The power of simulations has changed how decisions are made, but where is the technology going next and why?
Simulations have moved from being used for one-off capital investment-type decisions to now being used in real time to make regular operational decisions.
This advancement comes as processes have become more complex and granular, so we need assistance in understanding those details when making decisions. We now have the data and systems to involve simulation-powered digital twins in every one of these important decisions, meaning organisations can reach value quicker to gain competitive advantage.
Whereas before a simulation might have been focused on overarching objectives, like improving throughput or removing bottlenecks in a system, the combination with digital twins also makes it possible to demonstrate the most effective way to implement a process at a specific point in time.
For example, Ceva Logistics is a global supply chain management company that’s very involved in online retail and needs to manage high volumes of stock throughput where demands for the different lines can be very volatile. We worked with them to develop a simulation-powered digital twin that leveraged real-time data from its fulfilment centres so that they could plan ahead for resource allocation and minimise the risk of high volume volatility. They managed to save an average of 200 working hours per week and found a 2% increase in capacity thanks to far more accurate resource allocation to meet fluctuating demands. This simply wouldn’t have been possible without the combination of simulation and a digital twin.
For the simulation community, this represents a shift from one-off analysis to using simulation-powered digital twins for pin-pointdecision-making and at much greater frequency, which opens
up a world of opportunities to move operational research into operational excellence.
How do you believe simulation-powered digital twins will influence the field of operational research moving forward?
Businesses no longer have the luxury of just relying on long-term analysis to determine the best process. Why? Because in fastpaced sectors such as manufacturing, healthcare and logistics, the slightest disruption can impact daily or weekly workflows. Process owners have to make decisions in the here and now.
From an operational research perspective, it’s about adapting to the real time environment. How do we do what we do with ever-changing data sets and make that practical and feasible so decisions are made quickly?
The more frequently we can do this, the more granular decisions we’ll be able to influence. The aim is to go from making weekly decisions to daily ones, solving every problem not just the big ones.
What role will artificial intelligence and machine learning play in the future of simulation software and what impact will they have on traditional operational research methods?
At Simul8 we have always been innovators and driving technical firsts in simulation, so we’ve had machine learning incorporated in our tools for a while now. We’re also exploring the new forms of generative AI, and they have the potential to really open up what we do to a much larger audience.
Similar to how ChatGPT can support writing an article, it could become an analyst helping to develop simulations and analyse the results. This will accelerate build times and make operational research methodologies accessible to people from all sectors, process owners and decision makers, not just analysts.
It won’t replace us as specialists, however, because we are the ones to train the algorithms and understand the space. We need to monitor it to ensure it’s not making bad decisions, particularly in areas such as patient care.
Our mission has always been to take simulation from being something reserved only for the few and make it accessible to anyone that needs to make process decisions. AI will be a big enabler for this.