Digital twins: optimizing warehouse performance to reduce costs


Ecommerce is the business success story of the last few years. Faster processing and improved technology provide consumers with access to a wider range of products than ever. With behavior change accelerated by the Covid-19 pandemic, millions of consumers have switched their shopping habits to embrace the online marketplace. Global ecommerce sales reached around $6.3 trillion in 2023, an increase of 10.4% from the previous year. China makes up the bulk of online sales, with its revenue exceeding that of the rest of the countries in the top 10 combined, including the U.S., the U.K. and Japan. In fact, half of the 10 biggest ecommerce markets are in Asia, the most populous continent.

There’s room for more growth for ecommerce retailers but with the current slowdown in trade casting a shadow over economies around the world, there’s been a collective belt-tightening. More than ever, success hinges on the ability to manage the supply chain efficiently. It is essential to integrate technology into every aspect of logistics to reduce costs and optimize performance. Businesses need to ensure they continue to operate flexibly—able to manage both high and low volumes—in response to customer demand.

One tool to support reducing overheads for ecommerce retailers that’s emerging in this environment is the logistics digital twin (LDT), like the one that CEVA is currently trialing at a site in Großbeeren, Germany. Using a logistics digital twin will help fulfilment centers (warehouses) reduce hours and better allocate resources on the plant line to provide the necessary flexibility, while optimizing overheads and reducing cost.

CEVA fulfilment center

What is a digital twin?

Grieves, the inventor of the technology, defines a digital twin as:

… based on the idea that a digital informational construct about a physical system could be created as an entity on its own. This digital information would be a ‘twin’ of the information that was embedded within the physical system itself and be linked with that physical system through the entire lifecycle of the system.

In other words, a digital twin is a form of simulation modeling: a parallel digital entity to a physical system based in the real world. The digital and physical systems are linked via the flow of data between them in real time, so the digital simulation updates and is informed by the information available from a real, physical system.

And it’s not just about warehouses. The possible applications of this technology are extremely wide. For example, local governments could create a digital twin of traffic movement around a city to help improve its flow. The digital twin would receive all the information from traffic lights, traffic cameras and so on, to provide an accurate model, which would update in real time. This type of digital twin could provide forecasts to support planning, for example of how traffic would flow if traffic light phasing changed.


Benefits of a digital twin

In the digital age, data is invaluable. Access to data-driven insights allows companies to make informed decisions, forecast demand and adapt to changing customer preferences swiftly.

The logistics digital twin has several distinct benefits when it comes to warehouse operations. The first is in the planning and design of the facility. A logistics digital twin could use data from equivalent facilities to ensure the optimum layout, staffing and tech for product movement.

In an operational logistics warehouse, a digital twin could use information from the warehouse management system coupled with data on the physical layout of the warehouse, as well as updates on employee numbers and availability, to plot the most effective movement of product through the warehouse, the numbers of staff required at which stations and so on. The difference between a digital twin and a straightforward warehouse management system is that the digital twin is a more sophisticated tool that provides an additional level of forecasting and processing based on constantly updated real-world data.

A logistics digital twin would also receive input on any anomalies in its predicted performance. For example, if a product wasn’t available where it was expected to be or there was additional pressure on a particular picking area, the digital twin could analyze actual performance against its predictions and identify areas for improvement, which could then be applied in the facility itself.


CEVA's logistics digital twin project

CEVA Logistics prioritizes streamlined logistics operations, ensuring that products move seamlessly from suppliers to warehouses and, ultimately, to customers. CEVA’s commitment to optimizing transportation, warehousing and inventory management helps ecommerce businesses to reduce costs and improve delivery speed.

CEVA’s logistics digital twin project started at Größbeeren, near Berlin just over a year ago. The LDT operates with real-time data from a range of sources within and external to the warehouse. The core of the LDT is a discrete events Monte Carlo simulator engine. Essentially, it’s a black box that uses algorithms employing a mathematical model of repeated random sampling of data to obtain numerical results.

This central engine is located on site at Größbeeren, tracking, mapping and forecasting physical activity from data collected in real time from cloud and on-premise databases. It aims to help management make the right decisions about resource allocation inside the different processing areas of the warehouse.

The LDT processes and verifies the data it receives and then uses it to run scenarios, based on forecasts of the likely number of customers at given times. It combines the resulting predictions with the different KPIs relating to production and tracks the projected volumes against the actual volumes that get processed in the warehouse. This gives a complete picture of the optimum possible performance and how the different areas and factors within the warehouse are performing in relation to this optimum result.

Once the tool is operating at full strength, with trained operators in place, it will reduce unproductive time and help the control room track what’s happening on the warehouse floor, in

almost real time, to keep production to the forecasted plan. The insights will allow the warehouse team to improve its use of resources and increase the overall operating capacity of the warehouse.


Movement of data through the digital twin


As the schema indicates, the central LDT engine collects activity tracking data in real time from the warehouse management system database and other databases that the simulation then processes and verifies. Once the core engine of the digital twin has processed the data, it sends the information to a reporting tool (TableauTM) to be displayed in the form of charts in the control room. Here, a trained operator can make decisions relating to the resources needed for upcoming warehouse shifts.

The LDT displays its results via Tableau on an hourly basis so managers can compare the processed volumes and see the optimal allocation of people and products. Decisions made by the operator close the digital twin’s information loop, as these real-world decisions affect the data collected by the database.

Closing the loop between the real and the digital world is a key point in the operation of a digital twin. There should be a constant exchange of information between the two worlds.


Development of the digital twin industry

The ultimate aim of the digital twin is to allow people to make the best use of the data, to empower management to make informed decisions when running a facility, to optimize operations, to enhance visibility, and to improve overall efficiency. The result of this optimization will mean more efficient use of resources; in terms of energy, people and raw materials, all of which increases sustainability and reduces cost.

There’s huge potential for digital twin use to improve logistics operations worldwide. But the digital twin industry is still in its infancy. Many of the building blocks of the technology have been available for some time but it’s only the recent increase in process monitoring and real time data collection that has enabled the development of digital twin operations.

Standardization is crucial for the industry to take its next steps. CEVA’s logistics digital twin in Größbeeren is just one project. Other projects in different countries and companies will not necessarily use the same system design or methodology. Future developments have to focus on the integration of systems, with interoperability the buzzword of the moment. It’s essential that the various systems can all ‘talk’ to each other effectively.

The International Standards Organization (ISO) responsible for industrial product data, ISO TC184, is currently collecting business cases of digital twins to understand how best to transform this new concept into a mature tool. It’s only with standardization of the digital twin domain that single projects can grow into an easily developed, routine tool that is straightforward and cost effective to use. That level of development would then give rise to mass adoption by logistics operators, so the ecommerce industry could realize all of the cost and sustainability benefits.