Digital twin: a virtual mirroring and monitoring tool

What is a digital twin?

« A digital twin is a virtual representation of an object or system that spans its lifecycle, is updated from real-time data, and uses simulation, machine learning and reasoning to help decision-making ». (IBM)

In other terms, a digital twin enables to collect data and generate multiple scenarios in order to predict how a system will perform. Its mission consists in controlling decisions at operational level by reacting and adjusting to real-time events.

Benefits of digital twin

The first benefit is to gather in one digital place all the data collected (mainly via sensors) about different aspects of the physical object being studied. Operators then have a holistic view about the functioning - and dysfunction - of the original physical object.

This leads to a better monitoring of the object, with real time data collected to insure reliability and operational transparence with adjusted and recalculated situation (dynamic contextual informations).

The main goal of such a digital twin is to act as a super-simulator, by virtually replicating the entire environment of the object, hence enabling to run various simulations in order to study multiple processes in a holistic way.

Taking digital twins to the next level

The ability to easily run many different simulations on various aspects of a complex system is a huge step forward. But it’s still not enough to ensure that you are making the best decisions. With billions of possible scenarios to run, trying them all is not an option. To figure out what is the best decision to take, digital twins’ simulation features should be empowered with optimization solvers.

And this is the normal evolution in the Analytics Ladder: diagnostic and predictive analytics should be enriched with prescriptive analytics to take those insights to the next level by suggesting the best way to handle a future situation.

From the « what will happen » to the « what should I do », mathematical optimization technologies are the key, automating decision-making both on the operational, tactical, and strategic levels.

Digital twin or Modeling twin: making the right choice

What we call a « modeling twin » is a comprehensive mathematical formulation of a decision-making problem in order to lay the foundations for an efficient resolution powered by optimization technologies. Pretty often neglected or even ignored, this first step of setting the problem down correctly is yet fundamental to get to the right solution.

While digital twin and modeling twin follow the same logic — modeling a physical system — the fundamental difference between them lies in their ultimate aim. While a digital twin’s main purpose is to build an exact replica of a physical object to visualize, monitor, and simulate various situations, a mathematical modeling twin aims to abstractly formulate a specific business problem to solve, decide, and automate a situation.

• If you have a complex and wide project that you want to mirror digitally, then digital twin is the option that you should explore.
• If you are dealing with complex decision-making problems that you want to solve automatically, then modeling twin powered with advanced optimization solvers is the option that you should explore.

Of course, you might need both to virtually replicate an intricate system and solve within it various decision-making problems. Modeling twin and associated optimization solvers can enhance your digital twin to make it a powerful decision support ally.