Manufacturing Leadership Journal

As supply chain optimization involves multiple systems acting together with multiple input and outputs, it is fundamental to understand how the systems interact. With new technologies that enable the parallel processing of enormous data sets through pattern recognition engines, we are beginning to address these challenges by enabling modeling of broader digital twin ecosystems, examining the relationships and interactions between agents across supply chain networks including customers, factories, warehouses, suppliers, logistics, etc. We call this a “multiverse” approach, and we believe it allows us to evaluate potential opportunities that would be missed by logic-based decision platforms.

Due to improvements in computing, including in quantum computing, we are increasingly able to construct and imagine digital twin modeling environments in which we can effectively compute multiple internally consistent versions of the future – a “multiverse” of realities3. Using this approach, we can help choose what future to move towards versus being locked into decisions imposed by deterministic predictions that are rarely correct. Because multiverse modeling at its core allows infinite scenarios, we are starting to enable a new form of supply chain modeling where we can consider potential systems of systems effects across the extended network that heretofore could not effectively be represented or perhaps even imagined.

These types of models
not only break new ground
for understanding and op-
timizing supply chains, they
also will improve the modular digital twin
applications described above by incorpo-
rating systems effects into the applications.

Think, for example, of the systems effects of supply disruption or exchange-rate volatility on overall systems performance.

As time progresses and updates are received, we continuously compute the new physical realities encapsulated in our models and can rapidly identify new supply chain optimization opportunities and strategies, ranging from basic risk mitigation at the edge of the network to broad strategic pursuits of new cost positions or business models.

We are already deploying hundreds of thousands of digital twins in our supply chains and expect the performance impact to continue to improve as our self-learning algorithms in the digital twins are trained and refined through real world experience.

We believe that, ultimately, autonomous systems will naturally emerge in the near future as digital twin technology is paired with a robust industrial internet and global digital trust fabric, enabling concurrent information exchange across dispersed global supply chain networks.

The brave new worlds of possibilities in global supply chain optimization through digital twin technology may be frightening to some, but it is also exciting and will eventually become a cornerstone in how GE and other companies start thinking about and planning their transitions from traditional industrial companies to digital-industrial companies. While it is a long-term vision that will effectively play out over the next five to

10 years for most manufacturers, there are many individual steps that manufacturers can pursue today quite profitably, depending on their most critical cost, revenue, and busi-nessmodeldrivers. M

MANUFACTURING LEADERSHIP JOURNAL

..........................................................................................................................................................................................................................................

www.ManufacturingLeadershipCommunity.com ................... ................... ................... ................... ................... ................... ...................

"We are
starting to
see a journey
toward brave
new worlds of
global manu-
facturing and
supply chain
management
where digital
t win tech-
nologies and
strategies are
enabling new
approaches to
monitoring,
managing,
and optimiz-
ing increas-
ingly complex
global supply
chains."