Impact of Machine Learning on Supply Chain Management

This paper explores the impact of machine learning on supply chain management. It discusses machine learning types, reasons driving for machine learning, and its impacts on inventory control, transport networks, procurement, and customer relationship management.

Running Head: IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN
MANAGEMENT
Impact of Machine Learning on Supply Chain Management
Name:
Academic Affiliation:
Date:

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT 2
Introduction
Currently, business environment is facing stiff competition calling for great demand
uncertainty, greater risk of supply and increasing competitive intensity thus negatively impacting
the supply chain management of organisations. SCM is a concept that entails provision of the
accurate product, to the correct clients at the correct time, price and best property (Connelly,
Ketchen & Hult, 2013). Meeting all these requires planning, developing and disseminating
information across all the stakeholders : suppliers, manufacturers, retailer, transporters and
consumers. The association with all the stakeholders thus makes SC more information-intensive
requiring various key technologies for seamless flow of product and information. Thus
professionals have explored various ways of machine learning on the impact of the SC. The
current paper thus is an exploration of how ML can transform the supply chain.
Machine learning
Learning refers to the ability to gain knowledge through understanding particular skills
and following instructions. In regards to machines, learning is also a process that enables them to
perform better. Machine learning is a typical non-natural intellect that enhances the algorithm or
the software to learn and adjust without explicitly programmed to do so thus making the
technology to teach itself to improve operations (Bottou, 2013). Machine Learning is continual
of the Traditional Programming that entails both facts and database are fed on the processor to
harvest the outcome, unlike the current that requires both the facts and outcome run on the
processor to generate syllabuses. Each of the current machines contains different components
such as: Representations that shows how knowledge is represented such as in the tree decision

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT 3
tree Evaluation that enables the machines to make predictions and the optimisation that enables
the programs to generate processes.
Figure 1: Shows comparison between Traditional programming and Machine Learning,
obtained from (Bottou, 2013).
Types of Machine Learning
Machine Learning types are categorising into different categories:
Supervised and Unsupervised learning- the supervised learning refers to a situation where
computers are provided with examples of inputs that are marked with the required outputs to
enhance the capability of the computer to learn (LU, 2013). Thus supervised learning entails the
use of patterns to enhance the prediction of the values of the additional unlabelled data. The
common used supervised learning in organisations in the capability of the machine to filter
emails out of spam emails (Iosifidis, 2015).

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT 4
While unsupervised machine learning, refers to the situation whereby the computer is left to
identify the commonalities among the fed data. The unsupervised learning process aims to allow
the machine to find the desired pattern among the dataset and enable it to classify the identified
data accordingly. The unsupervised data thus is essential for the transactional data such as in
customers purchasing behaviour, essential in the Supply Chain Management.
Decision Tree Learning- Decision Tree Learning refers to the machine learning process that
entails the application of the visually represented decisions. Decision Tree Learning aims to
establish a predictive model that will predict the future targeted value based on the current inputs
variables that are determined through observation (Shere, 2017).
Deep learning process -aims at imitating the capability of the human brain in processing light
and sound stimuli into vision and hearing. The process is inspired by the biological neural
networks that consist of numerous artificial neural in layers that are composed of hardware and
GPUs (Jiung', 2017). The Deep Learning thus allows the machines to extract data through a
cascade of the nonlinear processing unit to provide artificial intelligence space, thus applicable I
recognising speech and images.
Reasons driving for Machine Learning
Numerous reasons require machine learning in current technology. Some of the reasons
are:
When the tasks are too complex to the program-In organisation, numerous activities such as
production are carried out by human that always have insufficient information on how to do the
same tasks. Some task such as speech recognition, driving and image understanding require the
art of machine (Wang, 2018).

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT 5
Tasks that are beyond the human capabilities -Some organisational tasks such as data sets are
always large to be comprehended by human capabilities. Analysis of such data is always
instrumental in providing clear direction to the organisation and when done by the human
workforce might take quite longer time. Thus machine learning can detect and interpret large
data(López, 2018). The learning process detects meaningful patterns that are large and complex
thus allowing fast processing of the data. Unlike the traditional supply chain the current, supply
chain has been digitalised with various aspects such as integrated planning and execution,
Procurement 4.0, Logistic visibility, smart warehousing, supply chain analytics and automated
B2C logistics.
Impacts of Machine Learning on Supply Chain Management
Machine learning impact on inventory control and planning
Machine learning enhances the analysis of large, diverse data sets to enhance inventory in
supply chain management. Inventory entails resources required in controlling and marinating
high customer service with a lowest substantial cost. According to Lolli et al., (2018) a single
unit cost about 15% to 35% of the product value in every year, and this needs to be controlled in
a competitive market. Such control is only achievable through having real-time information
entailed on the supply chain network. Supply chain management is currently faced with the
uncertainty of predicting the future product demands to have optimum production. Product
underproduction affects the flow of products as consumers receive insufficient products thus
making other customers look for alternative products.
While overproduction results in large wastages as some of the products end up unused,
posing large revenue loss to the company. The overproduction happens when a company

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT 6
overestimates product demand thus leading to large production that does not only affect the
production cost bust also raises the storage and labour cost. Thus proper forecasting ensures that
companies have the optimum supply at hand that can satisfy the market demand (Huiskonen,
2012).Through the application of technology that includes a wide range of statistical analysis
techniques, moving average and simulation modelling. Machine learning can take into accounts,
tracking and quantifying the existing factors over some time. Organisations with global
operations require sophisticated software systems that majors on product supply, demand and
market price fluctuations.
Machine learning has the capability of tracking stock in the warehouses. In numerous
organisations, inventory management as stocking and restocking becomes a challenge. Machine
learning thus can be effectively used in optimising the inventory storage thus offering transparent
supply chain communication between the responsible departments.
Al- techniques such as expert systems and ERP offer a new approach to inventory control
and planning problems. Machine learning has enabled the current Enterprise Resource Planning
to plan, design and forecast, an aspect that was not initially available in the new systems such as
EDI and the new versions of ERP. ERP can capture varied information in the SC thus enabling
the managers to estimate the most desirable level of output that will put the company in the best
competitive position in the market (Ogino, 2015). The ERP machine captures data and adjusts
calculations as well as predictions through different programs to enhance effective prediction and
planning for the required stock. For example in the current warehouses, the digital camera is
used to monitors stock level irrespective of the location of the warehouse, and the feeds
generated from the camera are analysed and used to create restocking alerts.

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT 7
Consequently, machine learning does not only help in tracking the stock but also helps in
optimising inventory management. Numerous companies consume much time in trying to
optimise their inventory which in many occasion result in unrealistic and unreliable information
(Reimann & Ketchen, 2017). Through the application of artificial intelligence, algorithms can be
adjusted and customised to fit the business to optimise inventory, especially business with
numerous distribution channels and locations.
Figure 1: Shows Iot-driven Inventory management, obtained from Lolli et al., (2018)

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT 8
Figure 2: Shows ERP system impact on supply chain management, obtained from (Ogino,
2015).
Machine learning impacts transport networks
The transport system has been the most significant aspect of supply chain management in
many organisations. Transport system enables the movement of raw products from the supplier
to the organisation and from the organisation to the customers. Machine learning thus aims at
reducing freight cost, enhance supplier delivery performance and reduce supplier risk in the
entire supply chain network (Abdulredaa et al., 2017). Machine learning can identify horizontal
collaboration synergies that exist in the supply chain networks such as vehicle scheduling and
routing problem, freight consolidation problem and intermodal connection problem, gas
distribution pipeline, traffic jam among many others.

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT 9
AI technologies can be applied to different systems with few modifications to adapt to
the desired transportation network design that can address different problems in traditional
technologies. The introduction of the meta-heuristic such as tabu search has the machines to
conduct some of the transportation tasks without the intervention of human being. Genetic
Algorithm has been applied in the biological concept and thus can be used in designing the urban
networks to ensure that roads are always free from traffic jam (Javaid & Siddiqui, 2018).
Through this, products and services are delivered to customers without delay or damages of the
product. Supply organisations also apply some AI to detect incidents such as accidents during the
transport system (Huiskonen, 2012). The machines are adjusted in such a way reports algorithms
to neural networks since they measure vehicle flow before and after the accident through data
collected from various sensors along the road.
Lastly, supply chain organisations use automated vehicle in making various product
delivery. Automated vehicle majorly relies on AI software and hardware that are deeply based on
the learning techniques (Pandey, Zhang & Jian, 2013). The hardware is composed of sensors and
computers system while the software is composed of navigations module, algorithm and
perception detect moving object. The vehicles are always taught how to drive while marinating
safe headways, discipline, control and lane to transform the transportation system (Sayed, 2014).
The application of the AV in the supply chain eliminates various barriers such as driver’s
reliability and vehicles accessibility. The vehicles are automated to move safely between another
vehicle on the road through the application of artificial intelligence partner recognition algorithm
along with sensor and the 3D cameras.
Machine learning and procurement

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT
10
Organisations have numerous supplier contracts that entail a large number of
transactions. The transactions at times are always prone to errors due to poor contact
management technology. Machine learning offers more comprehensive, complete understanding
and improved operations of the supplier contact with the entire penny spent as well as the ability
to plan for the future. Machine learning thus affects the procurement process through the
following process (Nijboer, Senden & Telgen, 2017):
Automated compliance- Machines can be taught and adjusted to have automated
compliance to enhance the capability of the machine to scan all the contracts. The applications of
NLP machine learning have the capability of extracting various suppliers’ information such as
names, renewal dates, signature dates among others that are significant from the procurement
process. Through this, the organisation can integrate the numerous contacts of suppliers within
the organisation alongside those that will be created in future to provide the right data; the
automatic renewal alerts to ensure that deadlines observed.
The automated procurement also increases efficiency by eliminating points that
consistently delay contacts. The application of NLP machine learning enables the organisation to
identify the correct suppliers that have matches the preferred qualities and characteristic of raw
material or product needs from one system thus saves both time and money and increase the
workflow of the SC (Telles & Ølykke, 2017). The automated and programmed predictive
analytic can identify when the majority of contracts are signed thus enhancing the procurement
process.
Machine learning enhances the negotiation process thus increases an organisation’s
saving. Numerous organisations tend to lose much money during the negotiations as the

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT
11
organisations focus much energy to explore on suppliers that suit their requirements (Luo, Yang
& Leung, 2015). However, through the Machine learning, organisations can search through their
existing contact list to match their profile thus highlight better opportunities for growing and
saving as well as capturing revenue. Machine learning also has the capability of providing
advanced supplier Quality Analyses to enhance various organisations to receive quality raw
materials. The algorithm analyses can identify suppliers that are providing the low-quality
product as well as the capability of the organisation to identify the accurate errors thus enhancing
quality products.
Machine Learning impacts on Customer Relationship Management
Every organisation's priorities customer trust to ensure its continual growth. To retain
customers, organisations must deliver what customers want at the right time, quality and price.
This is achievable through having various customers’ information as well as strong
communication to build a strong and long-term relationship (Amnur, 2017). Thus customer
relationship management has been the bets prerequisite to demand creation that affects all the
activities in the supply chain management. Customer Relationship management refers to all the
business practices that focus on improving service delivery, the building of the social bonds and
secure customers loyalty. Machine learning entails computational statistics that focuses on
prediction-making through the application of computer software. The computer software
measure customer relationship management by determine the Present Net Value of the profit
gained from various customers within a specific period. Currently, through Machine learning
have impacted customer relationship management in the supply chain in the following ways:

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT
12
Customer attraction- Through teaching the machines, programmatic advertising can serve to
produce customised adverts that can transform the customer’s behavior. Through the application
of AI, customised POS recommendations based on the purchasing patterns and histories are
provided to enhance customer’s decision-making process.
Customer profiling- Through the use of Machine Learning, supplychain organizations can
capture various customers data thus use Artificial intelligence to personalise various products.
The AI while analysing numerous transactions tends to focus on identifying various optimum
offers for individual customers.
Machine Learning and Order Picking Problems
Order picking is a significant aspect of the warehouse and inventory management
activates. The process consists of product retrieval from specific storage locations and delivering
them to customers according to the orders. The order picking process is always costly, labour
intensive activity and accounts from more than 50% of the warehouse operations. Insufficient
picking negatively impacts the supply since it leads to incorrect picking that affects customer
satisfaction and the business collapse (Charkhgard & Savelsbergh, 2016). Numerous systems
have been developed to enable the order picking process in the warehouses and can be classified
into four different ways: Picker to Part, Part to Picker, Sorting System and Pick to Box, however,
still these systems do not solve all the problems in the warehouse.
Through machine learning, these problems are eliminated. Machine learning optimises
the process in numerous ways such as reducing the number of steps that are taken during the
picking process thus saves both energy and eliminates any opportunity of damages and errors.
Secondly, picking multiple orders in the warehouse at times pose challenges to the workers as

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT
13
some end up forgetting or coming with wrong orders, thus through machine learning, the
machines can analyse the order and arrange them in the correct direction path to enhance the
picking process(Jakubiak, 2014). Additionally, the application of the classical heuristics that has
a u-shape enables the picker to transverse through any aisle with the least one item to be picked
as the picker advances.
Machine Learning reduces the potential risk such as fraud in supply chain
The movement of products from the manufacturing to the store is always faced with
different challenges such as climate, damages and fraud among many others. Some items are
always interchanged before they reach the final user thus learning to customer dissatisfaction.
Therefore the application of machine learning helps in automating the inspection through the
application of mobile technologies and inspectorio. Inspectorio refers to the machine learning
that addresses different frauds in the supply chain system thus resulting in transparency
(Caldwell, 2015). Thus, it eliminates the traditional inspection-model that usually used a pen and
paper. The inspectorio is significant to the supply chain since it enhances accuracy; speed and
inspection scale as well as provide real-time information.
Conclusion
In conclusion, business organisations face stiff competition leading to greater demand
uncertainty, greater risk of supply and increasing thus negatively impacting the supply chain
management of organisations. The business thus is required to continuously meet customer needs
through planning, developing and disseminating information across all the stakeholders such as
suppliers, manufacturers, retailer, transporters and consumers through machine learning.
Machine learning is a typical artificial intelligence that enhances the algorithm or the software to

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT
14
learn and adjust without explicitly programmed to do so thus making the technology to teach
itself to improve operations. There are different types of machine learning: Supervised vs.
Unsupervised, Decision Tree learning and deep Learning.
Machine Learning has great impact on the supply chain management through various
methods: Machine learning impact on the inventory control and planning such as control of over-
stocking, under-stocking as well as stock tracking. Machine learning impacts transport networks
such as the automated vehicle that enhance transportation system of items to consumers.
Machine learning has impacts on procurement process as well as Customer Relationship
Management. therefore, it is imperative that Machine Learning is inevitable and plays significant
role in the current organizational supply chain management as well as other sectors in the
organization.
References
Abdulredaa Kadhim, E., Haqqi Ismael, F., Challob Mohsen, H., & Hussein Jabbar, A. (2017).
Hybrid Approach for PRNGs using BBS and Dithering Technique. Indian Journal Of
Science And Technology, 10(27), 1-5. doi: 10.17485/ijst/2017/v10i27/113945
Amnur, H. (2017). Customer Relationship Management and Machine Learning Technology for
Identifying the Customer. JOIV : International Journal On Informatics
Visualization, 1(4), 12. doi: 10.30630/joiv.1.1.10

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT
15
Bottou, L. (2013). From machine learning to machine reasoning. Machine Learning, 94(2), 133-
149. doi: 10.1007/s10994-013-5335-x
Caldwell, T. (2015). Securing small businesses – the weakest link in a supply chain?. Computer
Fraud & Security, 2015(9), 5-10. doi: 10.1016/s1361-3723(15)30083-x
Charkhgard, H., & Savelsbergh, M. (2016). Efficient algorithms for travelling salesman
problems arising in warehouse order picking. ANZIAM Journal, 57, 166. doi:
10.21914/anziamj.v57i0.7666
Connelly, B., Ketchen, D., & Hult, G. (2013). Global Supply Chain Management: Toward a
Theoretically Driven Research Agenda. Global Strategy Journal, 3(3), 227-243. doi:
10.1111/j.2042-5805.2013.01041.x
Huiskonen, J. (2012). Service parts management: demand forecasting and inventory
control. Production Planning & Control, 25(6), 513-514. doi:
10.1080/09537287.2012.734449
Huiskonen, J. (2012). Service parts management: demand forecasting and inventory
control. Production Planning & Control, 25(6), 513-514. doi:
10.1080/09537287.2012.734449
Iosifidis, A. (2015). Extreme learning machine based supervised subspace
learning. Neurocomputing, 167, 158-164. doi: 10.1016/j.neucom.2015.04.083
Jakubiak, M. (2014). The influence of order picking zone’s configuration on the time of the order
picking process. Ekonometria, (3(45). doi: 10.15611/ekt.2014.3.10

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT
16
Javaid, T., & Siddiqui, D. (2018). Supply Chain Responsiveness and Supply Chain Performance:
The Role of Supply Chain Risk Management. SSRN Electronic Journal. doi:
10.2139/ssrn.3285077
Jiung', L. (2017). Deep Learning in Genomic and Medical Image Data Analysis: Challenges and
Approaches. Journal Of Information Processing Systems. doi: 10.3745/jips.04.0029
Lolli, F., Balugani, E., Ishizaka, A., Gamberini, R., Rimini, B., & Regattieri, A. (2018). Machine
learning for multi-criteria inventory classification applied to intermittent
demand. Production Planning & Control, 30(1), 76-89. doi:
10.1080/09537287.2018.1525506
López de Prado, M. (2018). The 10 Reasons Most Machine Learning Funds Fail. The Journal Of
Portfolio Management, 44(6), 120-133. doi: 10.3905/jpm.2018.44.6.120
LU, J. (2013). Semi-supervised multi-label classification algorithm based on local
learning. Journal Of Computer Applications, 32(12), 3308-3310. doi:
10.3724/sp.j.1087.2012.03308
Luo, X., Yang, Y., & Leung, H. (2015). Reward and Penalty Functions in Automated
Negotiation. International Journal Of Intelligent Systems, 31(7), 637-672. doi:
10.1002/int.21797
Nijboer, K., Senden, S., & Telgen, J. (2017). Cross-country learning in public procurement: An
exploratory study. Journal Of Public Procurement, 17(4), 449-482. doi: 10.1108/jopp-17-
04-2017-b001

IMPACT OF MACHINE LEARNING ON SUPPLY CHAIN MANAGEMENT
17
Ogino, T. (2015). Evaluation of Machine Learning Method for Intrusion Detection System on
Jubatus. International Journal Of Machine Learning And Computing, 5(2), 137-141. doi:
10.7763/ijmlc.2015.v5.497
Pandey, G., Zhang, B., & Jian, L. (2013). Predicting submicron air pollution indicators: a
machine learning approach. Environmental Science: Processes & Impacts, 15(5), 996.
doi: 10.1039/c3em30890a
Reimann, F., & Ketchen, D. (2017). Power in Supply Chain Management. Journal Of Supply
Chain Management, 53(2), 3-9. doi: 10.1111/jscm.12140
Sayed, A. (2014). Adaptation, Learning, and Optimization over Networks. Foundations And
Trends® In Machine Learning, 7(4-5), 311-801. doi: 10.1561/2200000051
Shere, L. (2017). A Relative Analysis of Multi-Relational Decision Tree Learning
Algorithm. International Journal Of Science And Research (IJSR), 6(1), 752-756. doi:
10.21275/art20164150
Telles, P., & Ølykke, G. (2017). Sustainable Procurement: A Compliance Perspective of EU
Public Procurement Law. European Procurement & Public Private Partnership Law
Review, 12(3), 239-252. doi: 10.21552/epppl/2017/3/7
Wang, H. (2018). Decision Tree Incremental Learning Algorithm Oriented Intelligence
Data. International Journal Of Performability Engineering. doi:
10.23940/ijpe.18.05.p3.849856