MACHINE LEARNING TECHNIQUES
FOR
SUPPLY CHAIN MANAGEMENT
FINAL CULMINATING
DEFENSE
Table of Content:
•Introduction
•Problem Statement
•Nature and Significance of the Problem
•Objectives of the Project
•Project questions
•Methodology and Limitations to the Project
•Literature Review
•ML Model
•Results
•Conclusion
•References
Introduction
•Supply chain management is the handling of the
entire production flow of a good or service.
•Vast network of facilities, including retailers,
distributors, transporters, manufacturers, and
customers.
•Challenges : Dynamic market conditions, changing
Consumer preferences, and the need for real-time
decision-making.
•Integration of Machine Learning has emerged as a
powerful solution.
•Ex: Procter & Gamble (P&G)
Problem
Statement
•The lack of utilization of machine learning
algorithms and models prevents organizations from
leveraging the full potential of data-driven decision-
making, optimization efforts, and overall
performance in their supply chain operations.
Nature and Significance
of the Problem:
• Problem : Manual process and Static
models.
• Consequences : Sub-optimal
decision making, increased costs
etc.
• According to a Bloomberg report,
the overall cost in the SCM has
decreased by 12% in last two years.
Objectives of the Project:
Develop machine learning Optimize inventory Improve supply chain
models for accurate demand management analytics and decision-
forecasting making.
Project Questions:
•What are the applications of Machine Learning and where should they apply in SCM?
•How does Machine Learning change supply chain management (SCM)?
•What are the specific domains within supply chain management where machine
learning models can be applied, and what outcomes and benefits can be achieved by
leveraging these models?
•What are the potential risks or limitations associated with the use of machine learning
in supply chain management, and how can they be mitigated?
•How do the outcomes of traditional methods compare to those achieved through the
implementation of machine learning models?
Methodology:
Identify Data collection Select Implementation Documentation
Specific Areas And processing ML Models (Data Pipeline including (Documented the
(Inventory Management, (Cloud Technologies for (Linear regression, Time storing, processing and process)
Logistics Optimization, Data – Aws, Azure, Gcp series analysis, Decision ML Model in cloud)
Resource Scarcity etc.) etc.) tress etc.)
Literature Review:
Tan and Le (2021) :
•Demand forecasting and Safety stocks in Inventory
management.
•Heuristic algorithm was introduced in order to identify seasonal
patterns more accurately. It helps companies plan production,
manage inventory levels, and prevent stockouts in the event of
unexpected demand or supply disruptions.
Chen and Wei (2020) :
•Demand forecasting and associated trends and gaps in
machine learning in the supply chain Industry.
•Concludes machine learning algorithms could provide better
accuracy and less computational cost for demand forecasting
than traditional forecasting models.
Implementing Data
Pipeline:
ML Model:
Python Libraries: Data:
• NumPy, Pandas, Seaborn, Matplotlib. •Data set into the three subsets.
• Import train-test split •training set - a subset to train a model.
• Import Linear regression •Validation set - To evaluate the model
•test set - a subset to test the model.
Results:
• ML applications can be utilized for demand forecasting, route optimization, inventory
management, supplier selection, and risk management.
• ML applications enables real-time insights, optimized routing, and streamlined operations.
• Businesses can achieve more accurate predictions, reducing excess inventory costs. In
logistics, ML can optimize routes, leading to cost savings and improved delivery times.
Risks:
•Include data security, model interpretability, and the need for skilled staff. These risks can be
mitigated by implementing robust security measures and investing in employee training.
Conclusion:
•Machine Learning in Supply Chain
Management reveals its vast potential for
efficiency, cost reduction, and informed
decision-making. While challenges exist,
the benefits, from accurate forecasting to
optimized logistics, are significant.
Emphasizing security measures, human
collaboration, and skills development
addresses concerns. In essence,
integrating ML in SCM is a strategic move
toward agile, data-driven, and efficient
processes, essential for navigating the
complexities of modern supply chains.
References:
•Chen, B., & Wei, J. (2020). Machine learning applications in supply chain management: A
systematic literature review. International Journal of Production Research, 58(7), 1977-1997.
•Li, X., & Huang, G. Q. (2021). Machine learning applications for inventory management in
supply chain operations: A review and future perspectives. Journal of Manufacturing Systems,
60, 300-317.
•Mishra, S., & Mankodiya, K. (2018). Internet of things (IoT) enabled supply chain
management. In S. Bandyopadhyay & K. Mankodiya (Eds.), Internet of things and big data
analytics toward next-generation intelligence (pp. 229-247). CRC Press.
•Panigrahi, B. (2020). Applications of machine learning techniques in supply chain
management: A systematic literature review. IIMB Management Review, 32(4), 356-369.
•Gupta, A., & George, M. P. (2020). Machine learning applications in supply chain management:
A review. International Journal of Advanced Operations Management, 12(3), 217-235.
Thank you
