Introduction and motivation

Ever since the concept of digital twins has evolved, the supply chain industry has been harvesting the power to yield a higher amount of learning, understanding, and reasoning. Digital twins are well known to provide a dynamic simulation model of the supply chain problems with real-time data to draw insights into the past, current, and future operations. The overall objective of the digital twin is to increase the efficiency of the supply chain pipeline to drive operations efficiently and meet service level KPIs. Digital twin enabled pipelines to have capabilities managing a global network of products from sourcing to servicing and beyond. In addition, it also accelerates the momentum to better manage the economic, social, and environmental impact of supply chains by driving in more business growth from increased fill rates, speed to market, and strategic sales. The business growth and yield of a higher amount of revenue culminate in improved on-time delivery and improved working capital.

The overall beneficial impact of digital twins does not remain confined to final product delivery, but its impact is also realized in manufacturing processes, inventory planning, inventory optimization, and product distribution. Most of the challenges related to the life-cycle of the product flow can be replicated in a virtual environment with necessary feedback loops. Supply chain companies now have the flexibility to test and discover suitable actions for emergencies, giving a significant breakthrough to control cost, plan and implement remedial actions and improve organizational stability.

Data should be assembled, validated and issues related to missing / duplicate data should be fixed, before feeding the data in the digital twins. The use of data transformation and data mining tools can expedite us of AI and machine learning (ML) to enhance data quality. To further enhance the quality of data in the transformation processes, process mining can be utilized to extract and analyze process data before generating the digital twin of the supply chain processes.

Now let us analyze the benefits that can be reaped in the supply chain industry from a closely coupled set of 4 digital twins.

Optimize overall supply chain processes

A 4-unit closely linked digital twin can help to identify businesses understand patterns, and model the predicted outcomes in an attempt to monitor end-to-end supply value chains. This can help to identify risks and threats emerging due to delays in the downstream channels or shortage of raw materials or labor. AI/ML-driven operations at four elements of the digital twin can further facilitate business continuity and transformations through pre-computation of benefits, savings, and potential ROI.  

Planning

Planning operations in each of the 4 units of the digital twin helps to ease process optimization by better planning with different suppliers, contractors, and 3^rd party vendors. Automation processes resulting from digital twins can further reduce manual intervention and reduce duplication of efforts. These specially crafted plans are capable of evaluating timelines specific to design conceptualization, product design, cost approval, PO orders, and shipping process. Each of the phases can be separately analyzed to identify sources causing prediction delay, and accordingly, help to deploy more personnel to address the issues and mitigate the delay. A virtual setup employing the digital twin can forecast the issues, strategies the remedial actions, and see the real impact.

Identify bottlenecks and provide recommendations

Digital twins give an in-detailed overview of processes to help in the identification of problems occurring in different stages of the supply chain with minor human intervention. The most important aspect of DT is agile problem resolution by collecting data from different sources (such as sensors). Even data collected from shipment can help to identify obsolete, damaged inventory, and performance issues in different transport mediums (land, air) and also in delivery trips.  Further recommendation plugins when plugged in with DT, can serve to discover alternate suppliers and delivery routes to meet customer expectations. Better recommendations along with the right benchmarking of product quality in inventories can yield a significant reduction of logistics costs.

Schedule delivery across different transport facilities

Digital twins help to access changes in supply and demand across downstream channels (manufacturers->distributors->retail stores) in different locations, proving a way to effectively manage stocks in a way that controls stock overflow and underflow. Data related to finished products and raw materials can alleviate shortages by leveraging real-time data where in addition to demand and supply, external environmental factors (such as weather, strikes, lockdowns across a city) can be factored in. The real-time data can also strategize the best transport options so as to avoid situations like port congestions, and provide the required deliverables to customers within stipulated dates.

Optimize inventory

A supply chain pipeline with a digital twin setup aids in better demand forecasting to avoid stock overruns and underruns and optimize overall production and storage cost at warehouses. It can lay a profound influence on both the “single-echelon” challenge (optimizing inventory in a single warehouse) and the “multi-echelon” challenge (optimizing inventory across the entire network) to aid in better reliable inventory management.

Predict the performance of packaging materials

Digital twins have a pivotal role to play in packing raw materials and finished goods so that they can be packaged within optimal volumes and shipped to the destinations. The digital twins can simulate package shapes and packaging material which enhances defect testing cycles before deploying them and decreases the financial as well as environmental cost.

Product traceability with AR/VR

AR/VR when infused into digital twin systems helps customers validate the authenticity of products. Customers become interested to buy safe products which are sourced ethically. Data coupled with AR/VR engines can provide detailed information on the exact origin of ingredients, along with the product processing and manufacturing information. AR/VR when hooked into digital twins not only helps in retrieving the entire brand story through their mobiles, it also aids in pushing relevant offers to customers, thus promoting sales and marketing.

Real-time tracking with IoT

When goods and products are in transit Real-Time transit Information can be obtained by placing IoT sensors and AR markers on containers and packages. This aids in monitoring merchandise from checkpoint to checkpoint till its arrival at the destination.

Efficient management of product cannibalization

The key aspect of digital twins is taking corrective actions to promote sales towards any untoward event like an imperfect market segmentation. In such instances, customers are found to cannibalize from a high price market segment to a lower price segment and vice versa. Deterministic AI/ML models can feed in change signals from Digital twins and apply optimal strategies for pricing and order quantity for brands experiencing customers’ cannibalization. A strive to match the demand for the product, based on the current price and key product features, always remains the key focus. The best way DTs can help here is by setting up a continuous feedback cycle that can balance the price elasticity. Effective matching of customer sentiments/behaviors with the right product will be able to reduce returns, pouring in due to customer dissatisfaction.

Market segmentation and new product introduction

A typical digital twin is able to capture trends and behaviors of customers and access the potential market segmentation in case of the introduction of a new product. For a seller, it becomes important to sense customer’ preferential and attitudinal traits with respect to buying and decide whether to alleviate cannibalization by sequential product introduction or to introduce differential priced products simultaneously to continue to generate revenue from both the high and low product.

Product substitution

This comes as an added feature in the supply chain pipeline before the product is shipped to its destination. The right strategy is to map customer and market segments and ship relevant products pertaining to the market. In case of a mismatch, the goal is to find similar products, recommend them to end-users, and plan for their delivery.

Sustainability

The rise and application of digital twins have led to the development of a more sustainable supply chain as more CXO initiatives are directed towards more flexible and resilient networks. This can not only respond effectively to global shocks and disruptions, such as the pandemic but also work on cost savings and sustainability by establishing inter-connectedness among the different digital twins. Studies indicate 48% of enterprises have successfully switched to more sustainable suppliers and partners in their supply chains, while 40% have engaged external sustainability consultants in their next phase of the journey. The end goal is to reduce waste and satisfy consumer demands for greener products through advanced AI and analytics Moreover other emerging technologies like blockchain, IoT, graph algorithms are there to establish the connection among the entities of the digital twin, help in better delivery efficiencies, provide better routing, save fuel costs, reduce carbon emissions and drive the sustainable growth.

Ethics – Privacy, Accountability, and Transparency

Security and having a framework to allow sensitive information flow among the supply chain units can tremendously boost global technological, military, economic, and geopolitical strength. In whatever way we use AI/ML to leverage the business use-cases, predict, learn and feed in the model outcomes, the objective is to make the model robust, interpretable such that the factors can be certainly explained to business. In order to enable sustained business growth by employing digital twins, AI/ML models designed should abide by privacy and regulatory laws so as to be accountable and transparent in the process.

References

• https://research.aimultiple.com/digital-twin-supply-chain
• https://mitsloan.mit.edu/ideas-made-to-matter/post-pandemic-supply-chains-retool-a-new-abnormal
• https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.134.7744&rep=rep1&type=pdf