in SCND. In 590 this context, "competitive" means that competition among rivals is explicitly considered when designing a SCN. Rivals' competitive actions may lead to lose market-share because, for example, clients buy the product to other suppliers [48, 129] . Therefore, quantities supplied by rivals are variables in the proposed model. Finally, Ghomi-Avili et al. [50] do not only design a competitive supply 595 chain, but they take into account green and closed-loop characteristics in a bi-objective fuzzy model that considers both operational and disruption risks. give to the negative consequences of operational and disruption risks, such as loss of customers and money, or even loss of lives due to the occurrence of natural or human-induced disasters. Besides, most papers that address some supply chain special case apply the model in a real-world case, which shows that theoretical models usually address generic SCNs. Finally, notice that some 605 papers considering real-world cases from the agricultural sector are not classified as belonging to agricultural special case. This is because these papers do not address any uncertain parameter specific of an agricultural supply chain. Hybrid sim-opt methods refer to the interaction between optimization and 610 simulation "to find near-optimal solutions to complex or stochastic optimization problems" [81] . In particular, this section analyzes the use of such methods for designing and assessing supply chains. Table 8 shows both the optimization and the simulation approaches used by each analyzed paper. Regarding optimization, MILP is the most used approach (53%). Regarding simulation, DES is the 615 preferred one (59%). The mix between these particular approaches is the most used hybridization (31% of the papers). All papers combine two methods, with the exception of Costa et al. [29], who hybridize three methods. They propose a sim-opt model to design a Colombian supply chain for bio-diesel production and distribution from palm oil feed-stock. Firstly, a deterministic simulation model is used to design the production pro-cess and obtain the production cost. This feeds a MILP model for locating production plants. Finally, a goal programming model is proposed to perform a micro-location considering social aspects. Fuzzy simulation Ji et al. [79] Exact methods are used by 59% of papers (Table 9 ). The other half uses which shows open opportunities for other techniques such as iterated local search [34] , tabu search [28] , or any other metaheuristic [47] or simheuristic [81] . Sim-opt papers make little use of scenarios with probabilities, as well as fuzzy 645 numbers ( They also model costs uncertainty through fuzzy numbers. Other modeled uncertain parameters are shown in Table 11 , such as cost (28% of the papers) and supply (25% of the papers). Nevertheless, demand is also the most addressed Purchase price Zhang et al. [166] objective is important here due to the environmental impact that the designed cold supply chain can yield. Two cases that require warehousing and transportation with a controlled temperature are analyzed through a hybrid sim-opt approach. to test the proposed approach. The increasing concerns about operational and disruption risks in SCND lead to use methods that model them as accurately as possible. As discussed in previous sections, designing and assessing resilient SCNs requires to take 705 into account many factors altogether, from uncertainty in several parameters of the optimization model that need to be considered (e.g., random occurrence of disruptions, random customers' demands or links availability, etc.) to dynamic conditions that might affect daily operations (e.g., continuous changes in availability of raw materials, transport costs, or customers' demands, among others). Consequently, solving methods need to be able to address all these characteristics of real-world SCNs together with the fact that, in most cases, SCNs constitute large-scale complex systems [46] . Modeling these concerns means: Although traditional SCND has focused on long-term and costly to reverse strategic decisions, there might be situations in which the design, re-design and assessment of a SCN needs to be performed in a fast way and several times in a short time period, thus calling for 'agile' SCND. This is the case, for example, of 745 emergency situations associated with natural or human-caused disasters, where a set of mobile facilities (e.g., first-aid centers) need to be quickly deployed over a territory and then re-allocated as new events happen (e.g., some areas get stabilized while others require more attention and resources). In this cases, decisions on the SCND (e.g., facility location, customers' allocation, or transport 750 modes) need to be made almost in real time and are required to be re-evaluated every few days or even hours. In these cases, the use of biased-randomized algorithms [57] constitute an interesting tool yet to be fully explored, specially when they are combined with parallelization techniques in order to generate high-quality solutions in real-time. Finally, with the advent of the internet of 755 things (IoT), real-time information obtained of sensors distributed by all the elements of the supply chain can be used to design SCN. Hence, the design process would not only become more efficient, but also much more reliable, since the IoT provides to supply chain managers a coherent stream of real-time data, by which they can develop flexible contingency plans and strategies to respond 760 to disasters [73] . Furthermore, as IoT allows the continuous self-assessment of the supply chain, it is possible to predict eventualities during supply chain operation. Thus, decision makers can assess and re-design the SCN in function of the time period of the disaster, providing a flexible and fast recuperation of the system, designing intelligent supply chains networks [31] . 765 This paper has provided a literature review on recent works related to the design of resilient supply chain networks via simulation and optimization approaches. A systematic literature review approach was followed. Our review shows that this is an emerging topic, which has been gaining 'momentum' during the last years. Simulation-optimization methods are specially designed to deal with considerations regarding uncertainty, time-efficiency and optimization of suitable criteria. The simulation side provides efficient tools to address uncertainty. Discrete-event simulation stands out in analyzed papers, although Monte Carlo simulation is also a useful tool. Nevertheless, simulation does not allow 775 itself to obtain optimal or near-optimal solutions. Therefore, the optimization side is intended to achieve them. Here, we find that most used mathematical approaches are stochastic programming, mixed-integer linear programming, fuzzy programming, and robust optimization. Hence, a hybrid simulation-optimization approach is particularly useful due 780 to the following facts: (i) both operational and disruption risks are subject to uncertainty in real-world problems. The use of a deterministic approach would lead to a supply chain design whose resilience is not coherent with this reality, increasing potential high economic losses after a disruption event; and (ii) instead of just measuring resilience or costs, as pure simulation would do, re-785 silience should be optimized in order to face risks properly. However, extremely resilient supply chains design would be highly expensive. A trade-off between resilience and any economic performance indicator would be very useful. Solving methods considering multiple objectives become relevant in these cases. Analyzed papers show that costs minimization is still the widely preferred demand is the most addressed uncertain parameter, which shows its relevance when designing and assessing supply chain networks. However, other parameters subject to uncertainty, such as costs, capacity, supply, node disruption, link disruption, and many others should also be considered together with demand. Clearly, these combinations are also open research lines, especially when con-sidering both operational and disruption risks. The uncertainty regarding these parameters is mostly addressed through probability distributions and scenarios with assigned probabilities. Fuzzy numbers are used to a lesser extent. Six supply chain design special cases were identified: sustainable, closedloop, agricultural, green, competitive and reverse supply chain. About 37% 805 of the papers address at least one of these. In addition, real-world cases are considered by most papers. Economic sectors, such as the automotive industry, the pharmaceutical industry, or the clothing industry are tackled. Most real-life supply chain networks are large-sized and associated problems easily become NP-hard as realistic constraints are included into the mathemati- The impact of supply network characteristics on reliability Capacity planning and warehouse location in sup-845 ply chains with uncertain demands Toward life-cycle reliability-, risk-and resilience-based design and assessment of bridges and bridge networks under independent and interacting hazards: emphasis on 850 earthquake, tsunami and corrosion Analysing supply chain resilience: integrating the constructs in a concept mapping framework via a systematic literature review. Supply Chain Management: An Interna-855 tional Where is supply chain resilience research heading? a systematic and co-occurrence analysis A genetic algo-860 rithm approach for multi-objective optimization of supply chain networks Supply chain optimization under risk and uncertainty: A case study for high-end server manufacturing A benders decomposition method for designing reliable supply chain networks accounting for multimitigation strategies and demand losses Should all literature reviews be systematic? Evaluation & Research in Education Robust and resilient strategies for managing supply disruptions in an agribusiness supply chain Development of a new approach for deterministic supply chain network design Seismic resilience of con-880 crete structures under corrosion Network-based approach to modelling uncertainty in a supply chain Resilience and sustainability of civil infrastructure: Toward a unified approach An evolution of uncertainty assessment and quantification Resilient food supply chain design: Modelling framework and metaheuristic solution approach A framework to quantitatively assess and enhance the seismic resilience of communities. Earthquake spectra Learnheuristics: hybridizing metaheuristics with machine learning for optimization with dynamic inputs Combining statistical learning with metaheuristics for the multi-depot vehicle routing problem with market segmentation Network 905 design and planning of resilient supply chains Resilience assessment of supply chains under different types of disruption Resilience metrics in the assessment of complex supply-chains performance operating under demand uncertainty Supply chain redesign for resilience using simulation Designing principles to create resilient supply chains Institute of Industrial and Systems Engineers (IISE) A unified framework for evaluating 920 supply chain reliability and resilience Optimal design of supply chain network under uncertainty environment using hybrid analytical and simulation modeling approach Building the resilient supply chain Logistical supply chain design for bioeconomy applications A decisional simulationoptimization framework for sustainable facility location of a biodiesel plant in colombia Fleet size optimization in the discarded tire collection process Improving supply chain resilience with employment of iot Design of clothing supply chain network 940 based on stochastic simulation Supply network resilience: a systematic literature review and future research Solving the deterministic and stochastic uncapacitated facility location problem: from a heuristic to a simheuristic Resilient solar photovoltaic supply chain network design under business-as-usual and hazard uncertainties Producing a systematic review The Sage handbook of organizational research methods chapter 39 Stochastic multi-objective production-distribution network design using simulation-based optimiza-960 tion Performance measures based optimization of supply chain network resilience: A nsga-ii+ cokriging approach Ripple effect in the supply 965 chain: an analysis and recent literature Modeling and analysing storage systems in agricultural biomass supply chain for cellulosic ethanol production Supply chain designs and management for biocrude production via wastewater treatment Re-975 silience and vulnerability in supply chain: Literature review. IFAC-PapersOnLine Proactive and reactive models for disaster resilient supply chain Marrying supply chain sustainability and resilience: A match made in heaven Responsive and resilient supply chain network design under operational and disruption 985 risks with delivery lead-time sensitive customers Recent advances in hybrid priority-based genetic algorithms for logistics and scm network design Handbook of metaheuristics volume 2 Designing a resilient competitive supply chain network under disruption risks: A realworld application Introduction to logistics systems planning and control A fuzzy pricing model for a green competitive closed-1000 loop supply chain network design in the presence of disruptions A multi-objective model for the closed-loop supply chain network design with a price-dependent demand, shortage and disruption Distinguishing between the concepts of supply chain agility and resilience. The International Optimal supply 1010 chain resilience with consideration of failure propagation and repair logistics An interdependent layered network model for a resilient supply chain Hybrid model to design an agro-food distribution network considering food quality Supply chain network design under uncertainty: A comprehensive review and future research directions Biased randomization of heuristics using skewed probability distributions: a survey and some applications Simulation-optimization techniques for closed-loop supply chain design 1030 with multiple objectives Supply chain network design using an integrated neuro-fuzzy and milp approach: A comparative design study A mixed resilient-efficient approach toward blood supply chain network design Robust design of blood supply chains under risk of disruptions using lagrangian relaxation. Transportation Re-1040 search Part E: Logistics and Transportation Review A new combination of robustpossibilistic mathematical programming for resilient supply chain network under disruptions and uncertainty: A real supply chain A systems approach to natural disaster resilience. Simulation Modelling Practice and Theory Robust global supply chain network design under disruption and uncertainty considering resilience strategies: 1050 A parallel memetic algorithm for a real-life case study A critical review on supply chain risk definition, measure and modeling Re-1055 search on the phenomenon of supply chain resilience: a systematic review and paths for further investigation Review of quantitative methods for supply chain resilience analysis. Transportation Research Part 1060 E: Logistics and Transportation Review A novel hybrid approach for synchronized development of sustainability and resiliency in the wheat network Simulation-based ripple effect modelling in the supply chain Revealing interfaces of supply chain resilience and sustainability: a simulation study Predicting the impacts of epidemic outbreaks on global supply chains: A simulation-based analysis on the coronavirus outbreak (COVID-19/SARS-CoV-2) case Viability of intertwined supply networks: sition paper motivated by COVID-19 outbreak The impact of digital technology and industry 4.0 on the ripple effect and supply chain risk analytics Literature review on disruption recovery in the supply chain Green and resilient 1085 design of electricity supply chain networks: a multiobjective robust optimization approach Resilient and sustainable supply chain design: sustainability analysis under disruption 1090 risks Designing a supply chain resilient to major disruptions and supply/demand interruptions Closed-loop supply chain network design under disruption risks: A robust approach with real world application A fuzzy programming approach 1100 for supply chain network design A resilient strategy for meat-food supply chain network design A review of simheuristics: Extending metaheuristics to deal with stochastic combinatorial optimization problems Simheuristics applications: dealing with uncertainty in logistics, transportation, and other supply chain areas A review of the literature 1115 on the principles of enterprise and supply chain resilience: Major findings and directions for future research Supply-chain analysis at volkswagen of america. Interfaces Hybrid optimization and simulation to design a logistics network for distributing perishable products Cost and environmental trade-offs in supply chain network design and planning: the merit of a simulation-based approach Integrated productiondistribution planning in two-echelon systems: a resilience view. Interna-1130 tional Optimal design and global sensitivity analysis of biomass supply chain networks for biofuels under uncertainty Modeling approaches for the design of 1135 resilient supply networks under disruptions A hybrid optimization/simulation approach for a distribution network design of 3pls Supply chain resilience: a systematic literature review and typological framework Decision support for integrated refinery supply chains: Part 2. design and 1145 operation Techno-economic optimization of ethanol synthesis from rice-straw supply chains Identification of key supply chain elements from the supply 1150 chain resilience viewpoint using the computer simulation and design of experiments Sustainable utilization and storage of carbon dioxide: analysis and design of an innovative supply chain A new resilience measure for supply chain networks. Sustainability, 9 A sample average approximation approach for supply chain network design with facility disruptions Exploring supply chain network resilience in the presence of the ripple effect Complex resource supply 1165 chains display higher resilience to simulated climate shocks Supply chain risk and resilience: theory building through structured experiments and simulation Resiliency optimization of biomass to biofuel supply chain incorporating regional biomass preprocessing depots A multi-objective optimization model for designing resilient supply chain networks Sustainable and resilient supply chain network design under disruption risks Complex network theorybased approach for designing resilient supply chain networks Adaptivity of complex network topologies for designing resilient supply 1185 chain networks An optimization-simulation approach to the network redesign problem of pharmaceutical wholesalers Robust optimization of the insecticide-treated bed nets procurement and distribution planning under uncertainty for malaria prevention and control Development of a mathematical model for sustainable closed-loop supply chain with efficiency and resilience systematic framework Investigating resilient 1200 supply chain design determinants using monte carlo simulation Eco-gresilient: Coalescing ingredient of economic, green and resilience in supply chain network 1205 design A hybrid MCDM-fuzzy multi-objective programming approach for a G-resilient supply chain network design Robust design of a sustainable and resilient bioethanol supply chain under operational and disruption risks Supply chain 1215 resilience for single and multiple sourcing in the presence of disruption risks Modeling the impact of unmet demand in supply chain resiliency planning Mitigating supply chain disruptions through the assessment of trade-offs among risks, costs and investments in capabilities The role of simulation and optimization methods in supply 1225 chain risk management: Performance and review standpoints Designing e-commerce supply chains: a stochastic facility-location approach A multi-objective approach for supply chain design considering disruptions impacting supply availability and quality Hybrid fuzzyprobabilistic approach to supply chain resilience assessment Reconciling supply chain vulnerability, risk and supply chain management Decision support for integrated refinery supply chains: Part 1. dynamic simulation Understanding the concept of supply chain resilience. The international journal of logistics management The new generation of operations research methods in supply chain optimization: A review Multi-echelon closed-loop supply chain network design and configuration under supply risks and lo-1250 gistics risks A biased-randomized metaheuristic for the capacitated location routing problem Supply chain resiliency: a review Decision-making 1260 models for supply chain risk mitigation: A review Supply chain performance measurement and evaluation: A mixed sustainability and resilience approach Resilient supply chain network design under competition: a case study Correlation between strategic and operational risk mitigation strategies 1270 in supply networks Modelling and analysing supply chain resilience flow complexity Supply chain resilience: Definitions and quantitative modelling approaches-a literature review Resilient supply chain design under operational and disruption risks considering quantity 1280 discount: A case study of pharmaceutical supply chain Retail supply chain network design under operational and disruption risks Cold supply chain design with environmental considerations: A simulation-optimization approach Developing a robust stochastic model for designing a blood supply chain network in a 1290 crisis: A possible earthquake in tehran A simulation-optimisation approach for supply chain network design under supply and demand uncertainties A stochastic programming approach for supply chain network design under uncertainty Developing resilient supply chains: lessons from high-reliability organisations. Supply Chain Management: 1300 Managing supply chain resilience to pursue business and environmental strategies Designing a resilient supply chain network for perishable products with random disruptions Performance indicators for supply chain resilience: review and conceptual framework Measuring supply chain resilience using a deterministic modeling approach Resilient design of biomass to energy 1315 system considering uncertainty in biomass supply Supply network modelling and simulation methodology. Simulation Modelling Practice and Theory Resilience in agri-food supply chains: a critical analysis of the literature and synthesis of a novel framework Resilience metrics and measurement methods for transportation infrastructure: the state of the 1325 art. Sustainable and Resilient Infrastructure Resilient network design of two supply chains under price competition: game theoretic and decomposition algorithm approach Perspectives in supply chain risk management. International journal of production economics Vehicles allocation for fruit distribution considering CO 2 emissions and decisions on subcontracting Simulation optimization in manufacturing analysis: Simulation based optimization for supply chain configuration design Incentivizing resilient supply chain design to prevent drug shortages: policy analysis using two-and multi-stage stochastic programs Supply chain resilience: definition, review and theoretical foundations for further study Network design using mix integer programming and monte carlo simulation in an international supply chain 1350 network A framework for designing robust food supply chains Toward a resilient holistic supply chain network system: Concept, review and future direction An improved voronoi-diagram-based algorithm for continuous facility location problem under disruptions A simulation-based robust optimization model for supply chain network design A computer simulation-based analysis of supply chains resilience in industrial environment An integrated two-layer network model for designing a resilient green-closed loop supply chain of perishable products 1370 under disruption Designing a resilient-green closed loop supply chain network for perishable products by considering disruption in both supply chain and power networks A simulation-based optimization method for solving the integrated supply chain network design and inventory control problem under uncertainty Hybrid algorithm for discrete 1380 event simulation based supply chain optimization Resilient hazardousmaterials network design under uncertainty and perishability Toward an integrated sustainable-resilient supply chain: A pharmaceutical case study Optimal design and operation for supply chain system of multistate natural gas under uncertainties of demand and purchase price A literature review on risk sources and resilience factors in agri-food supply chains Resilient supply chain design and operations with decision-dependent uncertainty using a data-driven robust optimization approach This work has been partially supported by the IoF2020 and the Erasmus+ Programs (2018-1-ES01-KA103-049767). We also acknowledge the support of the doctoral programs at the Universitat Oberta de Catalunya and Universidad de La Sabana (grant INGPhD-12-2020).