Digitalization in Small-Load-Carrier Management

Alexander Dobhan

, Lars Eberhardt

, Markus Haseneder

, Heiko Raab

Steffen Rabenstein

, Axel Treutlein

, Vincent Wahyudi

and Martin Storath

Technical University of Applied Sciences Würzburg-Schweinfurt, Ignaz-Schön-Straße 11, 97421 Schweinfurt, Germany

Lobster DATA GmbH, Bräuhausstraße 1, 82327 Tutzing, Germany

sprintBOX GmbH, Gerolzhöfer Straße 7, 97508 Grettstadt, Germany

TAF Industriesysteme GmbH, Ostring 28, 97228 Rottendorf, Germany

Keywords: Small Load Carrier, Returnable Transportation Equipment/Item, Return-to-Deliver, Computer Vision,

Localization, Citizen Development, Internet-of-Things.

Abstract: In this article, we describe our research on digitalization in the field of returnable small load carrier (SLC)

management. Our findings are the result of a collaboration between three companies and an academic insti-

tution. We apply various methods for modeling and analyzing digitalization measures that are already being

prototypically implemented and discuss them in terms of transparency, data quality, resource consumption

and costs. Our research enables academic researchers to build on real-world data and problems. For practi-

tioners, we offer concrete solutions to increase the level of digitalization in their organizations. Unlike most

other academic work to date, we focus on SLCs with their specific characteristics. This article could be the

starting point for a higher impact and a growing number of research activities on returnable SLCs to make

SLC cycles more efficient, which in turn will increase the sustainability of industrial packaging in general.

1 INTRODUCTION

The EU Packaging and Packaging Waste Regulation

facilitates returnable packaging for consumers. Re-

turnable packaging is also a CO2-reduced alternative

to single-use packaging in an industrial context (Coe-

lho et al., 2020). The term container as one type of

packaging refers to any container, from large inter-

modal containers to small boxes, while returnable

transportation equipment (or item) (RTI) usually does

not include intermodal containers, but pallets and

small boxes (Elbert & Lehner, 2020). SLCs belong to

both containers and RTIs. SLCs are stackable plastic

boxes that are smaller than pallets and are used both

https://orcid.org/0009-0006-6898-4394

https://orcid.org/0000-0003-4919-2608

https://orcid.org/0009-0002-2128-7963

https://orcid.org/0009-0004-8382-0326

https://orcid.org/0009-0008-3903-4159

https://orcid.org/0009-0001-6499-1056

https://orcid.org/0009-0008-5650-2161

https://orcid.org/0000-0003-1427-0776

in production and for transportation (Ziegler et al.,

2023).

An SLC can contain inlays or covers, which leads

to a wide variety of possible SLC-sets. In addition,

SLCs cannot usually be labeled with a reference to

themselves, as all the space is reserved for labels on

the SLC contents. Due to the large number and high

density of SLCs in a warehouse, complete tracking

with GPS, for example, is not possible. The large va-

riety of SLCs in combination with the low value of a

single SLC and the labeling challenges lead to poor

availability of data on SLC cycles. It follows that dig-

italization helps to improve decision quality in SLC

cycles. Since all of the previously listed attributes re-

late to it, we focus on the decisions in the SLC return-

848

Dobhan, A., Eberhardt, L., Haseneder, M., Raab, H., Rabenstein, S., Treutlein, A., Wahyudi, V. and Storath, M.

Digitalization in Small-Load-Carrier Management.

DOI: 10.5220/0013363200003929

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

In Proceedings of the 27th International Conference on Enterprise Information Systems (ICEIS 2025) - Volume 1, pages 848-855

ISBN: 978-989-758-749-8; ISSN: 2184-4992

to-deliver process, which begins with the return of

SLCs and ends with the delivery of refurbished SLCs

to the cycle partners. Decision tasks have some de-

grees of freedom to align the output with the respec-

tive goals of the company (Ferstl & Sinz, 2012). In

SLC management, decision-makers have to make a

trade-off between high availability of SLCs and low

inventories. We focus on the digitalization of decision

tasks. As digitalization improves transparency and

data quality, a lack of digitalization leads to less trans-

parency, e.g. in detecting machine issues, material de-

fects, shortages, or high inventories. These issues re-

sult in less use of returnable SLCs, which in turn leads

to less sustainability in packaging in general.

With the digitalization of SLC cycles, however, it

is possible to identify problems within the cycle more

quickly and determine stocks more effectively. Fur-

thermore, it helps to standardize processes and facili-

tates process visualization and transparency. Overall,

this can open up new opportunities for the use of re-

turnable SLCs, which means greater sustainability in

packaging. Therefore, we provide an answer to the

following research question:

How can decisions in the return-to-deliver process

of returnable SLCs be digitized?

Our research is anchored in the DIBCO project

funded by the State of Bavaria and which is being car-

ried out by four partners. Lobster DATA GmbH, as a

provider of logistics cloud software, sprintBOX

GmbH as a logistics service provider and SLC man-

agement specialist, TAF Industriesysteme GmbH as

a logistics system provider, and THWS as an aca-

demic research partner. According to our research,

this is the first academic paper that examines digital-

ization in the entire SLC return-to-deliver process in

detail and process-oriented (see section 2). Our find-

ings could help companies digitize their SLC cycles

to make them more efficient and increase their return-

able quota. Future scientific research can build on the

solutions we introduce below, improve them or offer

additional digital solutions for activities in the pro-

cess.

The organization of this article is as follows: After

the introduction, we provide an overview of recent re-

search in the field of digitalization in SLC manage-

ment. Then, we first describe our methodology and

then apply it to our use case. We then discuss the re-

sults from different perspectives and provide a sum-

mary and an outlook for future research.

2 STATE OF THE ART

To get an overview of the current research on digital-

ization in SLC management, we conducted a litera-

ture search with the following search term (digit* OR

automat*) AND ("Small Load Carrier" OR "Returna-

ble Transport*") in Scopus, Springerlink, and IEEE

transactions. To focus on recent results, we only con-

sidered articles and conference proceedings pub-

lished between 2019 and 2024. Scopus provided 10

publications, IEEE 2 publications and Springerlink

19 publications. The low number indicates that not

much research has been published on this topic. This

is due to the specific focus on SLCs or RTIs, which

leads to an exclusion of intermodal containers, to

which most research in this area refers. The relevant

literature can be categorized into 3 groups: Planning

which includes management and coordination activi-

ties, object recognition and sensors for monitoring the

process, and execution activities, in particular SLC

handling.

One planning task is the allocation of SLCs in the

SLC cycle. Elbert & Lehner (2020) solve this prob-

lem for pallets using an agent-based exchange plat-

form. Schneikart et al. (2024) discuss and partially

prove the viability of using returnable SLCs for three

use cases in the pharmaceutical industry. Cycle coor-

dination requires data on cycle inventory and lead

times of an SLC. In practice and in science, however,

there is a lack of corresponding data, which is mainly

due to a lack of data collection or an unwillingness to

share data with the cycle partners (Müller et al.,

2025).

One way to overcome these problems is to moni-

tor the SLCs using sensors. Bemthuis et al. (2023) use

pallet-specific data from temperature, vibration, and

GPS trackers to predict the state of individual SLCs

using decision tree models. Kreutz et al. (2021) focus

on the fill level of the individual SLCs. Gan (2019)

discusses an approach to find the best position for an

RFID tag on an RTI. While all of the findings in the

sensors category relate more to the SLC sensors

themselves, the articles in the object recognition cat-

egory contain findings on how to identify unlabeled

SLCs. Rutinowski et al. (2024) set out to create a da-

taset from real-world data for the recognition of lo-

gistics objects on a store floor. They conducted ex-

periments to create a large dataset for different logis-

tics objects including SLCs. Abou Akar et al. (2024a),

Abou Akar et al. (2024b), and Mayershofer et al.

(2021) aim to create synthetic datasets for SLCs. Be-

loshapko et al. (2020) used a Mask R-Convolutional

Neural Network to identify the bins.

Digitalization in Small-Load-Carrier Management

849

Another perspective of digitalization relates more

to the activities during process execution performed

to the SLC, such as handling, cleaning, or transport

(e.g. Blank et al., 2023). Overall, the literature review

shows the following:

- The digitalization of SLCs mainly relates to the

areas of object recognition, sensors, and handling.

- There is a lack of real data on object recognition,

but also on SLC cycle data in general.

- The investigation of sensors refers to the pallets

that store the SLCs and to the contents of the SLCs

instead of the SLCs themselves.

- In terms of activities, the reprocessing activities

such as cleaning are not digitally monitored or at least

the research does not address this monitoring.

- There is little research on the coordination (man-

agement) of SLC cycles, although SLCs have specific

characteristics described in the introduction.

- No article explicitly addresses the decisions in

SLC cycles, despite their importance for sustainabil-

ity in packaging.

- The detection of SLC defects is not part of the

digitalization approaches published so far.

Our research aims to improve decisions in SLC

cycles based on individual SLC data and the collec-

tion of real-world data in the three areas of planning,

execution, and monitoring.

3 METHODOLOGY

To answer the research question, we apply a use case

analysis as described in Eisenhardt (1989). The selec-

tion of companies was completed before the start of

the project. To collect data, we used expert inter-

views, participative observation, analysis of docu-

ments from our project partners (e.g. defect catalogs),

and project documents (e.g. requirements docu-

ments). From these, we extracted a BPMN model

(OMG, 2013) that describes the return-to-deliver pro-

cess from the arrival of soiled SLCs at the SLC depot

to the delivery of clean SLCs. Based on the BPMN,

we identified three relevant fields of decisions that

had a low level of digitalization before the project and

that we expected to significantly improve when digit-

ized. To map these decisions, we applied a modifica-

tion of the model by Dobhan & Zitzmann (2022),

which is based on Sieben & Schildbach (1975). Ac-

cording to them, a decision consists of 5 elements

(figures 2, 3, 4): Information gathering (1) refers to

the collection of all necessary data and information.

The objectives (2) include all relevant objectives for

the decision. The decision field (3) contains possible

decision options. The evaluated alternatives are listed

in the results matrix (4), while the selected alternative

is highlighted in the decision matrix (5). These activ-

ities are either implicitly conducted as thoughts or ex-

plicitly as a discussion on a sheet of paper, or within

a software. In order to examine the degree of digital-

ization and automation of a decision, it is necessary

to assign both degrees to each of these activities. In

information gathering, digitalization means that the

information is made available in digital form, while

automation means that all the required information is

automatically available to the decision makers in the

right format at a glance. Manual information collec-

tion in ERP systems means manual effort and reduces

the degree of automation to non-automated or only

partially automated. The same applies to the consid-

eration of objectives and the creation of the decision

field, the decision matrix, and the results matrix.

Figure 1: Return-to-Deliver Process for SLCs.

ICEIS 2025 - 27th International Conference on Enterprise Information Systems

850

According to Ferstl & Sinz (1998), automation is

represented by a square. In our application a filled

square means that the task is fully automated, a white

rectangle means that the task is not automated. A par-

tially filled square means that the task is partially au-

tomated. The same applies to digitalization, but with

a circle. The symbols on the left-hand side of the ac-

tivity box represent the status before the project,

while the activity on the right-hand side represents the

(prototypical) achieved status with the proposed solu-

tion. In addition to this semi-formal description, we

describe the measures taken in the use case in the text.

4 USE CASE ANALYSIS

Our use case analysis refers to the SLC management

for an SLC cycle with more than two partners, which

is mainly coordinated by sprintBOX GmbH. The so-

lutions described are implemented as prototypes in

the company. As technological partners, TAF Indus-

triesysteme GmbH and the logistics cloud software

provider Lobster Data GmbH supported the imple-

mentation of the required technological solutions.

The return-to-deliver process is mapped as BPMN in

figure 1.

We developed technological solutions for deci-

sions that are currently carried out manually and

partly digitally and for which suitable solutions are

not readily available. We identified key planning de-

cisions, a strongly experience-based decision during

process execution, and decisions during the process

monitoring. The technical solutions described below

are prototyped.

Planning Decisions. The main planning task for the

return-to-deliver process described above relates to

the decision on the order sequence for the three main,

partially decoupled activities of the process: cleaning,

assembly, picking. In our project, we focus on clean-

ing order sequencing. Currently, the planners perform

the sequencing tasks with paper and office software

based on data from the SLC management software.

The cleaning order is manually transferred to the store

floor. During the project, it turned out that no standard

software adequately met the requirements. Another

problem is the dynamics behind the SLC business.

After signing a contract with a customer, there is only

a short period of time (from 6 months to a year) to

implement processes in an often customer-specific

depot. This requires highly flexible systems or at least

a certain amount of development work. Therefore,

and because a MS Power App environment was al-

ready in place, we decided to use Citizen Develop-

ment. Citizen development means that non-IT person-

nel are enabled to take on development tasks (Binzer

& Winkler, 2022). We implemented a prototype for

sequencing cleaning orders based on a priority-based

algorithm. To transfer the results digitally to the store

floor, we have developed specific store floor views

that show the results of the sequencing and allow the

order to be started and stopped.

Figure 2: Digitalization and automation of sequencing.

With this solution, we are significantly increasing

the degree of digitalization and automation of plan-

ning decisions (figure 2). The sequencing mainly uses

demand, capacity, and inventory data as input. With

our solution, most of the data is automatically col-

lected and displayed in a software module for clean-

ing order sequencing. The software allows viewing

different order sequences and suggests the best one.

Decisions During Process Execution. The quality of

the SLCs is checked during the sorting activity and

after the cleaning process. Both checks were carried

out manually. We decided to develop a technological

solution that digitizes and automates defect detection,

built by TAF Industriesysteme GmbH and applied at

a sprintBOX depot. We chose to develop the solution

for clean SLCs because there are isolated components

after the cleaning machines, which simplifies the han-

dling of the SLCs for defect detection. Computer Vi-

sion (CV) as a method was at the center of the solu-

tion (Wahyudi et al., 2025; Ziegler et al., 2023). De-

fect detection (together with the SLC detection itself)

should take place within 3 seconds, which is the cycle

time of a cleaning machine. Another major challenge

was to distinguish defective from non-defective

SLCs. For example, it is difficult to distinguish a wet

SLC from an oily one. To capture the SLC images,

we implemented a portal with 5 RGB cameras. The

cameras were mounted on aluminum rods and parti-

tion walls with lights ensure that the lighting condi-

tions do not change. To find the most suitable model,

we compared several state-of-the-art anomaly detec-

tion models for a selection of the 34 most used SLCs.

A total set of 17,430 images was used for the experi-

ments. After tuning the hyperparameters, the Patch-

Core (Roth et al., 2022) model proved to be the best

Digitalization in Small-Load-Carrier Management

851

in terms of the Area under the Receiver Operating

Characteristic Curve (AUROC) with a value of 0.811.

Our solution meets the requirements of detecting ob-

jects and defects within 3 seconds in a laboratory en-

vironment. In the SLC depot, the conveyor roller was

modified to bring the SLCs into a position suitable for

the cameras. The SLC images are taken automati-

cally, the objectives are taken into account (not com-

pletely, but to a good extent), and the 3 steps to the

decision are carried out simultaneously, digitally and

automatically (figure 3).

Figure 3: Digitalization and automation of quality check.

Monitoring Decisions. Process monitoring takes

place in order to recognize target/actual deviations in

process execution and to make a decision on how to

react to these deviations. Without digitalization, mon-

itoring was mainly experience-based and happened

with the help of office documents supported by the

SLC management software for the transactions with

the cycle partners. The target values are specified in

the order plans, the digitalization of which we de-

scribed above.

Actual Values for Throughput Times. To get more

information about throughput times, we have devel-

oped an approach that involves tracking only a few

SLCs within a cycle. A complete data collection cov-

ering all SLCs in a cycle would be far too expensive,

as each SLC costs no more than 1 Euro. Following

Müller et al. (2025), we used the sample data as a ba-

sis to extend the data from there and perform a simu-

lation analysis to help us gain more insights into the

actual cycle throughput times. The technological re-

quirements for the tracking system mainly relate to

cost, precision, localization capabilities, and size. It

turned out that Apple Air Tags were the most suitable

technology. Using them, we collected the data for 10

runs in a test cycle (2 routes, 4 locations), enriched

the data with transportation times from navigation

apps, and used a PERT distribution to extend the data

and apply it to a Monte Carlo simulation (Müller et

al., 2025). For display and export of tracking data, we

developed an app for Lobster logistics.cloud.

Actual Values for the SLC Quantity. Knowing the

SLC quantity in the process helps to understand

throughput times, but also inventory levels. Since the

previously presented tracking approach mainly con-

siders the times between cycle locations and is only

designed for a small sample, it would be beneficial to

know the number of objects processed by the activi-

ties within the SLC depot. Before digitalization, the

number of SLCs was counted manually on a paper.

The data transfer from the paper to the system leads

to delays of several hours or even days. This

prompted us to combine object detection with defect

detection. In Wahyudi et. al. (2025), we propose a CV

based approach for classification. We use ConvNeXt

(Liu et al., 2022), which allows us to achieve an ac-

curacy of 100% for the same sample as for the defect

detection. We achieved that using the same portal as

for defect detection. To extract the detection data we

applied a prototypical IoT architecture via MQTT and

Web Services. Together with tracking data the object

detection data allows more insights into the current

process and inventory status.

Actual Values for the Machine Status. As there are

only a few cleaning machines in each depot, this is

also the most critical process of the entire depot. Be-

fore the project, the cleaning machine was only mon-

itored locally in the store floor. To improve this in or-

der to increase capacity and reduce downtimes, dash-

boards of the machine data were created and made

available to the central departments and site manag-

ers. To do this, we applied an IoT architecture that

connects the machine control software (in the proto-

typical case a Siemens Simatic S7) via VT Scada (a

Scada software). Even more important was the alert

management. As soon as a value was outside a certain

threshold, e.g. the temperature, an alert was dis-

played, which also enables documentation of the

cleaning machine's availability.

Figure 4: Digitalization and automation of monitoring.

In summary, we digitized the most relevant target

and actual values to improve the process monitoring

and enable decisions on how to deal with target/actual

deviations (figures 4). It changes the digitalization of

inputs. The trade-off between cost and SLC shortages

as well as developing, evaluation, and selection of the

reaction remains partially digital and not automated.

ICEIS 2025 - 27th International Conference on Enterprise Information Systems

852

5 DISCUSSION

Our research aims to develop solutions for the digi-

talization of the return-to-deliver process of returna-

ble SLCs to make the cycles more efficient and to

make companies use them instead of single-use pack-

aging. During our research, we developed prototypes

for each solution mainly addressing technological

feasibility for given constraints, such as the avoidance

of labelling each SLC or new tracking infrastructure

(other than the cameras). Therefore, our focus was not

on overcoming organizational barriers to new tech-

nologies such as machine learning (Schkarin & Dob-

han, 2022). Also, we cannot provide results on the

scalability of our solutions or the impact of large-

scale deployment. As with all case study analyses, in

addition to the findings from the literature review and

their embedding in academic research, it is mainly the

company-specific practical experience which guided

our development activities. Validity for other compa-

nies could not be proven in this article. There could

be other limitations in other companies. Furthermore,

the prototypes are still under development, which

leads to minor implementation issues, e.g. an increase

in inference time when the recognition SLC quantity

advances. Nevertheless, our scientific contributions

are clearly the following:

- In contrast to most previous work, we strictly fo-

cus on SLCs with their specific properties (see Elbert

& Lehner, 2020; Ziegler et al., 2023).

- We collected SLC-related data from a real envi-

ronment for localization technologies and CV. A lack

of both has been identified in previous work (Abou

Akar et al., 2024a; Müller et al., 2025)

- We discuss the application of citizen develop-

ment in SLC planning as a new suitable use case for

citizen development (Elshan et al., 2023).

Table 1: Impact of digitalization measures.

Planning Execution Monitoring

Transparency

Recognition of status (current plan is

available for all stakeholders) and

problems (delays or capacity shortage

is visible for all).

Facilitation of communication (re-

sults can be easily communicated to

the store floor and the users which

have access to order list).

Enabling decision-making (infor-

mation is displayed in an usable way,

the decision on order sequence is

simplified).

Recognition of status and

problems (defects are detected

automatically and recorded in

the system).

Facilitation of communication

(digital recording and commu-

nication of the defects ena-

bled).

Enabling decision-making

(the decision-making on de-

fects is completely handed

over to artificial intelligence)

Recognition of status and problems

(target/actual deviations are recog-

nized easier).

Facilitation of system performance

(digitalization of values enables to

check the system performance (actual

to target)).

Enabling decision-making (more in-

formation usually can improve deci-

sion-quality on reactions).

Data Quality

Accessibility of plans for all relevant

stakeholder every time in parallel.

Software ensures completeness of

data.

Concise representation through de-

velopment by users.

Consistency of data because of the

single source.

Timeliness as the plan is available

immediately after release.

Alignment with defect cata-

logues via training data

Consistency because decision

is made by software.

Objectivity because decision

is made by software.

Traceability at least regarding

responsibility and timestamp.

Unambigous data (ok, nok).

Accessibility of monitoring data for

all relevant stakeholder in parallel

Accuracy because of detailed actual

values

Believability, Objectivity as data

comes directly from sensors.

Timeliness as sensor data is availa-

ble immediately.

Traceability because data sources

are clear by design.

Resources

Avoidance of paper for communica-

tion, Reduction of inventory or

shortages. Less emergency trans-

ports or orders.

Avoidance of additional

transportation and inventory

of defective SLCs, Only defec-

tive SLCs are excluded.

Early detection of problems => less

or shorter machines stops => less

emergency transports or orders.

Avoidance of unnecessary resource

consumption.

Costs

Cost for Power App license, Non-IT-

resource cost, hardware cost

(screens etc.).

Camera portal: cameras plus

lighting (< 5k), Material and

personnel costs for building

the portal and changing mate-

rial handling.

Apple AirTags (25 Euros each plus

Apple Laptops), VT Scada license

(~10k euros) plus preparation and

configuration of existing machine,

Camera portal.

Digitalization in Small-Load-Carrier Management

853

- Furthermore, we slightly modified and applied

the approach of Dobhan & Zitzmann (2022), which

can be easily transferred to other decisions to examine

the degree of digitalization and automation.

The digitalization of business processes aims to

increase transparency, reduce resource consumption,

and improve data quality. On the other hand, digitali-

zation efforts incur costs for the implementation and

operation of digital solutions. We therefore shed light

on the impact of our solutions on transparency and

data quality as well as on resource consumption and

costs (table 1).

Transparency. According to Klotz et al. (2008),

transparency means that stakeholders understand the

necessary aspects and status of operations at all times.

In their study, they provide an overview of various

attributes of transparency. Specifically, each of our

digitalization efforts has the effects listed in table 1.

Data Quality. According to the extensive literature

review by Wang et al. (2024), data quality has various

dimensions, such as accessibility and timeliness. An

overview of the effects on these dimensions is given

in table 1.

Resource Consumption. The previous digitalization

effects also have an impact on resource consumption.

However, as we have only implemented our solutions

as prototypes so far, we do not yet have any detailed

figures on the impact on resource consumption in

day-to-day business. Nevertheless, we describe the

expected effects in general in table 1.

Costs. The implemented solutions are only proto-

types of a funded research project, which makes it dif-

ficult to estimate investment costs. Our research

shows that the current and future benefits of SLC dig-

italization should more than compensate for the costs.

6 OUTLOOK

Our research investigated the digitalization of the

SLC return-to-deliver process. The technological

application could be a blueprint for the digitalization

of the most important activities in the SLC return-to-

deliver process. In a next step, the developed

solutions should be distributed to more machines,

cycles, and locations in order to validate them for

mass use. The solutions introduced can make

returnable SLC cycles more efficient. This could lead

to a higher use of returnable SLCs, which in turn

increases sustainability in industrial packaging. From

a scientific perspective, our research contributes real

data and use cases in the context of SLC management,

which according to our research has rarely been

addressed before. It is the basis for further research

on the following topics.

- Future research should strive for fully automated

planning with automated event handling.

- Regarding sensor-based SLC time data, an

approach needs to be developed that combines

tracking data with data from the SLC management

software to determine an SLC target inventory. To

this end, it is interesting to investigate how SLC

management can benefit from process mining.

- It should also be analyzed how beneficial the

tracking of each individual SLC is. An SLC history

could help to understand the behavior in SLC cycles.

- In terms of CV, additional research is needed on

how to improve the results of defect detection

considering additional data and how to recognize the

degree of SLC soiling to derive information for a

decision on required cleaning activities for each SLC.

- Finally, from a strategic decision-making

perspective, the decision to digitize the SLC return-

to-deliver process should be analyzed by using

decision quality measures.

ACKNOWLEDGEMENTS

This research was funded by the research program

"IKT" of the Bavarian State (DIK-2105-0044 /

DIK0264), submitted by THWS. We would also like

to thank Sebastian Friedl, Fabian Freund, Christian

Marder, Rasha Mahodudi, Stefan Schramm, and Jan

Senner for their valuable prework on this study. For

grammar and vocabulary check, we applied

deepl.com, for proofreading in general instatext.io.

REFERENCES

Abou Akar, C., Abdel Massih, R., Yaghi, A., Khalil, J.,

Kamradt, M., & Makhoul, A. (2024a). Generative Ad-

versarial Network Applications in Industry 4.0 Interna-

tional Journal of Computer Vision, 132(6), 2195–2254.

Abou Akar, C., Tekli, J., Khalil, J., Yaghi, A., Haddad, Y.,

Makhoul, A., & Kamradt, M. (2024b). SORDI.ai:

Large-scale synthetic object recognition dataset gener-

ation for industries. Multimedia Tools & Applications.

Beloshapko, A., Knoll, C., Boughattas, B., & Korkhov, V.

(2020). KLT Bin Detection and Pose Estimation in an

Industrial Environment. In O. Gervasi, B. Murgante, S.

Misra, C. Garau, I. Blečić, D. Taniar, B. O. Apduhan,

A. M. A. C. Rocha, E. Tarantino, C. M. Torre, & Y.

Karaca (Hrsg.), Computational Science and Its Appli-

cations – ICCSA 2020 (Vol. 12254, p. 105–118).

Springer.

ICEIS 2025 - 27th International Conference on Enterprise Information Systems

854

Bemthuis, R., Wang, W., Iacob, M.-E., & Havinga, P.

(2023). Business rule extraction using decision tree ma-

chine learning techniques: A case study into smart re-

turnable transport items. Procedia Computer Science,

220, 446–455.

Binzer, B., & Winkler, T. J. (2022). Democratizing Soft-

ware Development: A Systematic Multivocal Literature

Review and Research Agenda on Citizen Development.

In N. Carroll, A. Nguyen-Duc, X. Wang, & V. Stray

(ed.), Software Business (Vol. 463, p. 244–259).

Springer.

Blank, A., Zikeli, L., Reitelshöfer, S., Karlidag, E., &

Franke, J. (2023). Augmented Virtuality Input Demon-

stration Refinement Improving Hybrid Manipulation

Learning for Bin Picking. In K.-Y. Kim, L. Monplaisir,

& J. Rickli (ed.), Flexible Automation and Intelligent

Manufacturing: The Human-Data-Technology Nexus

(p. 332–341). Springer.

Coelho, P. M., Corona, B., ten Klooster, R., & Worrell, E.

(2020). Sustainability of reusable packaging–Current

situation and trends. Resources, Conservation & Recy-

cling: X, 6, 100037.

Dobhan, A., & Zitzmann, I. (2022). Twin Transition: Sus-

tainable Digital Decision Making in Enterprise Re-

source Planning. Mobility in a Globalised World 2021,

137–160.

Eisenhardt, K. M. (1989). Building Theories from Case

Study Research. The Academy of Management Review,

14(4), 532.

Elbert, R., & Lehner, R. (2020). Simulation-based analysis

of a cross-actor pallet exchange platform. 2020 Winter

Simulation Conference (WSC), 1396–1407.

Elshan, E., Dickhaut, E., & Ebel, P. A. (2023). An investi-

gation of why low code platforms provide answers and

new challenges. https://scholarspace.manoa.ha-

waii.edu/items/227cc6b5-d9d1-4e76-9dec-

1afcd484576b

Ferstl, O. K., & Sinz, E. J. (1998). SOM Modeling of Busi-

ness Systems. In P. Bernus, K. Mertins, & G. Schmidt

(ed.), Handbook on Architectures of Information Sys-

tems (S. 339–358). Springer.

https://doi.org/10.1007/978-3-662-03526-9_15

Ferstl, O. K., & Sinz, E. J. (2012). Grundlagen der

Wirtschaftsinformatik. Oldenbourg Verlag.

Gan, O. P. (2019). Placement of Passive UHF RFID Tags

and Readers Using Graph Models. 2019 24th IEEE In-

ternational Conference on Emerging Technologies and

Factory Automation (ETFA), 640–645.

Klotz, L., Horman, M., Bi, H. H., & Bechtel, J. (2008). The

impact of process mapping on transparency. Interna-

tional Journal of Productivity and Performance Man-

agement, 57(8), 623–636.

Kreutz, M., Alla, A. A., Lütjen, M., & Freitag, M. (2021).

Autonomous, low-cost sensor module for fill level

measurement for a self-learning electronic Kanban sys-

tem. IFAC-PapersOnLine, 54(1), 623–628.

Liu, Z., Mao, H., Wu, C. Y., Feichtenhofer, C., Darrell, T.,

& Xi, S. (2022). A convnet for the 2020s. Proceedings

of the IEEE/CVF conference on computer vision and

pattern recognition, 11976–11986.

http://arxiv.org/abs/2201.03545

Mayershofer, C., Ge, T., & Fottner, J. (2021). Towards

Fully-Synthetic Training for Industrial Applications. In

S. Liu, G. Bohács, X. Shi, X. Shang, & A. Huang (ed.),

LISS 2020 (S. 765–782). Springer Singapore.

Müller, J., Eberhardt, L., Wahyudi, V., Storath, M., &

Dobhan, A. (2025). Towards an approach on location

data analysis for reusable small-load carriers. 58th Ha-

waii International Conference on System Sciences

(HICSS).

OMG. (2013). Business Process Model and Notation

(BPMN), Version 2.0.2,.

https://www.omg.org/spec/BPMN

Roth, K., Pemula, L., Zepeda, J., Schölkopf, B., Brox, T.,

& Gehler, P. (2022). Towards total recall in industrial

anomaly detection. Proceedings of the IEEE/CVF con-

ference on computer vision and pattern recognition,

14318–14328.

Rutinowski, J., Youssef, H., Franke, S., Priyanta, I. F., Po-

lachowski, F., Roidl, M., & Reining, C. (2024). Semi-

automated computer vision-based tracking of multiple

industrial entities: A framework and dataset creation

approach. EURASIP Journal on Image and Video Pro-

cessing, 2024(1), 8.

Schkarin, T., & Dobhan, A. (2022). Prerequisites for Ap-

plying Artificial Intelligence for Scheduling in Small-

and Medium-sized Enterprises. ICEIS (1), 529–536.

Schneikart, G., Mayrhofer, W., Löffler, C., & Frysak, J.

(2024). A roadmap towards circular economies in

pharma logistics based on returnable transport items en-

hanced with Industry 4.0 technologies. Resources,

Conservation and Recycling, 206, 107615.

Sieben, G., & Schildbach, T. (1975).

Betriebswirtschaftliche Entscheidungstheorie.

Wahyudi, V., Ziegler, C. C., Frieß, M., Schramm, S., Lang,

C., Eberhardt, L., Freund, F., Dobhan, A., & Storath,

M. (2025). A Computer Vision System for Recognition

and Defect Detection for Reusable Containers. Machine

Vision and Application, 36(2), 1-19.

Wang, J., Liu, Y., Li, P., Lin, Z., Sindakis, S., & Aggarwal,

S. (2024). Overview of Data Quality: Examining the

Dimensions, Antecedents, and Impacts of Data Quality.

Journal of the Knowledge Economy, 15(1), 1159–1178.

Ziegler, C. C., Ising, J., Dobhan, A., & Storath, M. (2023).

Computer Vision in Reusable Container Management:

Requirements, Conception, and Data Acquisition. Mo-

bility in a Globalised World 2022.

Digitalization in Small-Load-Carrier Management

855