| dc.description.abstract | In omni-channel supply chain, the process of selling, fulfillment and distribution is
always needed to be improved and optimized. For this reason, this thesis proposes an
approach to solve the problem from demand forecasting by machine learning to
distribution planning in supply chain with the objective optimizing the costs. Clustering
and getting underlying patterns to improve the forecast through neural network, then
linked to mixed-integer-programming to have a plan about production quantity,
distribution quantity problems, are the objectives of this paper. This paper investigates
how the features and the neural network affected on forecasting accuracy. Furthermore,
which factors impacted strongly on total cost during the process of production
distribution are also considered as final result. Finally, this thesis can propose a
horizontal planning for forecasting and distribution. | en_US |
