Page 68 - eBook_Proceedings of the International Conference on Digital Manufacturing V1

Proceedings of the International Conference on Digital Manufacturing –
Volume 1

Traditional SCM principles, that focus on cost reduction and
efficiency, are being challenged by the dynamic nature of
contemporary markets. SCM has evolved from a logistics function
to a central strategic business component. Today, organisations
leverage SCM to achieve efficiency, lower costs, and enhanced
customer service. This is achieved through the effective
management of material, informational, and financial flows. The
need for agility—the ability to rapidly adapt to unforeseen
changes—has become paramount (Christopher, 2016). Supply
chains must be customer-centric, resilient, and sustainable to
succeed in this environment. Integration and coordination are
essential for streamlining processes, reducing delays, and
lowering costs. Vertical integration aligns operations within the
organisation, while horizontal integration fosters strong
relationships with external suppliers and distributors. Integrated
enterprise resource planning (ERP) systems facilitate better
coordination across departments and stakeholders, helping to
reduce supply chain inefficiencies.

Machine Learning in Demand Forecasting

Accurate demand forecasting is essential for effective supply
chain management, directly impacting procurement, inventory
optimisation, and distribution strategies (Chopra & Meindl, 2016).
Forecasting is a basic technique in SCM, crucial for demand
management and ensuring stock availability. This involves both
qualitative methods, like expert judgment and market surveys, and
quantitative methods that use historical data to estimate future
demand. Traditional forecasting methods often struggle to capture
the nonlinearities inherent in market behaviour and fail to adapt to
real-time dynamics, leading to inefficiencies and increased costs.
Accurate demand forecasting allows companies to plan
production schedules, stock inventory, and allocate resources
efficiently. Machine learning (ML) offers a promising avenue for
addressing the limitations of traditional forecasting methods. ML
algorithms can model complex, high-dimensional datasets and
leverage ensemble learning techniques to mitigate overfitting,
leading to more accurate and robust predictions (Goodfellow et
al., 2016).

52