Application of neural network edges for a possible embedding into supply chain management

Temur Dzhuraev, BA (Hons)

Master's thesis

Abstract:

The purpose of this thesis is application of neural network advantages for a possible embedding into supply chain management. There were presented different types of neural networks fulfilling its own tasks but the chapter 1 was delved more into RNN and LSTM structures which have been practically considered and elaborated in subsequent chapters. The second chapter went into details about factors which must be taken into consideration during the neural network elaboration such as myriad types of optimization algorithms, distinction between standardization and data normalization procedures, overfitting and underfitting issues, evaluation indicators. The third chapter is dedicated to direct presentation of practical application of neural network in time series dataset. The neural network model has been designed on python language in the Jupiter Notebook software. At the end, the problem has been defined and the evaluation has been carried out regarding the data set specification and its features.

Abstract:

The purpose of this thesis is application of neural network advantages for a possible embedding into supply chain management. There were presented different types of neural networks fulfilling its own tasks but the chapter 1 was delved more into RNN and LSTM structures which have been practically considered and elaborated in subsequent chapters. The second chapter went into details about factors which must be taken into consideration during the neural network elaboration such as myriad types of optimization algorithms, distinction between standardization and data normalization procedures, overfitting and underfitting issues, evaluation indicators. The third chapter is dedicated to direct presentation of practical application of neural network in time series dataset. The neural network model has been designed on python language in the Jupiter Notebook software. At the end, the problem has been defined and the evaluation has been carried out regarding the data set specification and its features.

Keywords

RNN, Supply Chain Management, LSTM

Keywords

Supply Chain Management, LSTM, RNN

Language used: English

Date on which the thesis was submitted / produced: 12. 1. 2023

Thesis defence

• Date of defence: 16. 2. 2023
• Supervisor: Ing. Tomáš Malčic, Ph.D.
• Reader: David Holman, Ph.D.

ŠKODA AUTO a.s.

ŠKODA AUTO VYSOKÁ ŠKOLA o.p.s.

Master programme / specializace:
Business Administration / International Supply Chain Management