THESIS
2021
1 online resource (xv, 115 pages) : illustrations (some color)
Abstract
The product supply chain often includes the purchase of raw materials, production, product
distribution, and the sale of products. The exploration and analysis of the data in
the product supply chain are important for evaluating the operation of the product supply
chain, identifying potential problems, improving the supply chain, and adjusting the
strategy for product supply in response to market changes. However, challenges exist due
to the huge amount of data, the complicated relationship between different components
of the product supply chain, the increasing number and complexity of the machine learning
models used, and the uncertainty of the market. Prior studies on visualizing the data
in the product supply chain lack a detailed comparison of different algorithm outputs and
do not...[
Read more ]
The product supply chain often includes the purchase of raw materials, production, product
distribution, and the sale of products. The exploration and analysis of the data in
the product supply chain are important for evaluating the operation of the product supply
chain, identifying potential problems, improving the supply chain, and adjusting the
strategy for product supply in response to market changes. However, challenges exist due
to the huge amount of data, the complicated relationship between different components
of the product supply chain, the increasing number and complexity of the machine learning
models used, and the uncertainty of the market. Prior studies on visualizing the data
in the product supply chain lack a detailed comparison of different algorithm outputs and
do not show the reason for the difference between the outputs. What’s more, the uncertainty
of the market is not considered in these studies, which makes it difficult for users
to adjust the product supply chain in case of market changes. In this thesis, I propose
interactive visual analytics approaches for exploring and diagnosing multiple stages of
the product supply chain. Specifically, the approaches support the detailed comparison and selection of product demand forecasting models, enable the fast exploration and comparison
of different production plans, allow a quick adjustment to a production plan in
case of market changes, and help users to explore, inspect and diagnose the strategy for
the multistage distribution of products. The effectiveness and usability of the proposed
approaches are demonstrated by case studies with real-world datasets of the product supply
chain and interviews with experts working on managing the product supply chain in
large manufacturing companies.
Post a Comment