Towards supply chain risk analytics

Towards Supply Chain Risk Analytics Fundamentals, Simulation, Optimization... Unexpected deviations and disruptions, those are subsumed underthe notion of supply chain risk, increasingly

Towards Supply

Chain Risk Analytics

Fundamentals, Simulation, Optimization

ISBN 978-3-658-14869-0 ISBN 978-3-658-14870-6 (eBook)

DOI 10.1007/978-3-658-14870-6

Library of Congress Control Number: 2016945784

Springer Gabler

© Springer Fachmedien Wiesbaden 2016

This work is subject to copyright All rights are reserved by the Publisher, whether the whole or part

of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission

or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed.

The use of general descriptive names, registered names, trademarks, service marks, etc in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use.

The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made.

Printed on acid-free paper

This Springer Gabler imprint is published by Springer Nature

The registered company is Springer Fachmedien Wiesbaden GmbH

Von der Fakultät für Wirtschaftswissenschaften des Karlsruher Instituts für Technologie (KIT) genehmigte Dissertation

Tag der mündlichen Prüfung: 26.10.2015

Referent: Prof Dr Stefan Nickel

Korreferent: Prof Dr Francisco Saldanha-da-Gama

Prüfer: Prof Dr Wolf Fichtner

Vorsitzender: Prof Dr Rudi Studer

The pursuit of well-sophisticated solutions, derived from theory andmade applicable in practice, led me to FZI Research Center for In-formation Technology At FZI I was given the opportunity to work

on improvements for logistics systems, to learn technical content, toget to know real-world problems and to experience a fruitful workingenvironment

During that time I worked on the topic of Supply Chain Risk whichresulted in the underlying thesis However, this work would havenot been possible without the support and guidance of others, who

I want to thank here

I am indebted to Prof Stefan Nickel for being the supervisor ofthis thesis and a mentor to my research work – especially for hisfarsighted guidance, his productive ideas, and his sincere interest indiscussing (rebutting or exploring)

My thanks also go to Prof Francisco Saldanha-da-Gama for beingthe co-supervisor of this work – for his calm, his profound thoughtsand his ability to explain

Thanks are due to Prof Wolf Fichtner for being part of the nation committee and to Prof Rudi Studer for his sincere conduct

exami-as the chairman of the examination committee

Additionally, I thank my colleagues at the Logistics and SupplyChain Optimization Group at FZI, the Institute of Operations Re-search (IOR) as well as the Institute for Material Handling and Logis-

tics (IFL) of the KIT – Karlsruhe Institute of Technology for sharinggood and bad research times.

Priceless support and patient encouragement came from my familyand friends They all have never become tired to encourage me.Special thanks go to my parents, who have continuously offered me

a quite and safe place for retreat

Unexpected deviations and disruptions, those are subsumed underthe notion of supply chain risk, increasingly aggravate the planningand optimization of supply chains Over the last decade there hasbeen a growing interest in including risk aspects for supply chainoptimization models This development has led to the adoption ofrisk concepts, terminologies and methods defined and applied in abroad variety of related research fields and methodologies However,for the purpose of supply chain risk management the suitability ofrisk, as it is coined in these domains, is up for discussion.

The major contribution of this thesis is given by the development

of a profound conceptual basis of supply chain risk analytics andthe transfer of newly defined concepts for the modeling and opera-tionalization of supply chain risk within simulation and optimizationapproaches

The first part gives an extensive analysis of fundamental conceptsand approaches that surround research in the field of supply chainrisk management This includes a review of available concepts of risk

in general and supply chain risk in particular As supply chain risk

is either ambiguously or incompletely defined the literature reviewdoes not only critically revise existing approaches, but also identifiesessential aspects that drive the extent of supply chain risk Part Iprovides adjustments of commonly used concepts and offers a newdefinition of supply chain risk It is emphasized that it is the re-sponsibility of supply chain risk analysis to evaluate the interactions

of risk defining elements Having set the foundation for future proaches the new concept of supply chain risk analytics is coined

ap-Using mathematically based approaches, supply chain risk analytics

is tailored for the management of supply chain risk and associatedsub themes A discussion of the value of mathematically approaches

in the light of risk-aware solutions and a review of existing literaturewithin the field of operations research complete Part I

Consistently following the discussions and conclusions provided inPart I, Part II introduces a new simulation-based procedure for iden-tifying and assessing supply chain risks for a given supply chain, de-

noted by SimSCRF The approach builds on existing proprietary

sup-ply chain planning engines and applies methods from design of iments to determine weaknesses of the underlying supply chain Toalign the data model for supply chain planning with the simulation-based representation of a given supply chain, an object-oriented in-formation framework is presented The introduction of an additionallayer between planning engine and analysis algorithms offers the pos-sibility to easily switch between different planning engines and as aresult conduct risk analysis for different supply chain planning prob-lems An exemplary supply chain risk analysis is conducted on a realcase supply chain originating from the chemical industry The eval-uation provides insights on the existence of supply chain risk and itsextent as well as on potential conclusions for mitigation options.While the solution approach of Part II is characterized by the in-terplay of technical entities within a consistent process flow, PartIII focuses on the development of a risk-aware optimization modelfor supply chain network design problems Based on contemporaryresearch gaps identified for optimization approaches in Part I, PartIII deduces a mixed-integer two-stage stochastic programming modelthat extends the capacitated plant location problem and additionallyoffers the possibility to formalize and operationalize supply chain risk.With the goal to evaluate risk-aware solutions the concept of value

exper-of risk consideration is defined The evaluation exper-of the developed timization model discloses its usefulness in terms of providing risk-aware solutions and of approaching risk by stochastic programmingprinciples

op-Abstract VII

1.1 Motivation for Beginners 1

1.2 Advanced Risk 5

1.3 The Real Introduction: The Name of the Game 8

1.4 Outline and Course of Discussion 12

I Supply Chain Risk Concepts – Fundamentals 17 2 The Genesis of Supply Chain Risk 19 2.1 Logistics Innovations – A Blessing and a Curse 20

2.2 Supply Chain Disruptions 23

2.2.1 Environmental Disruptions 29

2.2.2 Economic Disruptions 30

2.2.3 Socio-Geopolitical Disruptions 33

2.2.4 Technological Disruptions 34

2.3 Coping with Risk 35

2.3.1 Enterprise Risk 35

2.3.2 Following the footsteps of Management 37

2.3.3 Identification needs Quantification –

Quantifi-cation needs Definition 40

3 A New Definition of Supply Chain Risk 43 3.1 The Evolution of Risk 45

3.2 Requirements for a Definition of Supply Chain Risk 47 3.3 Existing Approaches of Supply Chain Risk Definitions 48 3.4 Core Characteristics of Supply Chain Risk 50

3.4.1 Risk Objective 52

3.4.2 Risk Exposition 56

3.4.3 Risk Attitude 72

3.5 Re-defining Supply Chain Risk 74

4 Supply Chain Risk Analysis 77 4.1 The Risk of Supply Chain Risk Analysis 79

4.1.1 Biases of Risk Identification 83

4.1.2 Biases of Risk Countermeasures 88

4.1.3 Breaking of Biases 95

4.2 Main Elements of Supply Chain Risk Analysis 96

4.2.1 Analysis of Potential Triggers 96

4.2.2 Analysis of Performance Measurement 107

4.2.3 Analysis of Supply Chain Constitution 117

4.3 Tasks of Supply Chain Risk Analysis 127

5 Supply Chain Risk Analytics 131 5.1 Supply Chain Risk Analytics – Concept Definition 132 5.2 The Value of Supply Chain Risk Analytics 135

5.2.1 Risk Acceptance 137

5.2.2 Risk Reduction Measures 140

5.2.3 Risk Spreading Measures 149

5.3 Quantification Measures for Supply Chain Risk 152

5.3.1 Deviation Measures 153

5.3.2 Downside Risk 153

5.3.3 Expected Values 154

5.3.4 Probability and other measures 155

5.4 Risk-aware Supply Chain Optimization 157

5.4.1 Modeling Approaches 158

5.4.2 Solution Techniques 161

5.5 Research Gaps 161

II Supply Chain Risk Identification and Assess-ment – Simulation-based Framework 163 6 Simulation for Supply Chain Analysis 165 6.1 Simulation at a Glance 167

6.1.1 Basics 167

6.1.2 Technical entities of Simulation Tools 171

6.1.3 Simulation Paradigms 172

6.2 Simulation of Supply Chain Problems 175

6.3 Simulation and Optimization 176

7 Design, Metamodeling, and Analysis of Simulation Experiments 181 7.1 Meta-Models 184

7.2 Designs 187

7.2.1 Purpose of Design 188

7.2.2 Classic Factorial Designs 193

7.2.3 Design construction 196

7.3 Analysis 198

7.3.1 Regression Analysis 199

7.3.2 Method of Least Squares 200

7.3.3 Analysis of Variance (ANOVA) 202

7.3.4 Measures of Factor Effects 202

7.4 Illustrative Examples 206

7.4.1 A 24 Full Factorial Experiment for the Analy-sis of Production Characteristics 206

7.4.2 A 24−1 IV Fractional Factorial Experiment for the Analysis of Production Characteristics 211

7.5 Cautions with the Design of Experiments 213

8 A Simulation-based Approach for Supply Chain Risk

8.1 Requirements 218

8.2 A New Approach for Supply Chain Risk Analysis – Basic Models 220

8.2.1 Scenario-based Procedure 220

8.2.2 Screening Procedure 227

8.2.3 Procedure for Risk Quantification 229

8.3 Summarized Main Features of the Approach 233

9 Representative Master Planning Module for Supply Chains 235 9.1 Planning Tasks of Supply Chains 236

9.2 Mathematical formulation of a Master Planning Prob-lem 238

9.2.1 Determinants 241

9.2.2 Objective Function 247

9.2.3 Restrictions 248

10 A Conceptual Information Meta-Model for Supply Chains 261 10.1 Requirements for a Supply Chain Information Meta-Model 262

10.2 Related Work 264

10.3 Modeling Supply Chain Information 266

10.3.1 Concepts 266

10.3.2 Properties 268

10.3.3 Relations 268

10.3.4 Constraints 269

10.4 A Supply Chain Model 271

11 A Real Case Evaluation of the SimSCRF Approach 277 11.1 The Case 279

11.2 Contemporary Risk Quantification 281

11.3 A new View on Supply Chain Risk Analysis 286

11.4 Supply Chain Risk Analysis 288

11.4.1 Work Flow 288

11.4.2 Effect Analysis 300

11.4.3 Risk Line Identification 303

11.5 Conclusions and Outlook 308

III Strategic Supply Chain Risk Mitigation – Op-timization Approaches 311 12 Embedding Comprehensive Risk 313 12.1 Mathematical Model Formulations 316

12.1.1 Notations 316

12.1.2 The Risk-aware Capacitated Plant Location Problem (CPLP-Risk) 319

12.2 Illustrative Example 327

12.2.1 Data Input 328

12.2.2 Solution Plausibility 333

12.2.3 The Value of Risk Consideration 339

12.2.4 Quantification of Supply Chain Risk 343

12.3 Preliminary Computational Results 345

12.4 Model Extensions 354

12.4.1 Model Extensions for the Affected Supply Chain 354

12.4.2 Model Extensions for the Risk Objective 357

12.4.3 Model Extensions for the Risk Attitude 358

12.5 Conclusions and Outlook 360

13 Conclusions and Outlook 363 13.1 Conlusions 364

13.2 Outlook 369

1.1 A supply chain example 91.2 Outline 142.1 Global risks landscape – movements from 2012 to

2013 272.2 Economic damage in US$ billion caused by reportednatural disaster (1975-2014) 312.3 Risk perception and complexity of different logisticssystems 372.4 Process cycle of supply chain risk management steps 393.1 Analysis of supply chain risk definitions 513.2 Core characteristics of supply chain risk 523.3 Supply chain risk objectives 553.4 Literature analysis of supply chain risk categories 664.1 Formation of supply chain risk analysis 814.2 Relationship between events and their related conse-quences 834.3 Graphical visualization of functional relation betweenthe set of identified supply chain risks and potentialmitigation options 894.4 Perspectives of supply chain risk management 924.5 Traditional vs risk-aware planning process 944.6 Relationship between events and their related conse-quences 974.7 Elements of uncertainty profiles 101

4.8 Exemplary uncertainty profiles of supply chain tor values 1034.9 Uncertainty profiles of real-world supply chain factors 1044.10 Exemplary uncertainty profiles of supply chain fac-tor values influenced by different potential triggers 1064.11 Functionality and profitability of supply chain pro-cesses affect efficiency and effectiveness 1084.12 Exemplary uncertainty profiles of supply chain fac-tor values 1124.13 Acceptable performance deterioration levels of morerisk-averse and more risk-seeking decision makers 1144.14 Risk profiles 1164.15 Levels of structural metrics indicating supply chainresilience 1194.16 Exemplary extracts of a supply chain and a bill ofmaterial 1215.1 Example of a supply chain 1365.2 Example of a disruption propagating through theentire supply chain 1375.3 Impact of a disruption without supply chain riskmanagement 1395.4 Development of performance and extra expenses inthe presence of a disruption when supply chain riskmanagement actions are not installed 1405.5 Impact of a disruption in the presence of a disruptionwhen reactive countermeasures are installed 1435.6 Development of performance and extra expenses inthe presence of a disruption when reactive counter-measures are installed 1445.7 Impact of a disruption in the presence of a disruptionwhen flexible countermeasures are installed 1455.8 Development of performance and extra expenses inthe presence of a disruption when flexible counter-measures are installed 146

fac-5.9 Impact of a disruption in the presence of a disruptionwhen robust countermeasures are installed 1475.10 Measures applied for quantifying supply chain risk 1566.1 Process steps of simulation 1706.2 Technical components of a simulation 1726.3 Causal and temporal classification of simulation andoptimization integrating approaches 1787.1 Main elements of an experimental setup 1827.2 Relationship between input factors and output vari-able of the system and its simulated meta-model 1857.3 A response surface and a corresponding contour plot 1927.4 Interaction plots 2107.5 Residuals and normal probability plot of 24 factorialexperiment 2158.1 Relationship between supply chain risk defining en-tities 2228.2 Conceptual and simulation approach of a new supplychain risk analysis approach 2238.3 Basic process flow of a new supply chain risk analysismodel 2258.4 Process flow of the screening procedure within thesupply chain risk analysis model 2288.5 Process flow of the risk quantification procedure 2308.6 Risk line 2329.1 Supply chain planning matrix 2379.2 Software modules covering the SCP-matrix 2399.3 Set of nodes 2409.4 Visualization of the production and resource plan-ning Variables 2499.5 Visualization of modeling logic for substitute prod-ucts 253

10.1 Concepts of the deduced information meta-model forsupply chains 26710.2 Link-node constraint 26910.3 Capability-activity-product constraint 27010.4 Conceptual modeling of the supply chain task trans-port 27410.5 Conceptual modeling of the supply chain task pro-duction 27510.6 Conceptual modeling of the supply chain task storing 27611.1 Supply chain under investigation: physical view 28011.2 Supply chain under investigation: product view 28211.3 Relationship between frozen horizon, the emergence

of disruption information, and the start of tions of supply chain factors 29011.4 Planning reports for the reference supply chain and

modifica-a single-scenmodifica-ario supply chmodifica-ain: network view 29111.5 Planning reports for the reference supply chain and

a single-scenario supply chain: KPI view 29211.7 User input aligned to the work flow of the SimSCRF

model 29311.6 Conceptual modeling of the supply chain task pro-duction (replicate) 29411.8 Factor template for supply chain factor maximumproduction capacity 29611.9 Extract of the definition script for base data activityamount 29811.10 Extract of the definition script for base data activitycosts 29911.11 Definition script for the performance indicator activ-ity expenses 30011.12 Visualization of risk line quantification (A) 30411.13 Visualization of risk line quantification (B) 30511.14 Risk line for factor A and D 307

12.1 The core characteristics of supply chain risk SCR) hierarchy 31412.2 Development of stochastic parameters 33112.3 Development of overall capacity (gray) and demand(black) for the illustrative example 33212.4 Solution comparison of deterministic CPLP-T’ andstochastic CPLP-RISK model 34212.5 VSS and EVPI values for illustrative example withdifferent facility costs 34912.6 Development of VSS and EVPI values for illustrativeexample with different facility costs 35012.7 Development of computation time and related VSS

(CC-of the illustrative example with different facility costs 35212.8 Development of computation time of the stochasticprogram and its deterministic counterpart of the il-lustrative example with different facility costs 353

2.1 Opportunities and risks of prevailing logistics best

practices 22

2.2 Major disruptions from 1997 – 2011 29

3.1 Uncertainty model of uncertain parameter develop-ment 58

3.2 Supply chain risk categories 63

3.3 Supply chain vulnerability definitions 68

3.4 Risk attitudes considering approaches 73

4.1 Major disruptions and related countermeasures 91

5.1 Descriptive, predictive, and prescriptive analytics 133 5.2 Categorization of modeling approaches 159

5.3 Modeling approaches of the reviewed papers 160

5.4 Solution techniques of the reviewed papers 161

6.1 Simulation: terms and definitions 168

6.2 Summary of major characteristics of three main sim-ulation paradigms 174

7.1 Design matrix of a 24 full factorial design 207

7.2 Contrast constants for a 24 full factorial design 208

7.3 Effect estimates and sums of squares for the 24 fac-torial experiment 209

7.4 Analysis of variance table 211

7.5 Experimental design matrices for a 24−1

IV fractionalfactorial design 2127.6 Effect estimates and sums of squares for the 24−1

IV

fractional factorial experiment 2129.2 Decision variables of the supply chain planning model 2439.1 Sets and indices of the supply chain planning model 2449.3 Input factors of the supply chain planning model 24710.1 Property types deduced for the supply chain model 27211.1 Experimental design matrices for the 24 factorialscreening analysis 30211.2 Results of the 24 factorial screening analysis 30312.1 Sets and indices applied for the CPLP-T (c), CPLP-

T (c T) and CPLP-RISK (c Risk) model 31612.2 Deterministic and stochastic parameters applied forthe CPLP-T (c), CPLP-T (c T) and CPLP-RISK(c Risk) model 31712.3 Decision variables for the CPLP-T (c), CPLP-T (c T)and CPLP-RISK (c Risk) model 31812.4 Deterministic parameters and their values for theillustrative example 32912.5 Solutions of capacity extensions decisions for the il-lustrative example with target service level of 100% 33512.6 Solutions of capacity extensions decisions for the il-lustrative example with target service level of 95% 338

“Have Patience All things are difficult before they become easy.”

Saadi

1.1 Motivation for Beginners

Most of us use the concept of risk frequently in daily language “If you decide to do this, you risk suffering negative consequences.” This

sentence can be referred to a discussion with a colleague at work whentalking about new business strategies, or with your partner whenreflecting on financial precautions or health care, or with yourselfwhen deliberating about whether to eat fish in a sushi restaurant.But what exactly is risk? How can we value risk? How can riskchange our decisions? How can we improve our decisions throughthe consideration of risk? How is risk actually assessed? In terms

of percentage? With the means of a linear scale form 1 to 10 or byusing an ample scale from green to red? For the beginning we startwith a comparison of different risk situations by using relative termssuch as “lower” or “higher”

Imagine you want to assess the risk of not being able to deliver awedding cake to an exclusive restaurant, because you were seriously

© Springer Fachmedien Wiesbaden 2016

I Heckmann, Towards Supply Chain Risk Analytics,

DOI 10.1007/978-3-658-14870-6_1

injured by a car, when you cross the street without looking for proaching vehicles We construct the circumstances that define thissituation in more detail as follows: Assume the road to be a mainthrough road of a small village in the back country Generally, thevolume of traffic is high within the week early in the morning and inthe evening due to commuting, but it is only a small village such thatthe car frequency is moderate Four times a year a local car racing isheld in this region and the main through road of the village is part ofthe route The village youngsters have just recently got their drivinglicense and cruise through the village in the afternoons Additionally,the village is part of a field trial, whose overall objective is to reduce

ap-CO2-Emission Each household is endowed with at least one electriccar, which are by far more quite than diesel- or petrol-engined ve-hicles So, the probability of a car arriving just in the moment youwant to cross the street depends on the day of the year, the day ofthe week and the time of the day The probability of being hit by

an approaching car depends on your ear, your reaction, the drivingskills of the driver, and the weight of the cake If you want to crossthe street on the day of the car race, the probability of being badlyinjured is significantly higher compared to normal weekdays

The analysis of probability is a standard mean to assess risk, but isthe risk of not delivering a wedding cake and being injured by a carhigher or lower, if crossing the street is one of a series of dozen tasksthat you have to pass as a participant of an extreme obstacle course?This challenging run is of course illegal, but frequently attended byextreme athletes from all over the world Other obstacles on this runinclude passing a gill on a slackline and ends with a base jump from arock You are a person who can be described as an adrenaline junkie.Passing a road without looking for cars is nothing you would call arisk Hence, the risk of not being able to deliver a wedding cake andadditionally getting injured is lower for you than for a person thattackles tasks dutifully or a person suffering from the brittle-bone dis-ease, who both try to avoid any danger, when passing a street.How does the risk level change, if you are professional stunt-man justdelivering a wedding cake for your friend, who is a confectioner? You

are booked for years as a double for actors that have not the skillsnor the ability to ride a car accident out – especially not when hav-ing a wedding cake in their hands You have been trained for years

to overcome situations in which fast approaching cars hit you Foryou, crossing the through road of a backwater town while delivering

a wedding cake does not pose a challenge – as long as it is not theday of the car race So the risk of not being able to deliver and beinginjured is certainly lower for you than for a senior citizen But it ishigher, if you are on the run from a rabid dog The dog is wantedfor months as he has killed numerous animals in the neighborhoodand already attacked other people When you started your deliverywalk at the urging of your friend and went through the fields nearby,you startled the dog under a tree while he was eating a blackbird.During your career as a stunt-man you have witnessed several stuntswith mad animals You know that a confrontation with a mad oreven rabid animal can have serious consequences for your well-being,which you want to avoid The dog is fast and if you stop or slowdown before the road, he will certainly catch you The restaurant

is located on the opposite side of the road, if you run across thestreet (without looking), you will certainly find a shelter and be safe.Your individual objective has changed and you now want to ensure

or maximize your physical inviolability, but you also try to deliveryour friends cake anyway Nevertheless, crossing the street of a smallvillage is less dangerous than fighting with a rabid dog

How does the risk level change, if your objective changes again?What is the risk of being injured, when you are on the way to thehospice? You had a great time as stunt-man, but those days havelong since passed The last years you fought against a serious illnessand the doctors give you one or two months before your body willlose this fight During your stay at your friends house you decidedthat it is time to give up You are grateful for having such a goodfriend that you comply with his demand to deliver the wedding cake.When you passed the fields, you killed the rabid dog with a judochop and are now in thoughts while approaching the road on the day

of the car race The risk of not being able to deliver the cake and

being injured or even killed by a car is – with respect to what youhave to expect – considerable low.

And how does the risk of not being able to deliver a wedding cakeand being injured by a car while crossing the street of a backwa-ter town change, if all the aforementioned thoughts do not apply toyour situation, because you are sitting on the veranda of your ranchhouse in the outback of Australia just thinking about different types

of risk? As a child becoming a stunt-man has ever since been yourdream, but you decided to work as research scientist at university.You wrote your PhD thesis about the concept of risk A couple ofdays before submission you were sitting on some stairs at the Uni-versity, felt exhausted and thought of resigning All of a sudden abig nugget fell down on the street You had no idea, where it camefrom, but you felt blessed, finished your PhD and got a great amount

of money for this golden and heavy nugget You bought this ranchhouse in Australia, and became a dropout before you became a be-ginner Now you are sitting on a rocking chair, looking over thefields, and patting the head of your dog, who has an anti-rabies in-oculation Streets and cars are thousands of miles away So, the risk

of not being able to deliver a wedding cake and being injured by acar while crossing a street simply does not exist

But how about the risk being not able to deliver a didgeridoo to rope for the birthday party of your friend, because you were seriouslyinjured by a spider sized less than four centimeters while collectingfire wood in Australia?

Eu-This constructed case demonstrates that the level of risk changeswith the perspective and seemingly depends on specific circumstancesthat need to be known and evaluated prior to risk assessment Theconsideration of risk is often treated as exaggerated, while the riskconcept itself is almost always oversimplified As incidents like theEuropean ash cloud in 2011, Fukushima-Daiichi nuclear disaster in

2011, or the airplane crash in the French Alpes in 2015 highlight,reality is often more erratic and unimaginable than thinking up riskscenarios

1.2 Advanced Risk

In the last decade we witnessed numerous real cases of risk – smallerand bigger ones When these cases occurred we were stunned bytheir consequences or implications Force majeur or blow of fate?How could it happen? Why were we not prepared? Why did we notevaluate the situation correctly in advance? How can we prevent ordiminish such cases or their consequences?

We relate to some cases to evaluate in more detail the perspectivesintroduced in the previous section, which co-determine the level ofrisk Certainly, probability is one major concept of risk However,when talking about probability we immediately enter a highly dis-cussed field People have been arguing about the meaning of “prob-ability” for at least 200 years The major polarization of the notion

is between the objectivist1 and subjectivist2 schools [171, 223, 325].While objectivists believe that probabilities are real, subjectivists as-sume probabilities to reflect the degree of beliefs of rational persons.When applied in the context of risk, probability needs to be care-fully defined While the objective probability of an ash cloud forcingairplanes to stay on the ground was ever since equal to zero, therewas a certain plausibility, a subjective probability, we must affirm

at hindsight Throughout this thesis we follow [146], who combinesboth views and defines probability as subjective in the sense that itdescribes a state of knowledge rather than any property of the realworld; but it is objective in the sense that it is independent of thepersonality of the decision maker Two rational beings faced withthe same total background of knowledge must assign the same prob-abilities

Consider then the risk of a deadly airplane crash Often people start

to argue that the probability of deadly accidents is extremely low

1 Other notions related to the “objectivist’s” perception of probability are quentist”, the “aleatory” or the “physical” notion of probability.

“fre-2 Other notions related to the “subjectivist’s” perception of probability are the

“Bayesian” or “evidential” notion of probability.

compared to fatal car accidents Does that mean the risk is low, ifthe probability is low? When we talk about risk, there must be atleast a chance that “something” happens that may result in “some-thing” negative The problem with the comparison between deadlycar and airplane accidents is the objective of the decision maker.

If he wants to travel with safe transportation means, he could cide to use that mean which is endowed with the smaller probability

de-of deadly accidents However, these probabilities are based on theoverall amount of passengers Naturally decision makers are not in-terested in the welfare of others or of the majority of people, but intheir own welfare If the decision makers have an accident with thetransportation mean they are going to choose, how great is the prob-ability of dying? Clearly, the probability of dying while having anaccident is much higher, than the probability of a deadly accident Itseems to be necessary to use the right statistics, when arguing aboutthe security and the risk of using different transportation means

If the decision makers are the persons in charge of scrutinizing thedata and calculating the probability of a deadly plane crash, theyvalue the risk of dying in terms of absolute or relative numbers Con-trary, if the decision makers are the person in charge of the safety ofthousands of airline passengers, they decide whether this probabil-ity assessment necessitates new precaution actions, i.e the decisionmakers additionally value risk based on their subjective perceptionwith respect to their responsibility If the decision makers are privatepersons, like for example a parent, they value risk by reflecting theconsequences for his family Thus, compared to probability, the level

of risk seems to additionally depend on the degree of the individualinvolvement of the decision maker

The observation that risk is more than just one mathematical termalready resulted in misleading conclusions Right after the incidents

of Fukushima the former federal secretary for the environment, ture conservation, and nuclear safety stated that security and risk areneither a mathematical nor a static concept Instead, they are socialvalues which can change over time [190] Following this argumenta-tion the risk of a nuclear meltdown has tremendously increased solely

na-due to the public perception However, probability calculus indicatethat a nuclear meltdown can happen every 10.000 years per reactor.With more than 400 reactors worldwide a meltdown may happenevery 25 years [336] While the Chernobyl disaster occurred on 26April 1986, the Fukushima MCA began on 11 March 2011, roughly

25 years later So, probability calculations seem to work less, more cautiousness is needed when considering risk Althoughrisk does not seem to depend solely on mathematical calculations, itstill depends on probability

Neverthe-Residual risk refers to a part of the overall risk related to a certainsituation that is hardly predictable or manageable In this sense it

is often used as the scapegoat for everything the decision makers are

not able to manage or do not want to manage The world for

exam-ple faces the risk to be hit by comets, meteorites or even asteroids.When an asteroid impact occurs there might be no necessity for riskassessment, because there is nothing left to manage The walls ofFukushima plant were build to withstand waves of height 5.7m Itseems strange that even though mankind was able to build walls ofthis height in the 700th to 500th century before Christ birth that it

is the fault of residual risk, when the tsunami waves hit the coastoffshore the nuclear plant This waves might have been manageable

A meteor shower can (most probably) not be managed It seems that

it is possible to be better prepared, if risk is assessed appropriately.The prophylactic mastectomy of a famous celebrity in 2013 has longbeen discussed quite controversial in the public Is such a radical step

exaggerated, if only the medical history of close relatives indicates

increased cancer risk, but not the individual examination results? Is

it reasonable to intervene before becoming ill? Discussion whetherthis was a good or a bad decision will not come to an end What

is for sure is that the risk of dying of breast cancer has not onlydiminished considerably, but it has also became extremely low afterhaving had a mastectomy

The willingness to change and learn from risky situations recedeswhen normality returns Risk is more tangible, if it is not theoret-ically discussed, but if the decision makers are struck by its conse-

quences Risk becomes effective, if it intervenes with the personalenvironment of the decision makers and results in situations theycan not bear Then he realizes that he can live with the risk to missthe bus, but not with existential threats which may strongly affecthis lives People want to avoid nuclear meltdowns that destroy theenvironment they live in People want to avoid multi-million dollarlosses from security funds, in which they own shares and whose de-velopments should safeguard the financial situation of their pension.People do not want to be infected with life-threatening illnesses such

as Ebola If the only control is to live without nuclear power, out securities funds, and without traveling possibilities, then theyact accordingly However, as soon as these risks appear to be em-banked and manageable, decision makers start thinking, if we onlycould have cheaper energy, if we could only profit from risky securityfunds, if we could only travel without boundaries It seems that thelonger the immediate threat has passed, the smaller is the (perceived)risk

with-This discussion reveals that the level of risk is endowed with a highdegree of complexity This thesis is about risk, but it focuses on the

meaning and the influence of risk as well as a good risk treatment

within the field of supply chain management Therein, not all, butmost of the aforementioned thoughts are valid and applicable, too.The next section introduces the field of application and discloses howrisk affects the management of modern supply chains

1.3 The Real Introduction: The Name of the Game

A supply chain defines a network of organizations interlinked by

up-and downstream connections Each location is responsible for ferent processes and operations of value adding such as storing, pro-ducing, manufacturing or shipping The goods transfered between

dif-organizations and the goods sold to the final customer can be ucts or services Figure 1.1 shows a typical supply chain, its entitiesand flows Therein raw materials are procured from suppliers, prelim-inary products are produced at distinct or among several productionfacilities, shipped to warehouses for storage and then forwarded tofinal assembly, manufacturing or production By global and regionaldistribution centers or retailers the final goods are transferred to thecustomers Supply chains do not only ship materials or offer services,but they are also responsible for the transfer of financial and infor-mation flows More detailed descriptions of supply chains and theirdefining entities are provided by several standard works, see for ex-ample Chopra and Meindl [55], Simchi-Levi et al [274], and Stadtlerand Kilger [291].

prod-Figure 1.1: A supply chain example, see [291].

The ultimate purpose of a supply chain is to satisfy as many tomer orders as possible Decision makers need to balance the sup-ply and the demand by efficiently allocating resources to the (ex-pected) customer requests Therefore, a lot of decisions have to be

cus-carried out and coordinated These decisions comprise questions like

“Which products should be produced on which machine?”, “In which sequence should products be produced on a machine?”, “Which loca- tions should be responsible for the production of which products?”,

“Which transportation links and modes should be selected for ment?” or “Which production facility should be opened or closed?”.

ship-The more important a decision is, the better it has to be prepared [91].The preparation of supply chain decisions is done by supply chainplanning In accordance with the importance and the time horizon

of a decision, planning tasks are categorized upon three planning els: long-term (strategic), mid-term (tactical), and short-term (op-erational) planning Decisions of the strategic level deal with thedesign of a supply chain and balance different long-term effects overseveral years Tactical decisions concern the assignment of regularprocesses, operations, and flows to available resources They com-prise the determination of approximate quantities and aggregatedtime frames for each resource Decisions about daily activities arecarried out on the operational level

lev-In order to rein the complexity of decision making on each of theaforementioned planning levels it is necessary to abstract from real-

ity The simplified version of reality is called a model Supply chain

models are descriptions of supply chain activities and desired goals

A production process, for example can be described through mation about production capacity, variable production costs, or pro-duction time Information like requested product, requested amount,and delivery date describe an activity of ordering For planning pur-poses a supply chain model is created and solved by computer-basedapproaches

infor-Generally, planning can be regarded as a process of organizing, turing, and scheduling future activities Planning needs informationabout future developments to be able to consider restrictions and toachieve a favored goal A major concern for the planning of any task

struc-or system is the treatment of infstruc-ormation uncertainty Usually sion makers know about the uncertain development of some distinct

deci-information For example expectations about customer demand viate in most cases from the initial outlook Means for the prediction

de-of uncertain information comprise the consideration de-of historic izations or of expert knowledge Figure 1.1 provides an aggregatesnapshot of an exemplary supply chain In reality modern supplychains can be by far more complex Especially over the last decadessupply chains evolved into internationally-acting systems and aresince then caught in a crossfire of additional environmental influ-ences This evolution led to an increase of uncertain informationsand to a broadened range of uncertainty

real-In particular, incidences that lead to sudden and unexpected fications at different locations within the supply chain attracted theattention of decision makers Natural disasters such as earthquakesdestroy production facilities or roads, and forestall the possibility tosatisfy customer’s needs as promised Besides these so-called disrup-tions, unpredictable and slightly aggravating deviations also affectsupply chain’s goal achievement Exchange rate fluctuations, vari-ability of oil prices, or increased labor costs have the potential toreduce the profit margin and hence the competitive advantage ofsupply chain partners Thus, supply chain disruptions and unknowndeviations impede the availability of resources, the realization of theplan, and consequently the satisfaction of customer demand Theconsideration of perils that have the potential to derogate the sup-

modi-ply chain is carried out within the research field of supmodi-ply chain risk management.

In order to prepare for uncertainty supply chain models need to be

endowed with the information about uncertain developments ulation models, for example, re-enact decisions for supply chain pro-

Sim-cesses, operations, and flows Algorithms simulate random tions of the uncertain information, each of which provoke differentdecisions In the context of supply chain risk management simula-tion models provide mechanisms to understand the relationship be-tween alternative realizations of uncertainty and their consequences

realiza-on the supply chain However, with the goal not to evaluate known

solutions but to determine risk-aware planning solutions, matical optimization models and techniques are needed Those ap- proaches are often derived from the discipline of operations research.

They determine solutions by maximizing or minimizing a matically formulated goal, such as cost-minimization or service-level-maximization

mathe-Over the last years a considerable number of approaches has beenpublished within the field of supply chain risk management However,what is still missing is a structure that embraces these approaches

or serves as a guideline for future research The big interest forour critical review of supply chain risk [122] reveals the desire foranchors and guidance This thesis intends to provide further steps

on the research line of supply chain risk and strives for reducing thegap within the field of supply chain risk by approaching the answers

to the following questions:

• What is supply chain risk?

• How can simulation models support the understanding of the tent of supply chain risk?

ex-• How can optimization models determine risk-aware supply chain designs?

1.4 Outline and Course of Discussion

With the overall goal to answer the three questions posed in theprevious section this thesis consists of three parts:

I Fundamentals of Supply Chain Risk,

II Simulation-based Approaches for Supply Chain Risk tion and Assessment, and

Identifica-III Optimization Approaches for Strategic Supply Chain Risk gation

Miti-Figure 1.2 highlights the structure of this thesis The three parts areframed by an introductory discussion, Chapter 1, and a concludingdebate, Chapter 13 Each part consists of one or several chapters.The methodological purpose of each chapter within the overall con-text of this thesis is highlighted by the legend.

The chapters of Part I review existing literature and (re-)define sential concepts The discussion combines critical reviews of exist-ing literature with own thoughts The combination is intended toprovide a clear structure of different topics and to deduce new def-initions and conceptual approaches Contrary, the main chapters

es-of Part II are formed traditionally: The first two chapters providemethodological explanations and reviews, chapter three introducesthe new simulation approach, which is discussed by an evaluationchapter Additionally, two further chapters introduce accompanyingresearch Part III consists of one chapter, which deduces a new risk-aware stochastic optimization model for supply chain network designproblems

Chapter 2 introduces the genesis of supply chain risk It presentstrends that discover the vulnerability of modern supply chains to-wards uncertain deviations and classifies prominent disruptive trig-gers It motivates the need for supply chain risk quantification andtraces the path towards supply chain risk management Chapter 3provides a careful analysis of existing supply chain risk definitionsand related concepts, discusses rationales of risk concepts from se-lected domains of application, and derives core characteristics thatdrive nowadays supply chain risk understanding The chapter closeswith a new definition of supply chain risk The subsequent Chap-ter 4 consists of a theoretical discussion on common risk biases andmain elements of supply chain risk analysis Conceptual definitionsare provided along with logistical implications Chapter 5 coins thenew concept of supply chain risk analytics It deduces analytical in-sights into the capability of supply chain risk analytics to determinerisk-aware supply chains Risk measures, optimization models and

SimSCRF A real case evaluation of a representative supply chain

de-rived from chemical industry offers logistical insights and highlights

the usefulness of the developed approach compared to prevailing riskapproaches Chapters 9 and 10 represent accompanying research.Chapter 9 develops a representative master planning model that re-

places the (black box) proprietary planning tool within the SCRF approach Chapter 10 aligns the data model for supply chain

Sim-planning with the simulation-based representation of a given supplychain

With the goal to incorporate the new supply chain risk definition notonly in simulation-based, but also in optimization approaches, a risk-aware optimization model is developed during the course of Chapter

12 By extending the capacitated plant location problem a integer two-stage stochastic programming model is presented alongwith analytical and computational evaluations A new concept toanalyze potential risk-aware optimization solutions is defined as thevalue of risk consideration The usefulness in terms of operational-ization of supply chain risk and methodological appropriateness isdiscussed to close Part III

mixed-In summary, the purpose of this thesis is to provide:

• fundamental knowledge about elements of supply chain risk andtheir dependencies by defining supply chain risk, supply chain riskanalysis, and supply chain risk analytics,

• a new simulation-based procedure for supply chain risk tion and assessment by developing a framework for supply chainrisk analysis and deducing managerial insights within a real caseevaluation, and

identifica-• new mathematical models to determine risk-aware supply chaindesigns by considering supply chain risk as a holistic concept

Supply Chain Risk

Concepts – Fundamentals

“When anyone asks me how I can best describe my experiences in nearly 40 years at sea, I merely say, uneventful Of course, there have been winter gales, and storms and fog and the like, but in all my experience I have never been in any accident of any sort worth speaking about [ ] I never saw a wreck and have never been wrecked, nor was I ever in any predicament that threatened to end in

disaster of any sort.”

Edward John Smith Captain in command of the RMS Titanic

Supply chain officers may feel with Edward Smith After years of

uneventful supply chain management and after years of striving after

more efficient processes, unexpected and sometimes even devastatingevents have derogated supply chains A series of major disruptionslike Hurricane Katrina, piracy attacks offshore Somalia, global finan-cial crisis, flooding in Thailand, European ash-cloud, Japanese earth-quake and tsunami among others have revealed a missing prepared-ness within today’s supply and distribution networks [248] Thus,the management of so-called supply chain risks became an issue.This chapter briefly summarizes the genesis of supply chain risk Itstarts with explaining the trends that revealed the sensitivity of mod-

© Springer Fachmedien Wiesbaden 2016

I Heckmann, Towards Supply Chain Risk Analytics,

DOI 10.1007/978-3-658-14870-6_2

ern supply chains, it revises different classes of disruptive triggersand it traces the path from the consideration of disruptions to sup-ply chain risk management and the need for quantification.

2.1 Logistics Innovations – A Blessing and a Curse

The strategic influence of supply chain management on business formance – including not only overall logistics costs, but also cus-tomer satisfaction – is well confirmed by companies of almost allindustries Domestic firms can offer products worldwide, and hence,they compete not only with local companies on the domestic market,but also with international competitors they encounter on the worldmarket In order to remain competitive, while benefiting from logis-tics potentials, companies strive for improving and streamlining theiroperational processes [329, 334] As revealed by Computer ScienceCorporation in 2004, companies availed themselves of the implemen-tation of efficiency-increasing logistics innovations While 52% of theconsidered companies registered an increase in their revenues, 72%stated to benefit from the new developments implemented in theirsupply chains Besides the positive effects American AMR amongothers concluded that these new trends and strategies have a nega-tive counterpart, which is highlighted by the increasing number ofdisrupted supply chains [66, 129, 149, 334, 348] Highly efficient op-erations [329] expose supply chains to an environmental crossfire ofdifferent volatile influences While disruptions do not occur everyday,

per-supply chain strategies – also called risk drivers – that lead to

dis-ruptions do When a firm takes a pure cost minimization approach

in order to increase overall efficiency, it reduces excess capacity andinventory, which could make up for production losses caused by dis-ruptions Formerly isolated events within the supply chain can today

“escalate to wide scale network disruptions” [71]

Innovation Trend Opportunity Threat

Globaliza-tion LCCS 20% to 30% lower

ma-terial and labor costs

longer lead times

supply closer to ufacturing and cus- tomer sites in emerg- ing markets

man-increased risk of ply disruption and transportation capac- ity and performance issues

sup-exposure to new litical, security, regu- latory, tariff, and cur- rency risk

po-Outsourcingimprove operating

performance and service levels

limited visibility or control of service lev- els or selection and performance of sub- tier suppliers

lower operating costs Informa-

di-improve service and fill levels

Integrated Supply

reduce management burdens

increased reliance

on single supplier for broader range of materials and services access "one-stop

streamline tions

opera-increase risk of stockouts and manu- factirung disruptions due to supply or delivery glitches Supply

Base Rational- ization

improve spend age

lever-increased reliance on fewer or even sole- source suppliers

reduce management burdens

links performance to financial and opera- tional health of sup- plier

improve strategic supply relationships

increased likelihood of dual- or sole-sourced suppliers relying on single sub-tier sup- plier

Table 2.1: Opportunities and risks of prevailing logistic best practices

[see also 2, 63, 249, 329].

Popular logistics improvements, which can be derived from threemain sources: globalization, lean management principles, and in-creased information availability, are presented in Table 2.1 [2, 63,

249, 329]

Global sourcing enables companies to follow strategies like low-costcountry sourcing (LCCS), or outsourcing and off-shoring, all of whichenable companies to implement cost reduction actions and to fo-cus on their core activities Stretched lead times, limited visibility,and difficult communications can, however, decrease flexibility andresponse time in case of supply chain failures Lean managementprinciples like Just-in-Time or Just-in-Sequence and supply base ra-tionalization allow companies to synchronize supply with production