Disruption tails and post-disruption instability mitigation in the supply chain

9th IFAC Conference on Manufacturing Modelling, Management and 9th IFAC Conference on Manufacturing Modelling, Management and Control 9th IFAC IFAC Conference Conference on on Manufacturing Manufacturing Modelling, Management and Available online at www.sciencedirect.com 9th Modelling, Control 9th IFAC Conference on Manufacturing Manufacturing Modelling, Management Management and and 9th IFAC Conference on Modelling, Management and Berlin, Germany, August 28-30, 2019 Control Control Berlin, Germany, August 28-30, 2019 Control Control Berlin, Germany, August 28-30, 2019 Berlin, Berlin, Germany, Germany, August August 28-30, 28-30, 2019 2019 Berlin, Germany, August 28-30, 2019

ScienceDirect

IFAC PapersOnLine 52-13 (2019) 343–348

Disruption tails and post-disruption instability mitigation in the supply chain Disruption tails and post-disruption instability mitigation in the supply chain Disruption tails and post-disruption instability mitigation in the supply chain Disruption tails and post-disruption instability mitigation in the supply chain Disruption tails and post-disruption instability mitigation in the supply chain 1 2

Dmitry Ivanov 1, Maxim Rozhkov2 Dmitry Ivanov Maxim Rozhkov Rozhkov2222 Dmitry Ivanov Ivanov1111,,, Maxim Dmitry  Maxim Rozhkov Rozhkov Dmitry Ivanov Ivanov ,, Maxim 1 Maxim Rozhkov Dmitry of Business Administration 1Berlin School of Economics and Law,  Department of Business Administration 1Berlin School of Economics and Law,  Department Berlin School of Economics and Law, Department of Business Administration 1 Supply Chain Management, 10825 Berlin, Germany 1 Berlin School of Economics and Law, Department of Business Administration 1Berlin School of Economics Economics and Law, Law, Department Department of Business Business Administration Chain Management, 10825 Berlin, Germany Berlin SchoolSupply of and of Administration Supply Chain Management, 10825 Berlin, Germany Phone: +49 3085789155; E-Mail: [email protected] Supply Chain Management, 10825 Berlin, Germany Supply Chain Management, 10825 Berlin, Germany Phone: +49 3085789155; E-Mail: [email protected] Supply Chain Management, 10825 Berlin, Germany Phone: +49223085789155; E-Mail: [email protected] X5 Retail Group, Moscow, Russia Phone: E-Mail: [email protected] Phone: +49 +4923085789155; 3085789155; E-Mail: [email protected] X5 Retail Group, Moscow, Russia Phone: +49 3085789155; E-Mail: [email protected] X5 Retail Group, Moscow, 2 2 X5 Retail Group, Moscow, Russia 2 X5 Retail Group, Moscow, Russia X5 Retail Group, Moscow, Russia Russia Abstract: We study the coordination effects of production and ordering Abstract: We We study study the the coordination coordination effects effects of of production production and and ordering ordering policies policies in in the the supply supply chain chain (SC) (SC) Abstract: policies in the supply chain that experiences capacity disruptions and recovery. Discrete-event simulation methodology is used(SC) for Abstract: We study the coordination effects of production and ordering policies in the supply chain (SC) Abstract: We study the coordination effects of production and ordering policies in the supply chain (SC) that experiences capacity disruptions and recovery. Discrete-event simulation methodology is used for Abstract: We study the coordination effects of production and ordering policies in the supply chain (SC) that experiences capacity disruptions and recovery. Discrete-event simulation methodology is used for analysis with real company data and real disruptions. It has been revealed that disruption-driven changes in that and recovery. simulation methodology used that experiences capacity disruptions and recovery. Discrete-event simulation methodology is used for analysis with real real capacity companydisruptions data and and real real disruptions. ItDiscrete-event has been been revealed revealed that disruption-driven disruption-driven changes in that experiences experiences capacity disruptions and recovery. It Discrete-event simulation methodology is ischanges used for for analysis with company data disruptions. has that in SC behavior may result in post-disruption instability caused by backlog and delayed orders the accumulaanalysis with real data has revealed that disruption-driven in analysis with real company data and real disruptions. It has been revealed that disruption-driven changes in SC behavior result in in post-disruption post-disruption instability It caused by backlog backlog and delayed orders the the changes accumulaanalysis with may real company company data and and real real disruptions. disruptions. It has been been revealedand thatdelayed disruption-driven changes in SC result by orders accumulationbehavior of which whichmay in the the post-disruption period instability we call call as as caused “disruption tails”. A A transition of thesethe residues into SC behavior may result in post-disruption instability caused by backlog and delayed orders accumulaSC behavior may result in post-disruption instability caused by backlog and delayed orders the accumulation of in post-disruption period we “disruption tails”. transition of these residues into SC behavior may result in post-disruption instability caused by backlog and delayed of orders the accumulation of which in the post-disruption period we call as “disruption tails”. A transition these residues into the post-disruption post-disruption period destabilizes the normal normal operation modetails”. resulting in further furtherof delivery delays into and tion of post-disruption period we as A these tion of which which in in the theperiod post-disruption period we call call operation as “disruption “disruption tails”. A transition transition ofdelivery these residues residues into the destabilizes the mode resulting in delays and tion of which in the post-disruption period we call as “disruption tails”. A transition of these residues into the post-disruption period destabilizes the normal operation mode resulting in further delivery delays and non-recovery of SC performance. The results provide evidence that coordinated contingency policies need the post-disruption period destabilizes the normal operation mode resulting in further delivery delays and the post-disruption post-disruption period destabilizes the normal normal operation mode resulting in further further delivery delaysneed and non-recovery of SC SCperiod performance. The results results provide evidence thatresulting coordinated contingency policies the destabilizes the operation mode in delivery delays and non-recovery of performance. provide evidence that coordinated policies need to be be applied applied during during the disruption disruptionThe period to avoid avoid the disruption disruption tails. Copyright contingency © 2019 2019 IFAC IFAC non-recovery of The results provide evidence coordinated contingency policies non-recovery of SC SC performance. performance. The results provide evidence that that coordinated contingency policies need need to the period to the tails. Copyright © non-recovery of SC performance. The results provide evidence that coordinated contingency policies need to be applied during the disruption period to avoid the disruption tails. Copyright © 2019 IFAC to applied during the period to the disruption tails. © IFAC to be applied during the disruption disruption period to avoid avoid the disruption tails. Copyright © 2019 2019 IFAC to be applied during the disruption period to avoid the disruption tails. Copyright © 2019 IFAC © be 2019, IFAC (International Federation of Automatic Control) Hosting byCopyright Elseviersimulation, Ltd. All rights reserved. Keywords: supply chain, resilience, disruption, contingency policy, inventory, recovery Keywords: supply chain, resilience, disruption, contingency policy, inventory, simulation, recovery Keywords: supply chain, resilience, disruption, contingency policy, inventory, simulation, recovery Keywords: supply chain, resilience, disruption, contingency policy, inventory, simulation, recovery Keywords: chain, policy, recovery Keywords:1.supply supply chain, resilience, resilience, disruption, disruption, contingency contingency policy, inventory, inventory, simulation, simulation,The recovery INTRODUCTION considered considered as as “disruption “disruption tails”. tails”. The influence influence of of these these tails tails 1. INTRODUCTION 1. INTRODUCTION considered as “disruption tails”. The influence of these tails at the post-disruption time in the course of transition into the 1. considered as “disruption influence of these 1. INTRODUCTION considered as “disruption tails”. The influence of these tails at the post-disruption time tails”. in the The course of transition intotails the 1. INTRODUCTION INTRODUCTION considered as “disruption tails”. The influence of these tails at the post-disruption time in the course of transition into the normal operation mode is investigated. A comparison of SC Disruptions in the supply chain (SC) capacities may happen at the post-disruption time in the course of transition into the at the post-disruption time in the course of transition into the normal operation mode is investigated. A comparison of SC at the post-disruption time in the course of transition into the Disruptions in the supply chain (SC) capacities may happen normal operation mode is investigated. Apolicy comparison of SC SC operation with without contingency will Disruptions in the supply chain (SC) capacities may happen at factories and at facilities as operation is comparison Disruptions in capacities may happen normal operation mode is investigated. A comparison ofperSC operation with and andmode without contingencyA policy will be beof perDisruptions in the the supply supply chain (SC) capacities may such happen normal operation is investigated. investigated. Apolicy comparison SC at both both production production factorieschain and (SC) at logistics logistics facilities such as normal Disruptions in the supply chain (SC) capacities may happen operation with andmode without contingency willSC beof performed to compare the impact of disruption tails on operat both production factories and at logistics facilities such as distribution centers (DC) (Simchi-Levi et al. 2015, Gunoperation with and without contingency policy will be perat both and as and without contingency policy will be performed to with compare the impact of disruption tails on SC operat both production factories and at logistics facilities such as operation and the without contingency policy willSC beoperperdistribution centersfactories (DC) (Simchi-Levi et facilities al. 2015,such Gunat both production production factories and at at logistics logistics facilities such as operation formed to with compare impact of disruption disruption tails on ational and financial performance. Second, aa comparison of distribution centers (DC) (Simchi-Levi et al. 2015, Gunasekaran et al. 2015, Ho et al. 2015, Sokolov et al. 2016, formed to compare the impact of tails on SC operdistribution centers (DC) (Simchi-Levi et al. 2015, Gunformed to compare the impact of disruption tails on SC operational and financial performance. Second, comparison of distribution centers (DC) (Simchi-Levi et al. 2015, Gunformed to compare the impact of disruption tails on SC operet al. 2016, asekaran et al. 2015,(DC) Ho et(Simchi-Levi al. 2015, Sokolov distribution centers et al. 2015, Gunational and financial performance. Second, a comparison of SC operational and financial performance between an immeasekaran et al. 2015, Ho et al. 2015, Sokolov et al. 2016, Kamalahmadi and Mellat-Parast 2016, Jain et al. 2017, Reational and financial performance. Second, a comparison of asekaran et al. 2015, Ho et al. 2015, Sokolov et al. 2016, ational and financial performance. Second, a comparison of SC operational and financial performance between an immeasekaran et al. 2015, Ho et al. 2015, Sokolov et al. 2016, ational and financial performance. Second, a comparison of Kamalahmadi and Mellat-Parast 2016, Jain et al. 2017, Reasekaran et al.and 2015, Ho et al. 2015, Sokolov et2017, al. 2016, SC operational and financial financial performance between an immediate deactivation the plans and of Kamalahmadi Mellat-Parast 2016, Jain et al. Rezapour Altay He et 2018, Ivanov SC and performance an Kamalahmadi and Mellat-Parast 2016, Jain et al. 2017, ReSC operational andof financial performance between an immeimmediateoperational deactivation offinancial the contingency contingency plansbetween and installation installation of Kamalahmadi and Mellat-Parast 2016, et al. 2017, ReSC operational and performance between an immeIvanov zapour et et al. al. 2017, 2017, Altay et et al. al. 2018, 2018, HeJain et al. al. 2018, Kamalahmadi and Mellat-Parast 2016, Jain et al. 2017, Rediate deactivation of the contingency plans and installation of the normal operation policies after the capacity recovery and zapour et al. 2017, Altay et al. 2018, He et al. 2018, Ivanov 2018, Dolgui et al. 2018, Ivanov et al. 2019a,b). Disruption diate deactivation of the contingency plans and installation of zapour et al. 2017, Altay et al. 2018, He et al. 2018, Ivanov diate deactivation of the contingency plans and installation of the normal operation policies after the capacity recovery and zapour et al. 2017, Altay et al. 2018, He et al. 2018, Ivanov diate deactivation of the contingency plans and installation of 2018, Dolgui et al. 2018, Ivanov et al. 2019a,b). Disruption zapourDolgui et al. 2017, etIvanov al. 2018, et al. 2018, Ivanov the normal operation policies afterbethe the capacity recovery and aa usage of revival policy performed. These 2018, et al. Altay 2018, et al. al.He2019a,b). Disruption risks be by catastrophes, normal operation after capacity and 2018, Dolgui et Ivanov Disruption the normal operation policies after capacity recovery and usage of the the revival policies policy will will bethe performed. These experexper2018, Dolgui et al. al. 2018, 2018, Ivanovor et man-made al. 2019a,b). 2019a,b). Disruption the the normal operation policies afterbe the capacity recovery recovery and risks can can be caused caused by natural natural oret man-made catastrophes, 2018, Dolgui et al. 2018, Ivanov et al. 2019a,b). Disruption a usage of the revival policy will performed. These experiments aim at providing managerial insights on application of risks can be caused by natural or man-made catastrophes, political or These are rarely usage of revival will performed. These risks can be by or usage of the revival policy will be performed. These experiments aim at providing managerial insights on application of risks cancrises be caused caused by natural natural or man-made man-made catastrophes, usageaim of the the revival policy policy will be beinsights performed. These experexperpolitical crises or strikes. strikes. These disruptions disruptions are catastrophes, rarely locallocal- aaaiments risks can be caused by natural or man-made catastrophes, at providing managerial on application of contingency production and inventory control policies during political crises or strikes. These disruptions are rarely localized at the disrupted node and frequently propagate in the SC iments aim at providing managerial insights on application of political crises or strikes. These disruptions are rarely localiments aim at providing managerial insights on application of contingency production and inventory control policies during political crises or strikes. strikes. These disruptions are rarely rarely localaim atproduction providing and managerial insights onpolicies application of propagate in the SC iments ized at the disrupted node and frequently political crises or These disruptions are localcontingency inventory control during the period revival policies at the ized at the the disrupted node and and frequently propagate in the the SC contingency causing the ripple (Ivanov et 2014, and production and inventory control during ized at node frequently in SC contingency production and inventory control policies during the disruption disruption period and and revival policies atpolicies the transition transition ized at the disrupted node and frequently propagate in the SC contingency production andrevival inventory controlat policies during causing thedisrupted ripple effect effect (Ivanov et al. al. propagate 2014, Scheibe Scheibe and ized at the disrupted node and frequently propagate in the SC the disruption period and policies the transition time to after capacity recovery. causing the 2018, rippleIvanov effect (Ivanov et al.2019, 2014,Namdar Scheibeetand and Blackhurst 2017, al. period revival policies at causing the ripple effect et 2014, Scheibe the disruption period and and revival policies at the the transition transition timedisruption to normal normal operation operation after capacity recovery. causing the ripple effect (Ivanov et al. 2014, Scheibe the period and revival policies at the transition Blackhurst 2018, Ivanov 2017, 2018, 2018, 2019, Namdar etand al. the causing the 2018, rippleIvanov effect (Ivanov (Ivanov et al. al.2019, 2014,Namdar Scheibeet and timedisruption to normal normal operation operation after capacity recovery. Blackhurst 2017, 2018, al. 2017, Dolgui et al. 2018, Dubey et al. 2019, Dolgui time to after capacity recovery. Blackhurst 2018, Ivanov 2017, 2018, 2019, Namdar et al. time to normal operation after capacity recovery. Blackhurst 2018, Ivanov 2017, 2018, 2019, Namdar et al. time to normal operation after capacity recovery. 2017, Dolgui et al. 2018, Dubey et al. 2019, Dolgui Blackhurst 2018, 2017, 2018, 2017, Dolgui et etal. al.Ivanov 2018, Dubey et al. al.2019, 2019,Namdar Dolgui et et al. al. This This study study utilizes utilizes the the simulation simulation methodology methodology and and extends extends 2019a,b, Pavlov al. 2018, 2019). 2017, Dolgui et 2019, Dolgui et al. 2017, Dolgui et 2018, Dubey et al. 2019, Dolgui et al. This study and utilizes the simulation simulation methodology and extends 2019a,b, Pavlov etal. al.2018, 2018,Dubey 2019). et the strong growing literature on simulation applications 2017, Dolgui et al. 2018, Dubey et al. 2019, Dolgui et al. This study utilizes the methodology and extends 2019a,b, Pavlov Pavlov et et al. al. 2018, 2018, 2019). 2019). This study and utilizes the simulation simulation methodology and extends the strong growing literature on simulation applications This study utilizes the methodology and extends 2019a,b, 2019a,b, Pavlov et the strong and growing literature on simulation applications to disruption management (Schmitt and 2012, 2019a,b, Pavlov et al. al. 2018, 2018, 2019). 2019). In strong and literature on the strong and growing growing literature on simulation simulation applications to SC SC disruption management (Schmitt and Singh Singhapplications 2012, LewLewthe strong and growing literature on simulation applications In regard regard to to the the disruption disruption risks, risks, resilient resilient production production back-up back-up the to SC disruption management (Schmitt and Singh 2012, Lewis 2013, Ivanov 2017, Schmitt et al. 2017, Ivanov and In regard to the disruption risks, resilient production back-up and contingency inventory control policies became visible SC disruption management and Singh 2012, LewIn to resilient production to SC disruption management (Schmitt and Singh 2012, Lewis 2013, Ivanov 2017, Schmitt(Schmitt et al. 2017, Ivanov and RozhRozhIn regard to the the disruption disruption risks, resilient production back-up to SC disruption management (Schmitt and Singh 2012, Lewandregard contingency inventory risks, control policies became aaback-up visible to In regard to the disruption risks, resilient production back-up is 2013, Ivanov 2017, Schmitt2018, et al. Cavalcantea 2017, Ivanovetand and Rozhkov 2017, Paul and al. 2019, and contingency inventory control policies became visible research avenue over decade (Ivanov et 2010, 2013, Ivanov Schmitt et 2017, and contingency control policies became aaaa visible is 2013, Ivanov 2017, Schmitt et al. 2017, Ivanov Rozhkov 2017, Paul 2017, and Rahman Rahman 2018, Cavalcantea etand al. Rozh2019, and contingency inventory control policies became visible is 2013, Ivanov 2017, Schmitt2018, et al. al. Cavalcantea 2017, Ivanov Ivanovet and Rozhresearch avenue inventory over the the last last decade (Ivanov et al. al. 2010, is and contingency inventory control policies became visible kov 2017, Paul and Rahman al. 2019, Hosseini et al. 2019a,b). research avenue over the last decade (Ivanov et al. 2010, Schmitt and Singh 2012, Ivanov and Sokolov 2013, Raj et al. kov 2017, Paul and Rahman 2018, Cavalcantea et al. 2019, research avenue over the last decade (Ivanov et al. 2010, kov 2017, Paul and Rahman 2018, Cavalcantea et al. 2019, Hosseini et al. 2019a,b). research avenue over theIvanov last decade decade (Ivanov et al. al. 2017,et Paul and Rahman 2018, Cavalcantea et al. 2019, Schmitt and Singhover 2012, and Sokolov 2013, Raj2010, et al. kov research avenue the last (Ivanov et Hosseini al. 2019a,b). 2019a,b). Schmitt and Singh 2012, Ivanov and et Sokolov 2013, Raj2010, et al. al. Hosseini 2015, et 2016, Govindan al. Sawik 2017, et al. Schmitt and Singh and Sokolov 2013, Raj et Hosseini et al. 2019a,b). Schmitt and Singh 2012, Ivanov and Sokolov 2013, Raj et al. Hosseini et al. 2019a,b). 2015, Ivanov Ivanov et al. al.2012, 2016,Ivanov Govindan et al. 2016, 2016, Sawik 2017, Schmitt and Singh 2012, Ivanov and Sokolov 2013, Raj et al. The rest of this 2015, Ivanov etLücker al. 2016, 2016, Govindan et al. 2016, 2016, Sawik 2017, 2017, The rest of this study study is is organized organized as as follows. follows. Section Section 22 is is Mizgier 2017, et al. Despite significant pro2015, Ivanov al. al. Sawik 2015, Ivanov et al. 2016, Govindan et al. 2016, Sawik 2017, The rest toof ofcase-study this study studypresentation is organized organizedand as research follows. methodology Section 22 is is Mizgier 2017,et Lücker et Govindan al. 2018). 2018). et Despite significant pro- The devoted 2015, Ivanov et al. 2016, Govindan et al. 2016, Sawik 2017, rest this is as follows. Section Mizgier 2017, Lücker et al. 2018). Despite significant proThe rest of this study is organized as follows. Section 2 is devoted to case-study presentation and research methodology gress in theoretical studies and empirical principles to manThe rest of this study is organized as follows. Section 2 is Mizgier 2017, et 2018). Despite significant proMizgier 2017, Lücker et al. 2018). Despite significant prodevoted to case-study presentation and research methodology gress in theoretical studies and empirical principles to mandescription. Experimental settings production and Mizgier 2017, Lücker Lücker et al. al.and 2018). Despite significant pro- devoted to presentation and methodology gress in theoretical theoretical studies empirical principles to mandevoted to case-study case-study presentation and research research methodology description. Experimental settings with with production and distridistriage severe disruptions in the SC at proactive and reactive devoted to case-study presentation and research methodology gress in studies and empirical principles to mangress in theoretical studies and empirical principles to mandescription. Experimental settings with production and distridistriage severe disruptions in the SC at proactive and reactive bution disruptions are in the focus of Sections 3 4, gress in theoretical studies and empirical principles to mandescription. Experimental settings with production and age severe disruptions in the the SC SC at proactive proactive and reactive reactive description. Experimental settings with production and distridistribution disruptions are in the focus of Sections 3 4, rerestages, recent literature mostly assumed an immediate transidescription. Experimental settings with production and age severe disruptions in at and age severe disruptions in at and bution disruptions are in the the focus focus of Sections Sections 3 and and 4, 4, the restages, recent literature mostly assumed an immediate transi- bution spectively. Section 5 concludes the paper by summarizing age severe disruptions in the the SC SC at proactive proactive and reactive reactive disruptions are in of 3 restages, recent literature mostly assumed an immediate transibution disruptions are in the the focus focus of Sections Sections 3 and and 4, 4, the respectively. Section 5 concludes the paper by summarizing tion to a normal operation mode at the time of capacity rebution disruptions are in of 3 restages, recent literature mostly assumed an immediate transistages, recent literature mostly assumed immediate transiSection 5 concludes the paper bystudy, summarizing the tion to recent a normal operation mode at the an time of capacity re- spectively. most insights, limitations of and stages, literature mostly assumed immediate transiSection 5 the by summarizing the tion to a normal normal operation mode at the an time of capacity re- spectively. spectively. Section 5 concludes the paper by summarizing the most important important insights, limitations of this this study, and future future covery. aa full the spectively. Section 5 concludes concludes the paper paper bystudy, summarizing the tion to mode at of retion to aaaMoreover, normal operation operation modestabilization at the the time timeafter of capacity capacity re- most most important insights, limitations of this and future covery. Moreover, full system system stabilization after the capacicapaciresearch avenue. tion to normal operation mode at the time of capacity reimportant covery. Moreover, full system stabilization stabilization after theany capacimost important insights, limitations limitations of of this this study, study, and and future future research avenue.insights, ty has assumed considering resiimportant covery. Moreover, aaaa full system after capacicovery. Moreover, full systemwithout stabilization after the theany capaciresearch avenue.insights, limitations of this study, and future ty recovery recovery has been been assumed without considering resi- most covery. Moreover, full system stabilization after the capaciresearch avenue. ty recovery has been assumed without considering any resiresearch avenue. dues such as delayed orders and backlogs accumulated over ty recovery been without considering any resity recovery has been assumed assumed without considering any over resi- research avenue. delayed orders and backlogs accumulated dues such ashas ty recovery been assumed without considering any residues such as ashas delayed orders and backlogs accumulated over 2. the destabilized system during the disruption time. dues such delayed orders and backlogs accumulated over 2. PROBLEM PROBLEM STATEMENT STATEMENT dues such as delayed orders and backlogs accumulated over the destabilized systemorders duringand the backlogs disruptionaccumulated time. dues such as delayed over Two case studies 2. PROBLEM STATEMENT the destabilized system during the disruption time. are considered in 2. PROBLEM STATEMENT the destabilized system during the disruption time. 2. PROBLEM STATEMENT Two case studies are considered in this this paper paper in in regard regard to to the destabilized system during the disruption time. 2. PROBLEM STATEMENT the destabilized systemresearch during the disruption time. with the Two This case studies are distribution considered incapacity this paper paper in regard regardthat to both production and disruptions case studies are considered in this in to This study study closes closes the the research gap gap described described above above with the Two Two case studies are considered in this paper in regard to both production and distribution capacity disruptions that Two case studies are considered in this paper in regard to This studytocloses closes the research gap described above with the both objective reveal the dependencies among disruptions, production and distribution distribution capacity disruptions that This study gap above with happened in for the observed real This study the research gap described above with the production and disruptions objective tocloses revealthe theresearch dependencies among SC SC disruptions, This studyto closes the gap described described above with the the both both production andand distribution capacity disruptions that happened in reality reality anddistribution for which which capacity the authors authors observed that real both production and capacity disruptions that objective reveal theresearch dependencies among SC disruptions, contingency policies, and transition to the post-disruption happened in reality and for which the authors observed real objective to reveal the dependencies among SC disruptions, company operational policies and performance. objective to reveal the dependencies among SC disruptions, happened in reality and for which the authors observed real contingency policies, and transition to the post-disruption objective to reveal the and dependencies among SC disruptions, happened happened in reality and for which the authors observed real company operational policies and performance. in reality and for which the authors observed real contingency policies, transition to the post-disruption operational First, the analysis conducted in company operational operational policies policies and and performance. contingency policies, the contingency policies, and transition to the post-disruption post-disruption operational mode. mode. First,and the transition analysis is is to conducted in regard regard to to company contingency policies, transition the post-disruption company operational policies and performance. performance. 2.1. Production capacity disruption operational policies and performance. operational mode. First,and the in analysis is to conducted in regard regard to company disruption-driven changes SC behavior resulting in de2.1. Production capacity disruption operational mode. First, the analysis is conducted in to operational mode. changes First, the the in analysis is conducted conducted in regard regard to 2.1. Production capacity disruption disruption-driven SC behavior resulting in deoperational mode. First, analysis is in to disruption-driven changes in SC behavior resulting in de2.1. Production capacity disruption layed accumulation which be 2.1. Production Production capacity capacity disruption disruption disruption-driven changes in behavior resulting in disruption-driven changes the in SC SC behavior of resulting in dedelayed orders orders and and backlogs backlogs the accumulation of which can can be 2.1. disruption-driven changes in SC behavior resulting in delayed orders orders and and backlogs backlogs the accumulation of which can can be layed the accumulation of which be layed orders and backlogs the accumulation of which can be layed orders and backlogs the accumulation of which can be

2405-8963 © 2019, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved. Copyright 2019 responsibility IFAC 348Control. Peer review© of International Federation of Automatic Copyright ©under 2019 IFAC 348 Copyright © 348 10.1016/j.ifacol.2019.11.140 Copyright © 2019 2019 IFAC IFAC 348 Copyright © 2019 IFAC 348 Copyright © 2019 IFAC 348

2019 IFAC MIM 344 Berlin, Germany, August 28-30, 2019

Dmitry Ivanov et al. / IFAC PapersOnLine 52-13 (2019) 343–348

The case-study is based on a FMCG company that produces juices/beverages at proprietary and contracted plants in Russia (Ivanov and Rozhkov 2017). As the company name cannot be revealed, we provide some background information regarding the data collection. The data was collected in 2016 for the coverage period of 2013-2015. Data from the company’s owned plants and the subcontractor plants has been used. The real and forecasted demand data was collected using internal company reports. The data on inventory, production and shipment control was observed in the management information system supported by a joint analysis with managers in the respective departments at the factories and DCs.

in the every-day practice. The demand model is relatively generic with two sources of randomness and a seasonal factor, which is common for food industry. Seasonal demand variations of 50% within the planning horizon are considered according to the analysis of company data. Also, long-term demand changes with a duration of four weeks are possible where demand may vary by 20%. Both demand variation parameters can be described by a uniform or triangular distribution. Uniform distribution was implemented in the presented experiments with aggregated historical demand data for 60 periods having been used. Production planning is based on a discrete event simulation approach. If inventory at the DC reaches the reorder point, a new production order is allocated, the size of which is a multiple of the minimum lot size. The allocated order cannot be cancelled. Production planning considers lead time from the factory to the DC. If the computed production period of a batch (for both types of products) is reached, the orders enter the system and are allocated in the queues. If an order is waiting in the queue longer than the planning horizon at the DC, this order exits the system without any lost order costs. If the constraint on the waiting time is met, the order is transferred into the production module. Processing start times are based upon the computed production week. Early production, i.e. schedule smoothing, is not allowed.

We consider a fragment of the FMCG SC, i.e., a two-stage SC with five DCs and one factory that delivers a product “Juice” to the DCs. In the product line “Juices” with average shelf life of 270-360 days, the issues of product perishability play an important role. The ordering at the DCs is based on demand forecasts and exhibits the (Q, s, r) inventory control policy. At each period r (r ∈[1;1200]), inventory level y is forecasted for n periods. In a case where yis less than re-order point s, new order of the size O_r subject to minimum reorder quantity Q with a planned delivery period r+n is placed. If production is only possible for each m periods, the planning is done for the n+m-1 periods. The shipments are made with FEFO (first expired-first out) policy for both groups of customers.

Finally, SC resilience analysis is based on random disruptions resulting in a 50% decrease in production capacity. Disruptions are modelled as random events. The intervals between disruptions and their duration are subject to normal distribution. The assumption on normal distribution of disruption occurrence and recovery period is hypothetical and is partly confirmed with actual data.

Constraints on product perishability typically result in reductions in safety stock and increases in transportation frequency. The risks of product write-offs are considered in the SC planning subject to minimum service level. The targeted service level is 98.5%. The remaining shelf life limitations are aggregated according to contract agreements with major food retail companies operating in the Russian market (ranging from 62% to 70%). Consideration of the production capacity disruption risks may lead to safety stock increases. Production capacity at the factory is subject to random disruptions. In 2015 and 2016, production capacity breakdowns have been frequently observed. This resulted in delivery interruptions from the plants to DCs for the periods between two days and three weeks.

2.2. Logistics capacity disruption The second case-study is based on another company that produces non-perishable products for four regional markets. Without loss of generality, a fragment of the SC considered comprises three production plants and four regional DCs. The DC in region 1 crashed due to construction quality problems. A huge amount of inventory has been destroyed. At the day of the DC disruption, the experts estimate that the reconstruction of the DC will take about four months.

Regarding production capacity disruption, we focus on studying the impacts of production capacity disruptions on SC performance with consideration of a two-component demand structure and limited expiration dates. The discrete event simulation model in AnyLogic has been developed to simulate the SC dynamics and to determine the re-order, production and shipment quantities and times from DCs to the factory. Shipments, disruptions and recoveries are modelled as events. State charts and messages are used for information exchange between DCs and the factory. Mathematical formulation of this simulation model is presented in (Ivanov and Rozhkov 2017).

The company data was observed for the period of three years. Moreover, we observed the actual sales and service level data following the disruption at the DC described. A drastic decrease was observed in both sales and service level dynamics. The following control algorithms are implemented in the simulation model. Demand in the markets is considered as aggregated, normally distributed demand of all customers in this region characterized by a seasonal component subject to four periods. Weekly order placements from the markets to the DCs are considered.

The planning algorithms consider that the stock is deteriorating and that the batch, which can be shipped now, may be wasted in a few weeks. Base time unit is a week. This assumed that planning is done every week, but some parameters are measured in days. With the help of the company’s SC analysts, the heuristics have been developed, tested and used

Continuous review system is applied at the DCs. Backordering is allowed so that no orders can be lost. For simplification, an average lead time from DC to the market is considered assuming that all customers within the region will be reached during this lead time. According to demand genera349

Dmitry Ivanov et al. / IFAC PapersOnLine 52-13 (2019) 343–348

tion algorithms, orders are placed at the DCs. Subject to inventory-on-hand, safety stock, lead times, re-order point and the target inventory, shipments to the markets and replenishment from the factories is controlled.

345

the normal operation policies after the capacity recovery and a usage of the revival policy will be performed. These experiments aim at providing managerial insights on application of contingency production and inventory control policies during the disruption period and revival policies at the transition time to normal operation after capacity recovery. The contingency plan includes the installation of additional links in the SC which lead from the factories directly to the market 1. Back-up contractors, capacity flexibility (capacities of own plant in region 1) and usage of capacity of own other plants in neighborhood countries can be activated during the disruption period as a support to the contingency plan actions. These actions are considered as extended coordinated contingency policy (cf. experiments in Sect. 5.2).

Since in reality the company can widely use logistics service provider capacities along with the own fleet, no transportation capacity limitations are included. Because of the same reason and model simplification, no further restrictions on transportation control policies such as minimum or maximum load or aggregation periods are considered. Transportation time is computed in software automatically subject to distance and average truck speed. Real routes from anyLogistix software are considered in the computations. A set of key performance indicators (KPI) has been established to analyse the simulation results. The ELT (expected lead time) service level is the ratio of orders delivered within the “Expected lead time” to total orders. The expected lead time is a parameter set for each market. It measures the time between the placement an order at a DC and receiving the goods from the DC. Current backlog orders depict the currently unprocessed number of orders, i.e., the orders, which were received but have not been shipped yet. It is updated on daily basis in the cases if incoming orders is received, new shipment is sent, incoming orders is lost, incoming shipment is processed (in accordance with processing time set for the facility). Delayed orders show statistics on the quantity of orders, which failed to arrive within the specified expected lead time. The data is updated each time an order is delayed.

4. EXPERIMENTAL RESULTS 4.1. Production capacity disruption analysis In Fig. 1, the dynamics of customer order fulfilment is presented. Capacity

Inventory with disruptions

Delayed production orders

Non-fulfilled (lost) client orders

disruption tails 100

50.000 40.000

80

30.000 60

20.000

40

10.000

20 0

99 101 103 105 107 109 111 113 115 117 119 121 123 125 127 129 131 133 135 137 139 141 143 145 147 149 151 153 155 157 159 161 163 165 167 169 171 173 175 177 179

capacity availability, %

120

quantity

2019 IFAC MIM Berlin, Germany, August 28-30, 2019

0

weeks

Figure 1. Dynamics of customer order fulfilment

For the problem considered above, discrete event simulation model and a network optimization model have been developed in anyLogistix. anyLogistix is a simulation and optimization tool developed by AnyLogic Company. Optimization functionality of anyLogistix is implemented on the CPLEX basis in Network Optimization module. Simulation functionality in anyLogistix is based on discrete-event simulation with the agents that can be used either in a standard setting or be customized in AnyLogic. anyLogistix allows a wide range of experiments in regard to facility location planning, multistage and multi-period SC design and planning, inventory control, transportation control, and sourcing analysis in both deterministic and stochastic settings. Variation and comparison experiments have been performed (Ivanov 2019).

The disruption lasts 26 weeks starting in period #110. The recovery period is 30 weeks. Due to the fact that delivery from the factory to the DC is just ahead of the capacity disruption period, inventory at the distribution centre is available until period #117. In periods #120 and #122, two small deliveries from the factory to the DC can be observed since 50% of capacity still operates. After the capacity recovery, a number of delayed production orders are shipped to the DC creating higher inventory costs. After that, order allocation intensity changes again. High inventory levels increase writeoff risks and the system tries to allocate fewer production orders. In the case of delivery delays, penalties may be incurred. For example, in period #165, inventory reaches zero which implies lost sales. Therefore, it can be observed that a production capacity disruption causes both product shortage and write-off risks.

3. EXPERIMENTAL SETTING The first stage of experiments contains production disruption analysis with the focus on contingency inventory control policy impacts. The second part of experiments addresses DC disruptions focusing on analysis of survival policy impacts.

After capacity recovery, a number of delayed production orders are shipped to the DC creating higher inventory costs. High inventory levels increase write-off risks and the system tries to allocate fewer production orders. The observed SC dynamics can be named “disruption tails” in regard to the impact of redundant production-ordering system behaviour during the disruption period on the production-ordering system behaviour in the after-disruption period (Ivanov and Rozhkov 2017). Examples of SC redundant behaviour during the disruption period can be redundant production or deliveries downstream from the disrupted part of the SC or redundant order allocations to disrupted facilities.

At both stages mentioned above, the analysis is conducted in regard to disruption-driven changes in SC behavior resulting in delayed orders and backlogs the accumulation of which can be considered as “disruption tails”. The influence of these tails at the post-disruption time in the course of transition into the normal operation mode is investigated. A comparison of SC operations with and with-out contingency policy will be done to compare the impact of disruption tails on SC performance. A comparison of SC performance between an immediate deactivation of the contingency plans and installation of 350

2019 IFAC MIM 346 Berlin, Germany, August 28-30, 2019

Dmitry Ivanov et al. / IFAC PapersOnLine 52-13 (2019) 343–348

Lost orders increase significantly in some periods following production capacity disruption and stabilize shortly after capacity recovery. Delayed orders increase almost immediately after production disruption. The stabilization period is longer for lost orders than for delayed orders. Its start is closely tied to the time when inventory reaches its maximum level.

Fig. 2 depicts a significant decrease in the service level (measured in orders; the diagram represents the average) following the disruption at the day 91 and lasting till the DC1 recovery at the day 213. The delayed orders can be observed even in the post-disruption period by the disruption tails.

During disruption, the average inventory in the SC does not reach zero level since the factory is still operating at 50% of production capacity. After capacity recovery, an inventory peak can be observed that is related to disruption tails. Therefore, we conclude that if delayed orders are increasing under conditions of stabilized service levels, this indicates a significant inventory increase in the near future in the SC.

4.2.2. Extended contingency recovery policy Simulation experiments of this part have been conducted with the contingency policy containing three emergency sources: • back-up contractors • capacity flexibility (capacities of own plant in region 1) • using capacity of own other plants in neighborhood countries In line with the study by Ivanov (2017) we assume a time lag in between the disruption, activating the contingency policy capacities, and the first effects of these contingency policy operation. The emergency sources operate according to the following logic. No initial inventory is available. Two days after the DC1 disruption, the emergency sources start producing for the market 1. First deliveries to the market 1 arrive in about 18-20 days after the disruption date. The simulation results are presented in Fig. 3.

4.2. Logistics capacity disruption analysis 4.2.1. Initial contingency recovery policy Simulation experiments of this part have been conducted with the contingency policy that implies an installation of additional links in the SC from the factories to the market 1. These links are activated immediately after the DC1 disruption and function till the DC1 recovery. The results are shown in Fig. 2.

Fig. 2. Performance analysis of disruption impact with contingency recovery policy

Fig. 3. Performance analysis of disruption impact with extended contingency policies 351

2019 IFAC MIM Berlin, Germany, August 28-30, 2019

Dmitry Ivanov et al. / IFAC PapersOnLine 52-13 (2019) 343–348

When comparing Figs 2 and 3, an increase in service level, a reduction in the number of delayed orders during the disruption period, and an elimination of delayed orders after the disruption recovery can be observed in Fig. 3. The extended recovery policy allows to stabilize the order fulfillment dynamics resulting on positive effect on service level performance. The delayed orders accumulated over the disruption period do not influence the SC operations and performance since new contracting plants compensates for this with the help of an additional production capacity. This allows the service level to be recovered faster as compared to the usage of recovery policy only (Figs 2 and 3). This provide evidence of the disruption tail mitigation with the help of an extended coordinated recovery policy based on a production capacity increase in the post-disruption period. This observation indicates the necessity to align the normal operation policy and deactivation of the contingency policies.

347

the capacity recovery in the presence of the delayed orders and backlogs also results into the destabilizing the inventory and production systems and in the new delayed orders and backlogs. The results obtained suggest two managerial insights. First, contingency production and inventory control policies need to be applied during the disruption period to avoid the disruption tails. Second, special actions need to be developed for the transition time from the contingency plan to the disruption-free operation mode. We provide examples of contingency policies for production factory and DC disruptions. Disruption tails can be reduced by applying a contingency policy during the disruption time when the production control system cancels excessive DC orders waiting in the production system queues because of disrupted manufacturing capacities. Alternatively, the DCs need to adjust their ordering policy subject to the reduced production capacity to avoid long waiting times and the resulting delayed orders, backlogs, service level reductions, and the production-inventory control system destabilization after the capacity recovery.

5. CONCLUSIONS We studied inventory, service level and order fulfillment dynamics in the SC that experiences capacity disruptions and recovery. Discrete-event simulation models in AnyLogic and anyLogistix have been used for analysis with real company data and for real disruptions in production and logistics capacities.

In addition, of the companies operated with forecasted recovery dates, the inertness of the decisions on activation and deactivation of contingency plans frequently leads to the disruption tails. The disruption tails represent the residues from the disruption period such as backlog and delayed orders which may influence the SC operations and performance in the post-disruption mode. The extended coordinated contingency policy intends to mitigate the negative impact of these disruption tails and stabilize the SC control policies and long term performance impact (Ivanov and Dolgui 2018).

Despite significant progress in theoretical studies and empirical principles to manage severe disruptions in the SC at proactive and reactive stages, recent literature mostly assumed an immediate transition to a normal operation mode at the time of capacity recovery. Moreover, a full system stabilization after the capacity recovery has been assumed without considering any residues such as delayed orders accumulated over the destabilized system states during the disruption time. The performed experiments allowed to reveal new insights on the existence of so called disruption tails which are the repercussions of the backlogs and delayed orders accumulated over the disruption period.

Concerning limitations of this study, it needs to be pointed out that the main events in the model such as disruption start, full recovery, high inventory increase, system stabilization, product write-off and the resulting problems with service level are significantly distributed in time. In the simulation model, the impacts of these events on SC efficiency and service level can be estimated according to the final experiment results. In real life, such a retrospective analysis can only be applied conditionally to performance impact analysis. More flexible planning horizons (e.g., freeze time in production plus lead time to DCs) can be included in said analysis. Other planning algorithms and contingency policies will also influence improvements in this research field. Another limitation of this study might be related to the case-specific analysis and insights. Further research can include analysis of other industries and datasets. Moreover, analytical studies are needed to provide more generalizable theoretical results and practical recommendations.

Disruption tails occur in two cases. The first reason for disruption tails are non-coordinated inventory and production control policies, i.e., if DCs and customers continue placing new orders even if the production capacity is disrupted. These order are accumulated in the waiting lines of production systems and increase the delays and backlogs. After the capacity recovery, the production system needs to produce both the backlog and new incoming orders. If the recovered production capacity is lower than the total of the backlog and new incoming orders, this results in new delays and backlogs. If the recovered production capacity is sufficient to cover both the total of the backlog and new incoming orders, this results into a disproportionally high delivery quantities at the DCs. In turn, the DCs stop ordering because of high inventory which results in some periods in new shortages. The second reason for disruption tails is the inertness of the decisions on activation and deactivation of contingency plans. An immediate deactivation of a contingency policy after the capacity recovery if the performance is not recovered fully by that moment. An immediate deactivation of the contingency plans and installation of the normal operation policies after

References Altay, N., Gunasekaran, A., Dubey, R., Childe, SJ (2018). Agility and Resilience as antecedents of Supply Chain Performance under moderating effects of Organizational Culture within Humanitarian Setting: A Dynamic Capability View. Production Planning and Control, 29(14), 1158-1174. Cavalcantea, I.M., Frazzon E.M., Forcellinia, F.A., Ivanov, D. (2019). A supervised machine learning approach to data-driven simulation of resilient supplier selection in digital manufacturing. 352

2019 IFAC MIM 348 Berlin, Germany, August 28-30, 2019

Dmitry Ivanov et al. / IFAC PapersOnLine 52-13 (2019) 343–348

International Journal of Information Management, https://doi.org/10.1016/j.ijinfomgt.2019.03.004 Dolgui, A., Ivanov, D., Sokolov, B. (2018) Ripple Effect in the Supply Chain: An Analysis and Re-cent Literature. International Journal of Production Research, 56(1-2), 414-430. Dolgui A., Ivanov D., Rozhkov M. (2019). Does the ripple effect influence the bullwhip effect? An integrated analysis of structural and operational dynamics in the supply chain. International Journal of Production Research, in press. Dubey R., Gunasekaran A., Childe, S. J. Wamba S.F., Roubaud D., Foropon C. (2019). Empirical Investigation of Data Analytics Capability and Organizational Flexibility as Complements to Supply Chain Resilience. International Journal of Production Research, DOI: 10.1080/00207543.2019.1582820 Govindan, G., A. Jafarian, M. E. Azbari, T.M. Choi. (2016). Optimal Bi-Objective Redundancy Al-location for Systems Reliability and Risk Management. IEEE Transactions on Cybernetics, 46, 1735-1748. Gunasekaran A., Subramanian H., Rahman S. (2015). Supply chain resilience: role of complexities and strategies. International Journal of Production Research, 53(22), 6809-6819. He J., Alavifard, F., Ivanov, D., Jahani H. (2018). A real-option approach to mitigate disruption risk in the supply chain. Omega, https://doi.org/10.1016/j.omega.2018.08.008 Ho, W., Zheng T., Yildiz H. & Talluri S. (2015). Supply chain risk management: a literature review. International Journal of Production Research, 53(16), 5031-5069. Hosseini S., Ivanov D., Dolgui A. (2019a). Review of quantitative methods for supply chain resilience analysis. Transportation Research: Part E, 125, 285-307. Hosseini, S., Morshedlou, N., Ivanov D., Sarder, MD., Barker, K., Al Khaled, A. (2019b). Resilient supplier selection and optimal order allocation under disruption risks. International Journal of Production Economics, DOI: 10.1016/j.ijpe.2019.03.018. Ivanov D. (2017). Simulation-based ripple effect modelling in the supply chain. International Journal of Production Research, 55(7), 2083-2101. Ivanov D. (2018). Revealing interfaces of supply chain resilience and sustainability: a simulation study. International Journal of Production Research, 56(10), 3507-3523. Ivanov D. (2019) Disruption tails and revival policies: A simulation analysis of supply chain de-sign and production-ordering systems in the recovery and post-disruption periods. Computers and Industrial Engineering, 127, 558-570 Ivanov D., and Sokolov B. (2013). Control and system-theoretic identification of the supply chain dynamics domain for planning, analysis, and adaptation of performance under uncertainty. European Journal of Operational Research, 224(2), 313–323. Ivanov D., Rozhkov M. (2017). Coordination of production and ordering policies under capacity disruption and product write-off risk: An analytical study with real-data based simulations of a fast moving consumer goods company. Annals of Operations Research, https://doi.org/10.1007/s10479-017-2643-8 Ivanov D., Sokolov B., Pavlov A., Dolgui A., and Pavlov D. (2016). Disruption-driven supply chain (re)-planning and performance impact assessment with consideration of pro-active and recovery policies. Transportation Research: Part E, 90, 7‒24. Ivanov D., Sokolov B., Pavlov, A. (2013). Dual problem formulation and its application to optimal re-design of an integrated production-distribution network with structure dynamics and ripple effect considerations, International Journal of Production Research, 51(18), 5386-5403. Ivanov D., Dolgui, A. (2018) Low-Certainty-Need (LCN) Supply Chains: A new perspective in managing disruption risks and resilience. International Journal of Production Research, https://doi.org/10.1080/00207543.2018.1521025.

Ivanov D., Tsipoulanidis A., Schönberger J. (2019a). Global Supply Chain and Operations Management, Springer, 2nd Ed. Ivanov, D., Dolgui A., Sokolov B. (Eds) (2019b). Handbook of Ripple Effects in the Supply Chain. Springer, New York: Ivanov, D. (2018). Structural Dynamics and Resilience in Supply Chain Risk Management. Springer, New York. Ivanov, D., Dolgui A., Sokolov B., Ivanova M. (2017). Literature review on disruption recovery in the supply chain. International Journal of Production Research, 55(20), 6158-6174. Ivanov, D., Sokolov B., Kaeschel J. (2010) A multi-structural framework for adaptive supply chain planning and operations control with structure dynamics considerations, European Journal of Operational Research, 200(2), 2010, pp. 409-420. Ivanov, D., Sokolov, B., Dolgui, A. (2014) The Ripple effect in supply chains: trade-off ‘efficiency-flexibility-resilience’ in disruption management, International Journal of Production Research, 52:7, 2154-2172. Jain, V., S. Kumar, U. Soni & C. Chandra (2017). Supply chain resilience: model development and empirical analysis. International Journal of Production Research, 55(22), 6779-6800. Kamalahmadi M. & Mellat-Parast M. (2016) Developing a resilient supply chain through supplier flexibility and reliability assessment. International Journal of Production Research, 54:1, 302-321. Lücker, F., Seifert R.W. & Biçer I. (2019) Roles of inventory and reserve capacity in mitigating supply chain disruption risk. International Journal of Production Research, 57(4), 1238-1249. Mizgier K. J. (2017) Global sensitivity analysis and aggregation of risk in multi-product supply chain networks, International Journal of Production Research, 55(1), 130-144. Namdar, J., Li, X., Sawhney R. & N. Pradhan (2018). Supply chain resilience for single and multiple sourcing in the presence of disruption risks. International Journal of Production Research, pub-lished online. Paul S., Rahman S. (2018). A quantitative and simulation model for managing sudden supply delay with fuzzy demand and safety stock. International Journal of Production Research, DOI10.1080/00207543.2017.1412528 Pavlov A., Ivanov D., Dolgui A., Sokolov B. (2018) Hybrid fuzzyprobabilistic approach to supply chain resilience assessment. IEEE Transactions on Engineering Management, 65(2), 303-315. Pavlov A., Ivanov D., Pavlov D., Slinko A. (2019). Optimization of network redundancy and contingency planning in sustainable and resilient supply chain resource management under conditions of structural dynamics, Annals of Operations Research, DOI: 10.1007/s10479-019-03182-6 Raj R., J. W. Wang; A. Nayak; M. K. Tiwari; B. Han; C. L. Liu; W. J. Zhang (2015). Measuring the Resilience of Supply Chain Systems Using a Survival Model. IEEE Systems Journal, 9(2), 377 – 381. Rezapour, S., Farahani, R., and Pourakbar, M. (2017). Resilient supply chain network design under competition: a case study. European Journal of Operational Research, 259(3), 1017-1035. Sawik T. (2017). A portfolio approach to supply chain disruption management. International Journal of Production Research, 55(7), 1970-1991. Scheibe K.P., & Blackhurst, J. (2018). Supply chain disruption propagation: a systemic risk and normal accident theory perspective. International Journal of Production Research, 56(1-2), 43-59. Schmitt A.J., & Singh M. (2012). A quantitative analysis of disruption risk in a multi-echelon sup-ply chain. International Journal of Production Economics, 139(1), 23–32. Schmitt T.G., Kumar S., Stecke K.E., Glover F.W., Ehlen M.A. (2017). Mitigating disruptions in a multi-echelon supply chain using adaptive ordering. Omega, 68, 185-198. Sokolov B., Ivanov D., Dolgui A., Pavlov A. (2016). Structural quantification of the ripple effect in the supply chain. International Journal of Production Research, 54(1), 152-169.

353