AI-based demand forecasts for logistics, trade and production

© AdobeStock/Andrey Popov

The expected customer demand is an essential factor for companies when it comes to the correct planning of e.g. production processes or transports. And the better these forecasts, the better one can manage one's own resources - which in turn is not only interesting from a financial and organisational point of view, but also for sustainability reasons.

Demand forecasts of various kinds are already used today in many corporate areas, for example in inventory management, in staff scheduling or in personnel demand planning. Data-driven methods for forecasting demand have become established here in the last two decades. In this context, statistical, univariate methods such as exponential smoothing or ARIMA models were superior to all other methods in terms of forecasting accuracy for a long time. Now, however, the development of software, hardware and algorithms has advanced so far that forecasting methods based on machine learning (ML) have recently significantly surpassed the solutions of statistical benchmarks.

Application-oriented forecasting with the help of machine learning methods

There is still a lot of research and adaptation work to be done in this field, especially when it comes to actually transferring the theoretical models - as desired in our research field - into concrete applications. That is why an interdisciplinary team from the fields of mathematics, statistics and engineering is working on ML-based demand forecasts for logistics, trade and production in order to be able to provide appropriate software solutions for the concrete requirements there.

For this, on the one hand, the evaluation and benchmarking of the developed methods is important; the central question is: Do complex methods actually also provide a higher forecast accuracy in the present application case? For this purpose, the forecast must be compared with the actual development on the basis of historical data and the forecast error calculated from this. This can then be used to evaluate different forecast models. On the other hand, knowledge about the associated processes and levels of a forecast can also help to improve the forecast. For this purpose, our researchers look for structures in concrete demand forecasting questions that translate the real world into mathematics, so that they can be described in mathematical models and then implemented in algorithms. This enables coordinated and accurate forecasts.

 

From forecasting to decision support

Furthermore, the aim is not only to make point forecasts, but to quantify forecast uncertainties. For this purpose, different neural network architectures are used for time series forecasting and cross learning is applied, i.e. learning in a model on the basis of many time series at the same time. With the help of stochastic optimization, different predicted scenarios can be optimized simultaneously so that the best decision is found in relation to the expected value. This means that sales forecasts can also be linked to the resulting decision, so that concrete, operative decision support systems are created with this method, which companies can in turn integrate into their process infrastructures for decision-making - either automatically or by expert decision.

Our research in the context of AI-based demand forecasting

With this understanding in mind, the scientists in the Center for Applied Research on Supply Chain Services are researching to bring AI-based demand forecasts into application in logistics, trade and production.

Spare parts forecast for final storage

Towards the end of the life cycle of durable products, their manufacturers face a problem: Suppliers of product components cease production of these parts, but at the same time, manufacturers still want to be able to supply spare parts throughout the life of the machines. Accordingly, the manufacturers of the machines are forced to stock a large quantity of spare parts to ensure that spare parts availability is guaranteed until the end of the life of products. To do this, manufacturers must estimate the all-time demand for these spare parts as accurately as possible. In the research field of AI-based demand forecasting, the Center for Applied Research on Supply Chain Services is working on a forecasting method with which this all-time demand can be predicted in a data-driven manner. To do this, the curve of the product to be forecast is estimated from the curves of similar (historical) products. The researchers are currently working on a Bayesian formulation of the problem using growth curve models

Demand forecasts for inventory management for wholesalers

One of the biggest cost and success drivers in wholesale is warehousing. On the one hand, retailers strive to offer their customers the highest possible availability; on the other hand, high inventory levels or the subsequent scrapping of goods that have been left behind are associated with high costs. Currently, many wholesalers plan manually and use - if at all - univariate, statistical forecasting models. These forecasts are then used to derive orders from suppliers with the help of inventory heuristics. In the research field, the machine learning models that have been successful in recent years are transferred to applications in wholesale. The focus is particularly on the quantification of uncertainty in the forecasts and the combination with stochastic optimization. In this way, wholesalers can be offered solutions for inventory management that allow largely automated decisions.

Freight volume forecasts for decision support in road freight transport

Many planning problems in road freight transport are based on forecasts of expected freight volumes. While there are already many publications on the formulation of optimization models for these planning problems, data-driven forecasts on which these optimization methods are based have hardly been researched so far. In practice, data-driven methods for forecasting freight volumes have hardly been used so far - apart from individual cases - so there is great potential here. The Center for Applied Research on Supply Chain Services is therefore developing models for freight volume forecasts for road freight transport at various levels and applying them to freight companies.

Find out more about the Center for Applied Research SCS

 

Success and added value thanks to data

Sustained success in a changing world: this is the vision of the Center for Applied Research SCS

 

Scientific expertise in the reference process

The methodological expertise developed in our fields of research is informed by our specially developed Reference Process for Digital Transformation. Read about what this initiative means to us, and how we can use our expertise to comprehensively support companies.