Increasing Resilience in Production Networks: A Practical Approach

Based on Scenario Planning and Simulation-Based Capacity Analysis

David Kunz

, Tim Maisel

, Andreas Kunze

and Jörg Franke

Institute of Factory Automation and Production Systems (FAPS), Friedrich-Alexander-Universität Erlangen-Nürnberg,

Egerlandstraße 7-9, 90158 Erlangen, Germany

Siemens Healthineers AG, Siemensstr. 3, 91301 Forchheim, Germany

Keywords: Resilience, Simulation, Scenario Planning, Simulation-Based, Capacity Analysis, Global Production

Networks, International Supply Chains, Case Study.

Abstract: In the current global economic landscape, companies with an international presence face the challenge of

ensuring that their production networks are not only efficient but also resilient to unpredictable events. Recent

technological advancements and the close integration of global production networks have been increasingly

disrupted. During times of global crises, it becomes evident that traditional approaches are no longer sufficient.

Therefore, the focus is shifting from reactive measures to proactive prevention. This paper presents a novel

approach for increasing resilience in a production network based on a combination of systematic foresight of

unpredictable events using scenario planning and a simulation-based capacity analysis for the identified

scenarios. To demonstrate and validate the application of the proposed approach, a case study for the

production network of a large German healthcare company is conducted and presented.

1 INTRODUCTION

The integration of globally spread-out production

sites and rapid technological advancements have led

to increased optimization of resource utilization.

However, this has also increased vulnerability to

interruptions in production. To combat these issues,

the focus has shifted from reactive measures to

proactive prevention. (Schollemann et al., 2022)

The term resilience is frequently used in both

natural and social sciences (Hoffmann, 2017), and has

more recently been applied to organizations and

production networks. Resilience is commonly

defined as the measure of the persistence of systems

and their ability to absorb change and disturbance, as

defined in (Holling, 1973).

Overall, resilience is not only about getting back

to the initial state after a failure occurs but also means

to adopt to the changing circumstances (Rydzak et al.,

2006). This especially relevant as some failures may

be inevitable in a complex and dynamic world.

Organizational resilience depends on the

organization and its circumstances. For example, a

financial company may encounter significant

https://orcid.org/0009-0002-8801-8231

https://orcid.org/0000-0003-0700-2028

challenges due to a loss of trust, whereas a produc-

tion-focused company may face more substantial

difficulties in the event of a key supplier’s failure

(Seville, 2008). (McManus, 2007) proposes an

expanded definition of organizational resilience that

encompasses coping with both day-to-day business

problems and longer-term change-related issues.

According to their model, resilience in manufacturing

organizations is contingent upon three factors:

situational awareness, effective management of key

vulnerabilities, and adaptability in a complex,

dynamic, and interconnected environment.

(Zhang & van Luttervelt, 2011) describe resilien-

ce in the context of manufacturing and production

systems which are depicted as a network that contains

not only directed but also undirected connections.

The authors describe five different types of failures in

a production system: oversatisfaction of demand,

inability to satisfy demand, unavailability of

resources to meet demand, damaged infrastructure,

and operations damaging internal systems. For the

authors, organisational resilience differs from the

resilience of a production system in the sense that a

158

Kunz, D., Maisel, T., Kunze, A. and Franke, J.

Increasing Resilience in Production Networks: A Practical Approach Based on Scenario Planning and Simulation-Based Capacity Analysis.

DOI: 10.5220/0012830100003758

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 14th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2024), pages 158-169

ISBN: 978-989-758-708-5; ISSN: 2184-2841

stable state is not a necessary condition for the

success of a production system.

Global Production Networks (GPN) are open-

ended systems with complex links and multiple

independent actors whose goals may be in direct

conflict. They depend on a variety of internal and

external influencing factors that change dynamically

and require adaptions of the design of GPNs (Váncza,

2016). Therefore, managing GPNs involves not only

organizing production sites within the company,

which are often heterogeneous and require

differentiated consideration (Erlach et al., 2023), but

also managing strategic partners, suppliers, and

external influences from global political and

economic sources (Henderson et al., 2002).

For these reasons, companies require practical

approaches for the assessment of their vulnerabilities

and identification of potential avenues for increasing

their resilience based on their specific circumstances.

This paper argues for the close integration of GPN

simulations into the analysis of potential strategies for

increasing resilience through scenario planning. This

integration provides a more comprehensive under-

standing of the relationships within the model, reveals

the underlying assumptions and produces quantitative

estimates for developed scenarios (Paich & Hinton,

1998). In order to reduce the time and resources

required for simulation and evaluation, it is essential

to identify the most relevant and crucial scenarios at

the outset.

2 STATE OF THE ART

A variety of approaches for simulating GPNs are

discussed in existing literature. (Peukert et al., 2023)

present an approach for optimizing responses to

disturbances in GPNs using simulation models and

control circuits. The authors emphasize considering

production- and logistics related countermeasures

first. Proactive strategies are evaluated based on a

simulation model of the production network, with

experiments comparing the performances under

different conditions: (1) without disruptions and

countermeasures, (2) with disruptions, and (3) with

disruptions and countermeasures.

(Alexopoulos et al., 2023) introduce the

framework 'FLEX4RES', which supports the

reconfiguration of production networks to achieve

resilient production value chains. The introduced

platform enables the integration of live data from

production based on Gaia-X and Asset Administrati-

on Shell.

(Tan, 2020) compares various complex systems

approaches for modelling and simulating supply

chains to enhance their resilience. The evaluation is

based on modelling real supply chains and developing

mitigation and contingency strategies based on

hypothetical scenarios.

The authors in (Ivanov, 2018) investigate

disruption propagation in supply chains to increase

resilience with consideration of sustainability factors

and employ simulation to assess the impact of various

factors.

(Carvalho et al., 2012) presents a case study on

the redesign of a supply chain of a Portuguese auto-

maker to increase resilience using simulation. The

study examines two common strategies, the creation

of redundancies and of flexibilities, and evaluates six

scenarios. The simulation is restricted to the supply

chain and does not consider the impact of material

shortages on production processes.

Scenario Planning is a popular approach to

managing uncertainty in strategic planning. The focus

is on creating awareness and preparing for uncertainty

and disruptions (Cordova-Pozo & Rouwette, 2023)

which is why it has been applied to identify ways to

improve resilience in disaster mitigation (Debnath et

al., 2024).

To summarize, the existing literature mainly

focuses on simulating production networks or exami-

ning the impact of selected scenarios and counter-

measures on production networks. As resources for

increasing resilience are limited, it is essential to

identify and prioritize the most critical elements of a

system (Balakrishnan & Zhang, 2020). The

applicability of the presented approaches for

production companies seeking to enhance their

resilience is limited due to the lack of consideration

given to whether the examined failures in the

production network are relevant to the specific

companies. On the other hand, current approaches for

using scenario planning focus on areas outside of

production networks.

This publication aims to address these issues by

providing practitioners with guidelines for how to

identify areas of potential for increasing resistance in

production networks and to evaluate the scenarios and

possible countermeasures using simulation.

3 METHODS AND APPROACHES

The above examples of different strategies for

increasing resilience in a production network clearly

show that no strategy has so far been useful as a basis

for the stated objectives. Among other things, there is

a lack of reference to the capacity performance of a

production network, the rapid adaptation and

mapping of a complex GPN and the quantitative

analysis of various resilience strategies. For these

Increasing Resilience in Production Networks: A Practical Approach Based on Scenario Planning and Simulation-Based Capacity Analysis

159

reasons, a separate approach for increasing resilience

in a production network is introduced in this

publication using simulation-based capacity analyses.

As a full-factored resilience optimisation of a

GPN is a highly resource-intensive process, the

objective of the introduced approach is to initially

identify the most relevant levers. The scenario

development approach was selected to determine

these levers because of its systematic approach and

widespread application. The proposed approach

consists of three phases, as shown in Figure 1.

After specifying the project, scenarios are syste-

matically developed based on scenario planning

technique by (Gausemeier et al., 1996) in the second

phase. The identified scenarios are then modelled,

and possible countermeasures are developed and

evaluated in the third phase. Following (Cordova-

Pozo & Rouwette, 2023) this approach can be charac-

terized as following the probabilistic modified trend

school using a cross-impact analysis technique.

Figure 1: Proposed three-phase approach for increasing

resilience in production networks with forecasting

scenarios (blue) based on (Gausemeier et al., 1996) and

modelling and evaluating the scenarios (green) extended by

artefacts.

3.1 Concept for Forecast Generation

The developed approach for the forecasting of scena-

rios is based on the general scenario planning process

outlined in (Gausemeier, Fink, & Schlake, 1996).

This process has been adapted to meet the

requirements of highly dynamic and specific

production environments, refer to Figure 1.

One of the main differences is that the developed

scenarios are transferred to simulations where impact

analysis, contingency planning and robustness

planning are integrated, see chapter 3.2 for details.

The second major difference is that specialized

software catalogues and databases are not used. This

ensures that the developed forecasts are tailored

precisely to the specific circumstances of the

respective production network and are not influenced

by predetermined catalogues and databases.

Additionally, the generation and selection of base

scenarios and the visualization of the future space

should not be restricted by specialized software.

The division of tasks represents the third

significant difference. In (Gausemeier et al., 1996)

most steps are performed by an external scenario

team. The proposed process in contrast is

characterized by the cooperation of three essential

groups of actors: the project lead, main stakeholders

and the extended group of experts. The project lead is

in charge of organization, structuring and

visualization of results, whereas the main

stakeholders consist of (internal) costumers and end

users of the developed scenarios and simulations

which are very involved in the creation of the

scenarios and application of the results. The extended

group of experts is made up of experts from various

specialized areas of the production network. Ideally,

members of this group are selected by the project lead

and main stakeholders without restraints. All groups

can be employed internally in the production network

or externally, e.g. by suppliers, customers or

consulting companies, as not every production

network has the capacity for an internal scenario team

or is allowed to disclose the sensitive information

externally. The following chapters describe the

procedure for preparing the forecasts in detail.

3.1.1 Project Description

At the beginning of the project, the project description

should give a precise definition of the target values

and the intended utilization of the scenarios within the

project’s context. This involves specification of the

goals for the scenario planning as well as the

simulation project in close cooperation with the main

stakeholders. For the former, it is recommended to

follow the specification guidelines for scenario

planning in (Fink & Siebe, 2016) which include the

target audience, forecasting goals, scenario field

parallelization, scenario depths and forecasting

timescale. The necessary specifications for the

simulation project should be defined, e.g. based on

Project description

Design field analysis

Raw scenario

creation

Future space

mapping

Project plan

& specification

Design field

Grouped infl.

factors

Connectivity

matrix

Projections

Consistency

matrix

Raw scenarios

Future space

Scenarios with

description

Artefacts

Identification of influencing factors

(in joint workshop)

Connectivity analysis

(asynchronously)

Consistency analysis

(asynchronously)

Selection of projections

Extended

group

of experts

Main

stakeholders

Project lead

Scenario description

Development of Scenarios

Modelling of base

production network (PN)

Including experiments for

scenarios and countermeasures

Main

stakeholders

Project lead

Simulation

expert(s)

Evaluation of potentials

for increasing resilience

Modelling & Evaluation

Model with

scenarios

Understanding

of levers for

increasing

resilience

Simulation

model of PN

Planning

SIMULTECH 2024 - 14th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

160

VDI 3633. This step also involves selecting the

members of the extended group of experts and

creating the project plan.

3.1.2 Design Field Analysis

The aim of the design field analysis is to identify the

components that constitute the scope of action. In the

following steps, these components are analysed in

specific scenarios. Therefore, it is crucial to identify

them and their boundaries precisely. The focus of the

design field analysis can vary depending on the

project scope, e.g. a company, product, technology or

global design field. (Gausemeier et al., 1996)

In the context of a production network, the design

field analysis focuses on identifying the sub-

processes to be included in the scenario project and

those lying outside the project framework.

For instance, sub-processes such as production

sites and suppliers within the production network may

be included in the scenario project, as their future

development is crucial for the network's performance.

Sub-processes outside the project scope may include,

for example, external customers or the development

of new products, as these areas have little influence

on the capacity of a production network.

In addition to the central sub-processes, the design

field analysis also identifies focus areas. A focus area

represents the specific points within a design field to

which scenario development should be aligned.

While focus areas can overlap within the network, it

is crucial to differentiate them clearly in their

objectives within the production network. One

example of such an area might be the 'availability of

qualified personnel', while another could be the

'attractiveness of the location'. Despite the clear

demarcation of these areas, it is possible that they

influence each other.

The design field analysis is conducted by the core

project team which includes the project lead(s) and

the main stakeholder group.

3.1.3 Identification of Influencing Factors

The third step in the process involves identifying

influencing factors through a workshop including the

extended group of experts. Participant are assigned to

a specific focus area based on their experience and

position in the company. Each participant indepen-

dently collects influencing factors for their assigned

focus area through brainstorming.

The participants then share and discuss their

results, categorizing their collected influencing

factors into grouped influencing factors within each

design field. Grouped influencing factors are made up

of very similar or identical influencing factors and

reflects the focal points within a focus area. In the

third step, the grouped influencing factors are

collected amongst all groups and refined and

developed further. The workshop aims to develop

comprehensive factors for each focus area, by

identifying detailed influencing factors. The process

involves brainstorming, categorization and refine-

ment to set comprehensive priorities.

As preparation for the following steps it is

essential to have a shared definition of the groups of

influencing factors. Therefore, a description is

created for each group of influencing factors after the

workshop.

3.1.4 Connectivity Analysis

To identify the areas of influence with the greatest

impact on the production network, a connectivity

analysis consisting of a matrix of direct influence

(MDI) and a significance matrix is carried out. The

proposed method recommends an asynchronous

approach instead of performing the analysis in a

workshop, as exchange between the participants is

not required which allows more extended experts to

participate in this step.

An MDI is used to assess the influence of the

grouped influencing factors on each other by evalua-

ting the extent to which the influencing factor in the

row influences the influencing factor in the column.

Based on the evaluation, the four characteristic values

active sum, passive sum, dynamic index and momen-

tum index can be calculated in the completed MDI.

In addition to the behaviour of the influencing

factors, the importance of each factor is assessed

using a relevance matrix, which identifies the more

relevant influence for each pair of influences.

With the two matrices, a connectivity and rele-

vance score is calculated for each influencing factor.

Their ranked results are visualized as Connectivity-

Relevance-Grid in Figure

2, based on which key

factors with high connectivity and high relevance can

be identified. Furthermore, other possible key factors

with high connectivity but low relevance and vice

versa can be determined. (Fink & Siebe, 2016)

When identifying the key factors within a

production network, the key factors with a high

degree of relevance and a high degree of connectivity

should be used for a targeted prioritization and focus

on the performance of the production network for the

further process. Key factors with high relevance and

low interconnectedness should not be neglected as

they can have a major impact on the performance of

Increasing Resilience in Production Networks: A Practical Approach Based on Scenario Planning and Simulation-Based Capacity Analysis

161

the production network despite their low connectivity.

Influencing factors with high connectivity but low

relevance can be neglected for the rest of the process,

as they influence the other factors but have little

discernible impact on performance.

Figure 2: Connectivity-Relevance-Grid based on (Fink

& Siebe, 2016).

3.1.5 Creation of Future Projections

Building on the identified key factors, possible states

for each are determined and described in the form of

future projections. The creation of future projections

is a joint task of the project lead and main

stakeholders. A variety of information sources are

used, including internal information on risk

management in the production network, current

trends, company reports and external information

sources such as scientific reports and publications.

Three future projections are developed for each key

factor, each representing a positive, a neutral and a

negative development of the key factor.

These projections serve an essential role in

informing future decisions, as subsequent scenarios

are developed and refined based on them. The future

projections are deliberately exaggerated in order to

describe the projections in as much detail as possible.

It is very important that these projections are distinct

and unambiguous, reflecting potential developments

of key factors without emphasizing the causes or

consequences. Each projection for a key factor should

cover the full range of possible developments and be

compatible with the other key factors. Failure to do

so, may result in inconsistencies in the subsequent

analysis when establishing consistency between the

key factors. (Fink & Siebe, 2016; Gausemeier et al.,

1996)

Particular attention should be paid to adapting the

descriptions to the production network’s specific

circumstances in order to avoid an overly general or

overly dramatic presentation of the future scenarios.

3.1.6 Consistency Analysis

The main objective of the consistency analysis is to

assess the consistency of the future projections for

various key factors. Each participant rates the

consistency of the future projections between the

different key factors. The assessment is done for pairs

of key factors and on a scale from 1 (complete incon-

sistency) to 5 (strong mutual support). It is important

to emphasize that this assessment only evaluates the

consistency between the respective pairs of future

projections and not their probability. The individual

consistency assessments are then summarized by

averaging to form an overall consistency analysis.

Various possible projection bundles can be created on

the basis of this analysis, whereby a projection bundle

consists of the combination of exactly one future

projection for each of the key factors. (Fink & Siebe,

2016; Gausemeier et al., 1996)

The consistency analysis is carried out by all

members of the project team and the ratings entered

into a matrix. As three future projections are created

for each key factor, 3

projection bundles must be

examined as part of the consistency analysis.

3.1.7 Creation of Raw Scenarios

Due to the large number of projection bundles, it

becomes necessary to reduce them. If at least one

combination within the scenario bundle is evaluated

as inconsistent, it is excluded from further considera-

tion. Here, the use of support through algorithms is

essential. The probability of the projection bundles

occurring deliberately plays no role here, as this

process is more about thinking ahead than making

concrete predictions. In a second step, the remaining

projection bundles are sorted according to their

consistency values. The consistency value of a

projection bundle is the sum of the consistency

ratings of all the future projections contained in this

bundle.

3.1.8 Future Space Mapping

To effectively communicate the results and provide a

quick overview of the structure of the resulting raw

scenarios, the raw scenarios are presented graphically

as part of the future space mapping. Various display

options may be suitable, such as diagrams, portfolios,

biplots, shapes, colours and symbols, which can be

used flexibly depending on the application. (Fink

& Siebe, 2016; Gausemeier et al., 1996).

Relevance (Rank)

Connectivit

(

Rank

)

Clear

key factors

Possible additional

key factors with

high relevance

Key factors

with high

connectivity

Key

factors

SIMULTECH 2024 - 14th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

162

3.2 Derivation of a Simulation-Based

Strategy

The methodically prepared qualitative scenarios by

themselves are not sufficient to comprehensively

analyse the complex and dynamic interactions

between the various key factors of a production

network and evaluate future strategical decisions.

Simulations can be used to convert these qualitative

scenarios into quantifiable estimates by depicting

complex interactions over time and thus supporting

decisions. In addition, a simulation model allows to

quantify the effects of changing variables on the

overall system by conducting multiple experiments.

Figure 1 showcases the proposed approach for

integrating the developed scenarios into simulations

of the production network for creating a holistic

simulation-based strategy which is loosely based on

(Shannon, 1998). Note that the extended group of

experts from chapter 3.1 is replaced by simulation

expert(s) which may also be a member of the main

stakeholders or experts. The following chapters

introduce the involved steps in more detail.

3.2.1 Modelling of Base Production Network

In the first step of modelling and evaluation, the base

model of the GPN is created by identifying the

relevant variables and specifying their relationships.

A graphical representation of the simulation model

and its results can facilitate communication with

internal and external stakeholders. It is important to

maintain a simple structure of the simulation to avoid

losing focus. Collecting the relevant information and

data is the most time-consuming aspect of modelling

(Kunz et al., 2023). Therefore, it is advisable to

involve the relevant parties early on and begin to

consider collecting information on the network even

before the final scenarios are determined.

Continuous verification and validation (V&V)

must be employed in the creation of the simulation

model. The simulation results must be plausible and

the simulation model checked for its representative-

ness for the real production network in relation to the

defined target. This usually involves the integration

of main stakeholders and key experts of the

production network.

3.2.2 Modelling of Scenarios and

Countermeasures

Once the base simulation model is verified and

validated, the scenarios and countermeasures are

implemented to enable the comparison of the base

network, with scenario and countermeasures, also see

(Peukert et al., 2023). The exact implementations

may vary depending on the specific network,

scenarios and software but it is advised to realise them

as modules for each key factor which can then be

reused across scenarios.

Feasible countermeasures are collected

throughout the entirety of the proposed process as

they come up and included here. To enable users to

influence the severity and timings of scenarios and

countermeasures, the creation of a suitable interface

is key. An appropriate visualisation of the scenarios’

results is key for understanding and discussing them

(Feldkamp et al., 2020). V&V should be employed

throughout this process.

3.2.3 Identification of Potentials for

Increasing Resilience

Using a suitable interface, simulation expert(s), main

stakeholders and other key experts are able to test the

scenarios by changing variables including effects and

timings. Users have the opportunity to test different

strategies and their impact in preparation or response

to these scenarios. This is an iterative process where

new solutions may arise during the assessment and

are implemented and tested again. Depending on the

project goals, this exploration may additionally

involve optimization algorithms.

The comprehensive analysis and modelling of the

production network and scenarios enables a deeper

understanding of the process dynamics and helps to

identify and optimize potential bottlenecks and

inefficient elements in the production chain. More

informed strategic decisions can be made based on

the explored solution space.

4 APPLICATION AND RESULTS

The proposed approach is implemented for a practical

case study in a part of the production network of a

large German company. The following chapters give

an overview of the results of this application and the

lessons learned.

The case study examines a major product line in

a German healthcare company. Production is distribu-

ted across three locations in Germany, the United

Kingdom, and China. Some products are exclusive to

one site, while others are produced at multiple

locations.

4.1 Forecasting Scenarios

In the first stage, the fundamental scenarios to

Increasing Resilience in Production Networks: A Practical Approach Based on Scenario Planning and Simulation-Based Capacity Analysis

163

increase resilience are determined following the

approach outlined in section 3.1.

4.1.1 Project Description

During the initial stage of the project, the project lead

and main stakeholders defined its context and objecti-

ves. The project's results are intended for internal use

within the company. They aim to include a

methodical elaboration of the main influences on the

production network by internal experts and a

transparent representation of the same.

Furthermore, it is considered crucial to test

potential resilience-enhancing measures through

simulation-based analyses to inform the planning of

various projects. Alternative countermeasures within

the possibilities of the production should be

demonstrated for the developed scenarios and quanti-

tatively and objectively evaluated. The alternative

scenarios should relate to network capacity and

demonstrate various capacity related options.

The depth of the scenarios should be based on

industry scenarios and appear plausible and not too

generic within the production network. With a time

horizon of five years, it is particularly important to

disregard unplanned everyday situations within the

production, such as machine breakdowns, or new

product developments. Rather, the focus should be on

unforeseen future situations that have not been

adequately considered and for which no

countermeasures have been developed yet. This

strategic focus aims to ensure the long-term resilience

of the production network and identify potential risks

at an early stage.

The acceptance criterium for the simulation is a

realistic portrayal of the production network. The

focus is on the production network with its production

sites with modelling up to basic processes. Inputs for

the simulation are market conditions, existing orders,

delivery times, production capacity, and lead times.

Its outputs are the production capacity, production

volumes, and lead times. The results should be

visualised in graphs. Furthermore, the members of the

extended group of experts from process planning,

purchasing, supply chain management and

production planning are determined at this stage.

4.1.2 Scope Analysis

As described in section 3.1.2 a careful definition of

the specific processes and focus areas is required to

further specify the scope of the analysis. The

processes are derived from an overview of the

production network, while the focus areas are defined

through discussions with the main stakeholders. In

the process, potential risks in other global companies

with comparable production networks were

considered, as well as specific product-related risks.

The key focus areas identified are geopolitical

factors, the supply chain, and production operations.

In this context, geopolitical factors refer to capacity-

dependent influences within the network. The supply

chain includes all factors between the sites, while the

production operation includes all internal factors

within the sites.

4.1.3 Identification of Influencing Factors

The identification of influencing factors is based on a

workshop including the extended group of experts as

described in chapter 3.1.3. Ten people from the

company and university participated in the workshop

which was conducted as a hybrid event. The

participants were assigned to the respective focus

areas according to their expertise. Overall, nine

unique influencing factors are identified and

duplicates such as 'availability and productivity of

employees' assigned to a single focus area, see Figure

3 for a summary.

The identified and clustered influencing factors

were then described in more detail by the project lead

in close coordination with the main stakeholders. For

example, 'Reliability and quality of suppliers' is

specified as 'The reliability and quality of suppliers

are of fundamental importance. A global company is

dependent on its suppliers for materials and services,

so close co-operation and evaluation of suppliers is

necessary.'

Figure 3: Resulting grouped influencing factors (grey)

allocated by focus area (light green).

4.1.4 Connectivity Analysis

The connectivity analysis was conducted using the

collected and specified influencing factors. The

members of all groups filled out a MDI and relevance

matrix, the results of which are summarized in Figure

4. The key factors with high relevance and

connectivity are trade policy, suppliers, and operating

resources.

Design Field of Production Network

Geopolitical Factors

Raw Materials

Supply Chain Production Operation

Operating Resources

Trade Policy Production Site

Political Stability Employees

Logistics

Suppliers

Market Factors

SIMULTECH 2024 - 14th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

164

Figure 4: Results of the connectivity analysis visualized as

Connectivity-Relevance-Grid by rank. Labelling based on

Figure 3.

As stated in section 4.1.4, we considered influencing

factors with high relevance but low connectivity, such

as access to materials and raw materials (C) and

productivity and performance of employees (G).

However, we ultimately excluded them from further

analysis because the focus of the analysis is on the

connectivity of influential factors, and these two

factors are already part of existing strategic

considerations and measures.

4.1.5 Creation of Future Projections

Future projections are generated by outlining a

positive, neutral, and negative scenario for each key

factor. Special consideration is given to the unique

circumstances of the production network to prevent

an overly general or exaggerated presentation. Table

provides an example of the three projections for the

key factor ‘quality and performance of suppliers’.

Table 1: Projections for the key factor 'quality and

performance of suppliers' with one positive, one neutral,

and one negative development.

Quality and Performance of Suppliers

Improved supplier quality through transparent

exchange of information. The innovative

strength of suppliers leads to new and improved

roducts and services that further optimise

production processes.

Continuous supplier optimisation for stable

roduction. The innovative strength of suppliers

remains constant, continuing to provide the

same quality of products and services that help

to maintain current production levels.

Impairment of production plans by unreliable

suppliers due to internal or external factors. A

lack of innovation among suppliers means that

they are unable to keep pace with changing

requirements and technologies.

4.1.6 Consistency Analysis

The consistency analysis is performed for the key

factors along with their descriptions. A matrix of

projections ranging from total inconsistency (1) to

strong mutual support (5) is evaluated by the partici-

pants of the initial workshop. The averaged results are

presented in Table

The range of values and relative consistency of

the evaluations suggest a well-developed set of key

factors and their projections. The medium values,

ranging from 2 to 4, are mostly independent of each

other. The main focus for further analysis is on the

pairs with high connectivity as this indicates a highly

connected and complex dynamic of the key factors.

Table 2: Resulting influencing factors allocated by key

factors. Positive projections indicated by plus, neutral by

circle and negative by minus symbols.

KF Raw. Mat. Suppl. Op. Res.

Key factor (KF) Dev. + ◯ – + ◯ – + ◯ –

Raw Materials

(Raw. Mat.)

◯

–

Suppliers

(Suppl.)

+ 4,7 2,3 1,5

◯ 2,2 3,8 4,2

– 2,5 4,0 1,0

Operating

resources

(Op. Res.)

+ 3,5 2,3 4,0 4,0 2,7 2,0

◯ 1,5 4,0 1,8 1,8 4,7 4,2

– 1,8 3,7 2,0 2,0 3,8 4,8

4.1.7 Creation of Raw Scenarios

In the next step, raw scenarios are created from the

key factors and their consistency. An algorithm is

used to create projection bundles that combine future

projections of each key factor. In this case, there are

27 projection bundles (3³). Any bundles with an

evaluation consistency smaller than 3 are considered

inconsistent and excluded from further analysis.

Based on this, ten realistic projection bundles remain.

The six projection bundles with the highest

consistency score contain predominantly positive or

neutral future projections. This suggests that the

extended expert group is generally optimistic about

the future development of the key factors.

4.1.8 Future Space Mapping

To infer future spaces and effectively communicate

the acquired results they are visualised in this step. A

scenario map, which is common in scenario planning,

is not suitable in this case as the projection bundles

cannot be arranged in a meaningful way. Instead,

13579

Connectivity (Rank)

Relevance (Rank)

Potential Key

Factors with

high Relevance

Clear Key

Factors

Not Relevant

Factors

Legend

Increasing Resilience in Production Networks: A Practical Approach Based on Scenario Planning and Simulation-Based Capacity Analysis

165

each key factor is indicated by a symbol coloured

according to the projected development, see Figure 5.

The future space is grouped according to its

influence on the capacity of the production network,

consistent with the previously defined project goals.

If all projections are positive (bundle 1), the capacity

flexibility is high. For bundles with one neutral

projection (2A, 2B, and 2C) or two neutral projec-

tions (3A, 3B, and 3C), the capacity flexibility is

medium to high. If all projections are neutral

(bundle 4) or negative (bundle 5), the capacity

flexibility is low to medium.

This differentiation allows for precise adjustment

of the specific capacity needs, based on which

different strategies for increasing resilience can be

assigned. In a future space with limited to moderate

capacity flexibility in the production network,

resilience can only be improved by effectively

overcoming the challenges within the network.

Increasing the resilience in future spaces with high

capacity flexibility is not essential as the production

network already encompasses an inherently high

flexibility to react to unexpected situations. Therefore,

the further focus is on increasing resilience for the

future space with medium to high capacity flexibility.

These projection bundles are the foundation for the

subsequent steps that demonstrate ways to enhance

resilience in production networks supported by

simulation.

Figure 5: Future Space Mapping for the developed

projection bundles.

4.2 Integration of Simulation-Based

Capacity Analysis

This section outlines the systematic implementation

of the approach for integrating simulations into the

scenario planning, as described in section 3.2. This

includes conducting a simulation-based capacity-

analysis and the derivation of strategies for increasing

resilience.

4.2.1 Modelling of Production Network

The production network was modelled in Plant

Simulation using discrete-event simulation network

due to existing licensing and the ability to build on

previous modules and experiences.

The simulation is structured in three levels. On

the process level, variables and objects which depict

critical production processes such as bottlenecks or

unstable processes are managed. This covers specific

aspects which focus on single steps and activities of

the production process. The site level is broader and

introduces variables and objects which impact the

lead times of the entire process flow, shift systems or

supply cycles of each specific site. At the network

level the sites and suppliers are connected to each

other. Variables of higher levels are influenced by

lower levels, e.g. the lead time for the specific site

depends on the critical processes within the site. This

can also lead to delayed deliveries at the network

level which underlines the complex relationships

between levels and sites. Figure 6 provides an

overview of the simulation model at the site level.

Figure 6: Overview of simulation model at network level.

The model contains three functional base

modules. The supplier module contains external and

internal procurement which are simulated as simple

module for uncritical suppliers or as complex net-

works. It acts as supply component for the production

modules and provides them with components such as

raw materials, semi-finished goods and other materi-

als. The production modules model the production

sites with networks including the individual

production processes. The functional modules enable

High

capacity

flexibility

Medium to high

capacity

flexibility

Low to medium

capacity

flexibility

Quality and

performance of

suppliers

Availability and

efficiency of operating

resources

Availability of

materials and raw

materials

Legend:

Positive development

Negative development

Neutral development

2B2A 2C

3B3A 3C

SIMULTECH 2024 - 14th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

166

the configuration of scenario parameters with a dialog

window and visualizes simulation results and reports.

The order management facilitates the

redistribution of orders across the production network

by providing them to the relevant production sites

which then initiate the production processes. The

production networks order the required components

from the supplier modules and networks where

varying delivery concepts such as supermarket or

just-in-sequence are implemented.

A set of fixed-value tests demonstrates that the

simulation accurately mirrors the production

network, with a maximum deviation of only 13% for

the combined network, also see Figure 7. Averaged

over an entire fiscal year (FY) the simulation deviates

by 0.42 %.

In addition to the fixed value tests, the simulated

production network is verified through a review in

which the management, client, and supplier of the

simulation model checked the model’s inherent

consistency and specification fulfilment. Further

V&V techniques are difficult to implement in this

context as real-world data on the impact of

disruptions is scarce (Tan, 2020).

Figure 7: Relative deviation between actual and simulated

delivery of products over a fiscal year with blue trend line.

4.2.2 Modelling of Scenarios and

Countermeasures

Once the base model of the production network is

verified and validated the scenarios and

countermeasures are implemented. As described in

section 3.2.2, the implementation is based on a

module for each influencing factor. These modules

are triggered based on the selected scenario and initia-

lize scenario-specific changes.

Out of the implemented scenarios, scenario 3C is

presented in detail in the following. Table 3 presents

an overview of how the development of the key

factors in this scenario is realised in the simulation

model. To model the neutral development of the key

factor quality and performance of suppliers the

delivery of products for one crucial supplier was

changed to a normal distribution over a set timeframe

to simulate unreliable delivery times. During this, the

components are no longer delivered on a daily basis,

but follow a normal distribution. The results of this

simulation are shown in Figure 8.

The normal distribution of deliveries from a key

supplier in the period from the 12

month in FY1 to

the 10

month in FY2 has a significant impact on the

overall throughput of the production network. The

volatile fluctuations are a plausible result of a

normally distributed delivery from a main supplier.

Once the delivery problems have been resolved, the

throughput returns to the previous level with a slight

delay. Minor differences between the results of the

basic simulation and the scenario simulation are due

to the dynamics of the simulation and negligible. The

positive development of the key factor 'availability

and efficiency of operating resources' has no impact

due to the neutral development of the other key

factors.

Table 3: Realisation of scenario 3C in simulation model.

Key Factor

Develo-

pment

Simulation Model

Availability and

efficiency of

operating

resources

Positive

Increased availability of

blocks in production sites

Availability of

materials and raw

resources

Neutral

Materials and

components are produced

steadil

Quality and

performance of

suppliers

Neutral

One crucial supplier de-

livers the products with a

normal distribution for

set timeframe.

Figure 8: Results of a simulation run for scenario 3C.

In response to the events of scenario 3C described

above, a reactive countermeasure is implemented in

the form of a second supplier. This additional supplier

is now responsible for delivering about ten percent of

the critical items. The attached chart in Figure 8

Scenario duration

Januar 25

April 25

Juli 25

Oktober 25

Januar 26

April 26

Juli 26

Oktober 26

Januar 27

April 27

Juli 27

Delivery Rate

Y1Q1

Y1Q4

Y1Q2

Y1Q3

Y2Q1

Y2Q2

Y2Q3

Y2Q4

Y3Q1

Y3Q2

Y3Q3

Base simulation Scenario

Scenario +

Countermeasure

Increasing Resilience in Production Networks: A Practical Approach Based on Scenario Planning and Simulation-Based Capacity Analysis

167

shows that this measure was able to mitigate the

negative impact, but not fully compensate for it.

Nonetheless, the performance of the production

network is able to reach stable condition sooner and

is therefore more resilient. However, there are other

associated costs, such as increased expenses and

existing joint strategic partnerships with the supplier.

The key factors alignment can have varying

impacts which are considered in the simulation, e. g.

instead of changing the rate of deliveries, the quality

of products can be varied, changing the rate of

rework, to characterize the development of the

suppliers. This step also involves visualising the

results and creating an interface to allow a simple

manipulation of variables.

4.2.3 Identification of Potentials for

Increasing Resilience

In the final step, the potentials for increasing

resilience are evaluated based on the created

simulation model by inserting countermeasures for

the specific scenarios. This allows observation of the

impact of potential measures and the time required to

reach a stable condition again.

Members of the project team and other key

experts are enabled to make more informed decisions

by evaluating the impact of the previously identified

scenarios on the production network and trial

potential countermeasures, such as increasing safety

stock, and observing the impacts, and combinations

thereof. This is an iterative process where new

countermeasures are devised and implemented into

the simulation model.

4.3 Lessons Learned

A number of lessons were learnt during the imple-

mentation and evaluation of the case study. It is

recommended to involve the extended experts often

and at an early stage in order to avoid scenarios that

are too specific or too general. This also helps to

improve access and quality of data, both of which are

crucial for the quality of the results. The consulted

experts should come from as many different

backgrounds and departments as is reasonable.

The process enabled all parties to gain a more

comprehensive understanding of the potential risks to

the production network under review. This

understanding was deepened by the systematic testing

of countermeasures in the developed model.

Users gained an understanding of the system's

inertia and reaction speed, and how the production

network reacts to different scenarios and possible

countermeasures. A flexible interface for adapting the

scenarios and countermeasures is key for enabling

this evaluation.

In accordance with the project's defined goals, the

costs associated with the introduction of certain

countermeasures and optimisation algorithms have

not been included in the simulation. However, their

addition is being considered for future projects.

At this point, it must be emphasized that a

simulation model is not able to accurately calculate

the future. Rather, it serves to point in the right

direction and to understand the production network

and its reactions to changes. The implemented

scenarios and countermeasures must always be

critically examined and V&V performed before

measures are put into action.

5 CONCLUSION AND OUTLOOK

The complexity of today's GPNs has made them

vulnerable to production disruptions. Therefore,

recent focus has shifted from reacting to disruptions

to proactive prevention. Existing literature does not

adequately address the needs of organisations aiming

to increase their resilience.

The proposed approach focuses on increasing the

resilience in a three-phase process.

In the first phase the project’s goals are specified.

After which the second phase involves the identifyca-

tion of the most pertinent scenarios, which are

subsequently subjected to simulation. This approach

allows for a more focused and efficient allocation of

modelling and evaluation resources, directing them

towards the most critical aspects of the production

network.

In the conducted case study, the involved

individuals gained a deeper comprehension of the

interactions within the considered production

network and the potentials for increasing resilience.

They were also able to experiment with the

consequences of various preventive and corrective

measures, thereby enhancing their decision-making

abilities.

Future research should concentrate on conducting

additional case studies with the objective of further

validating the proposed process and developing

guidelines for estimating the associated costs for an

implemented countermeasure with reasonable efforts.

ACKNOWLEDGEMENTS

We would like to thank Siemens Healthineers Ag for

SIMULTECH 2024 - 14th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

168

their continued collaboration and support throughout

the project. The research was supported by VDI/VDE

grant “BayVFP Förderlinie Digitalisierung” with the

funding reference number DIK0279/01.

REFERENCES

Alexopoulos, K., Weber, M., Trautner, T., Manns, M.,

Nikolakis, N., Weigold, M., & Engel, B. (2023). An

industrial Data-spaces Framework for Resilient

manufacturing value chains. Procedia CIRP, 116, 299–

304. https://doi.org/10.1016/j.procir.2023.02.051

Balakrishnan, S., & Zhang, Z. (2020). Criticality and

Susceptibility Indexes for Resilience-Based Ranking

and Prioritization of Components in Interdependent

Infrastructure Networks. Journal of Management in

Engineering, 36(4), Article 04020022. https://doi.org

/10.1061/(ASCE)ME.1943-5479.0000769

Carvalho, H., Barroso, A. P., Machado, V. H., Azevedo, S.,

& Cruz-Machado, V. (2012). Supply chain redesign for

resilience using simulation. Computers & Industrial

Engineering, 62(1), 329–341. https://doi.org/10.

1016/j.cie.2011.10.003

Cordova-Pozo, K., & Rouwette, E. A. (2023). Types of

scenario planning and their effectiveness: A review of

reviews. Futures, 149, 103153. https://doi.org/

10.1016/j.futures.2023.103153

Debnath, R., Pettit, C., van Delden, H., & Perez, P. (2024).

Collaborative modelling for goal-oriented scenario

planning: A resilience planning case study in the

context of greater Sydney. International Journal of

Disaster Risk Reduction, 100, 104205.

https://doi.org/10.1016/j.ijdrr.2023.104205

Erlach, K., Berchtold, M.‑A., Kaucher, C., & Ungern-

Sternberg, R. (2023). Gestaltung resilienter

Produktionsnetzwerke mit Agilitätsbefähigern.

Zeitschrift Für Wirtschaftlichen Fabrikbetrieb, 118(4),

217–221. https://doi.org/10.1515/zwf-2023-1047

Feldkamp, N., Bergmann, S., & Strassburger, S. (2020).

Knowledge Discovery in Simulation Data. ACM

Transactions on Modeling and Computer Simulation,

30(4), 1–25. https://doi.org/10.1145/3391299

Fink, A., & Siebe, A. (2016). Szenario-Management: Von

strategischem Vorausdenken zu zukunftsrobusten

Entscheidungen. Campus Verlag.

Gausemeier, J., Fink, A., & Schlake, O. (1996). Szenario-

Management: Planen und Führen mit Szenarien (2.,

bearb. Aufl.). Hanser.

Henderson, J., Dicken, P., Hess, M., Coe, N., &

Yeung, H. W.‑C. (2002). Global Production Networks

and the Analysis of Economic Development. Review of

International Political Economy, 9(3), 436–464.

http://www.jstor.org/stable/4177430

Hoffmann, G. P. (2017). Organisationale Resilienz:

Kernressource moderner Organisationen. Springer

Berlin Heidelberg; Imprint: Springer. https://permalink.

obvsg.at/AC12324391 https://doi.org/10.1007/978-3-

662-53944-6

Holling, C. S. (1973). Resilience and Stability of

Ecological Systems. Annual Review of Ecology and

Systematics, 4(1), 1–23. https://doi.org/10.1146/

annurev.es.04.110173.000245

Ivanov, D. (2018). Revealing interfaces of supply chain

resilience and sustainability: a simulation study.

International Journal of Production Research, 56(10),

3507–3523. https://doi.org/10.1080/00207543.2017.13

43507

Kunz, D., Weber, J., Barth, M., & Franke, J. (2023). Virtual

Commissioning and the use of Extended Reality and

Automated Testing: A survey of industry. Advance

online publication. https://doi.org/10.22032/dbt.57812

(20. ASIM Fachtagung Simulation in Produktion und

Logistik 2023, p. 241).

McManus, S. (2007). Organisational Resilience in New

Zealand [PhD thesis]. Department of Civil Engineering,

University of Canterbury, New Zealand.

Paich, M., & Hinton, R. (1998). Simulation Models: A Tool

for Rigorous Scenario Analysis. In L. Fahey (Ed.),

Learning from the future: Competitive foresight

scenarios. Wiley.

Peukert, S., Hörger, M., & Zehner, M. (2023). Linking

tactical planning and operational control to improve

disruption management in global production networks

in the aircraft manufacturing industry. CIRP Journal of

Manufacturing Science and Technology, 46, 36–47.

https://doi.org/10.1016/j.cirpj.2023.07.009

Rydzak, F., Magnuszewski, P., Sendzimir, J., & Chlebus, E.

(2006). A Concept of Resilience in Production Systems.

In Proceedings of the 24

Int. Conference of the System

Dynamics Society, Nijmegen, Netherlands: Radboud

Universiteit Nijmegen.

Schollemann, A., Wiesch, M., Brecher, C., & Schuh, G.

(2022). Resilience Drivers in Next Generation

Manufacturing. In F. T. Piller, V. Nitsch, D. Lüttgens,

A. Mertens, S. Pütz, & M. van Dyck (Eds.),

Contributions to Management Science. Forecasting

Next Generation Manufacturing (pp. 119–128).

Springer International Publishing. https://doi.org/10.

1007/978-3-031-07734-0_8

Seville, E. (2008). Resilience: great concept but what does

it mean? presented at the Risk Intelligence and

Resilience Workshop, Wilmington, USA.

Shannon, R. E. (1998). Introduction to the art and science

of simulation. In 1998 Winter Simulation Conference.

Proceedings (Cat. No.98CH36274) (pp. 7–14). IEEE.

https://doi.org/10.1109/WSC.1998.744892

Tan, W. J. (2020). Modelling and simulation of supply

chain resilience using complex system approaches.

https://doi.org/10.32657/10356/137144

Váncza, J. (2016). Production Networks. In T. I. A. f. Produ,

L. Laperrière, & G. Reinhart (Eds.), CIRP

Encyclopedia of Production Engineering (pp. 1–8).

Springer Berlin Heidelberg. https://doi.org/10.1007

/978-3-642-35950-7_16829-1

Zhang, W. J., & van Luttervelt, C. A. (2011). Toward a

resilient manufacturing system. CIRP Annals, 60(1),

469–472. https://doi.org/10.1016/j.cirp.2011.03.041

Increasing Resilience in Production Networks: A Practical Approach Based on Scenario Planning and Simulation-Based Capacity Analysis

169