For thirty years, I’ve had the privilege of working amongst people applying the latest technological advances to protecting their countries.

One such advance is taking place now in the development of Synthetic Environment (SE) technology, where data and models converge to provide a simulation of elements of the world around us. SEs help us understand the world as it is, and help us explore the world as it could be, offering a crucial competitive edge to organisations that need to operate effectively in complex, fluid and often highly ambiguous situations.

Purposeful virtual worlds for faster decision making

SEs provide a safe virtual proving ground to create and test all sorts of ideas, from policy choices about the size and shape of the Armed Forces, to designs of new systems and technologies and exploring how best to use them. When decision makers have solidified those ideas into policies, plans and designs, SEs can be used to evaluate and improve their effectiveness. Applied to national resilience and security, SEs can be used to train people safely and economically, and to explore possible courses of action in a virtual world, before taking action in the real one.

But there’s a catch: as I found when colleagues in Dstl were supporting decisions on counter-insurgency operations in Afghanistan, building synthetic environments to support major decisions can take years. Understanding which factors are most pertinent to the question in hand, building mathematical representations of them and gathering sufficient data to validate those representations is a challenging and laborious process. And, as I’ll explain in a later blog in this series, the process for making complex decisions to design or manage policies, systems or organisations is rarely predictable and linear. It’s seldom possible to know in advance everything that an SE will need to include to support those decisions.

From “a model of everything” to “quickly modelling what matters”

Decisions like these follow more of a spiral path than a straight line from question to answer. This constant flux often leads decision makers to ask in advance for a “model of everything”, able to answer any question they might pose in future. But the truth is that, while it sounds like a great idea, we can’t model absolutely everything in minute detail, all at once. The world is too complex, and patterns too unpredictable, for any such model to be feasible.

Instead, we need to be more agile. We need to find a way to rapidly identify key areas of focus, spin up a synthetic environment quickly that contains those elements of a system that really matter – for example, the effect of cyber attacks on critical national infrastructure – and then hone in, interrogate it, zoom out, adjust and repeat.

As things stand, the process of creating and deploying synthetic environments to capture this richness and complexity takes years. We often don’t have years. We need to move much faster: days, hours, if not minutes. Such acceleration requires a transformation in the way we build decision support tools that allow decision makers to deepen their understanding of and thereby influence over fluid, fast-moving situations at the “speed of relevance”.

Clearing the path for faster decision making

The focus of my research team’s role at Improbable is the future thinking that clears the path for this kind of transformation. Our Myridian programme, backed by substantial internal investment, aims to tackle the seemingly intractable problems that need to be overcome to build a strategic capability for scientifically reliable simulation.

I’m privileged to lead a diverse team of researchers and software engineers, collaborating with a broad network of world-class partners. It takes a special mixture of disciplines, including mathematicians, computer scientists, engineers, complexity scientists, physicists, statisticians, software engineers and operational researchers to approach challenges like this. Our different perspectives are our greatest strength. This is amplified through our partnerships with leading academics in Bristol, Oxford, Leeds, Durham and Manchester Universities, the Alan Turing Institute and with like-minded companies.

But in getting the latest technological innovation into the hands of the organisations we serve, Improbable recognises that research is just the tip of the spear. The spear is pushed home through co-creating with our product and engineering teams, sat alongside our commercial partners and the user community. Myridian will operate in conjunction with SPRITE+ and Resilience Beyond Observed Capabilities network+ teams. We are excited about how this supports the work of leading researchers and decision makers to help the UK prepare for security threats in the coming decades.

This forthcoming series of blogs will shine a light on some of our work as we find ways of transforming the process of creating synthetic environments by:

• applying problem structuring methods to identify the strategically pertinent factors for the problem at hand;
• developing a range of new approaches to rapidly integrate relevant models into a SE, so that they can run at the scale, speed and fidelity needed to more realistically reflect the complexity of the modern world;
• using tools such as program synthesis to quickly and efficiently build new models of human behaviour to help the Cyber Physical Infrastructure become a Cyber Physical Social Infrastructure, and
• finding new ways of calibrating SEs to fit real-world data and assembling the logic to understand their validity.

Improbable has a busy – and stimulating – year ahead. Combining models effectively, making sure they’re performant, and ensuring their validity are just the start of what’s in store. Through original research which translates innovation into real-world application, we aim to create a new generation of multi-domain, multi-use synthetic environments that provides a transformational capability for the security and national resilience of the UK and our allies.