Demand forecasting and production planning tools

The Food Storage and Distribution Federation is supporting a project which is part funded by the Technology Strategy Board to improve demand forecasting and production planning within the chilled and fresh food environment.

Partners, including DeMontfort University, MSA, C4FF and Zest Produce are working together to produce a software program utilising Artificial Neural Networks to enable companies to improve their planning and forecasting.

This is one of thirty major collaborative business-led food processing and manufacturing development projects which have received more than £11 million worth of government funding, securing the UK's future as a leader in food technology. Including investment by the participating companies the total value of the projects is in excess of £23 million.

Rising costs and increased regulation around greenhouse gas emissions and waste pose a significant challenge to the UK's food manufacturing and processing sector.

Figure 1. Neural network forecast for next period

This research and development activity supported by government funding will look to improve production planning and demand forecasting in the chilled and fresh food sector. It will also assist companies in reducing waste within the supply chain by reducing supply chain scheduling inefficiencies.

This particular project will concentrate on the chilled and fresh food sectors of the economy and will look to benefit manufacturers, wholesalers, distributors and end users including hotels, restaurants, and schools.

The project will produce an intelligent demand management system to construct a Master Production Schedule based on demand forecasting taking into account, amongst others, data such as:

• weekly sales patterns
• weekend sales
• monthly cycles in line with receipt of salaries
• seasonal demand changes in line with such variables as temperature, holidays and festivals
• promotional demand peaks
• television listings
• sales trends arising from such factors as declining living standards
• the weather
• random fluctuations.

Research has shown that Neural Networks (NNs) when trained appropriately are the best forecasting solutions in comparison with conventional statistical techniques. The project's partners are working on the design and development of novel artificial neural networks to predict the demand of chilled and fresh food products and help companies to make informed decisions. This collaborative R&D project aims to improve the ability of Refrigerated & Shelf-Life Constrained Food Supply Chains to achieve the significantly higher levels of resource efficiency and due date adherence required to cope with current and future exponential increases in levels of product customisation, demand turbulence and supply chain uncertainty within UK and global markets. The project will also look to reduce waste within this specific supply chain area.