Worldwide uncertainty a chance to be more efficient and resilient, say academics

Companies and governments are working with Imperial to make smart decisions that improve efficiency, resilience and expected value in uncertain times.

With global issues like US–China relations, Brexit and the coronavirus pandemic looming large, decision-makers face major uncertainty both in their long-term planning and their operations.

But by building flexibility into their plans and deploying digital tools like algorithms, models and simulations, decision-makers can accommodate uncertainty and even take advantage of opportunities provided by the ability to adapt to unfolding events.

Academic insights on decision-making under uncertainty have been revealed in a new long-form feature on Imperial Stories, which offers an accessible introduction to some of the work Imperial researchers are doing in partnership with businesses and policymakers. The feature looks at a range of industries and problem spaces including postal deliveries, new energy systems, healthcare scheduling, space and finance.

Academic and industry insights

Imperial academics Professor Ruth Misener and Dr Michel-Alexandre Cardin also shared their insights in a panel discussion earlier this month alongside Maria Angeles Diaz of Agilent Technologies, a chemical analysis and life sciences company that has a long-standing partnership with Imperial.

The panellists, who were brought together by the Imperial Business Partners membership network, discussed frameworks for decision-making under uncertainty and the digital tools that can be deployed to support it. The event was chaired by Professor Peter Childs from the Dyson School of Design Engineering. 

Flexibility in long-term projects

Dr Michel-Alexandre Cardin, Senior Lecturer in Imperial’s Dyson School of Design Engineering, said that organisations sometimes commit themselves to major long-term engineering projects, for example the implementation of waste-to-energy systems, without building in flexibility because they make deterministic assumptions about the future or want to take advantage of economies of scale. But this makes them highly vulnerable to unexpected events and may mean losing out on the opportunity to take advantage of unexpected upsides.

Our world is facing important challenges and also opportunities in decades to come from climate change, healthcare, geopolitical tensions. Dr Michel-Alexandre Cardin

He said that long-term decision-making should instead be carried out in phases, for example, with the ability to adapt plans built in the system design at an early stage. This can allow them to respond better to changing conditions, become more resilient towards uncertainty and risks, and take advantage of upside opportunities.

Dr Cardin said: “Our world is facing important challenges and also opportunities in decades to come from climate change, healthcare, geopolitical tensions. Our systems designed today may be highly vulnerable and may not allow us to capture these opportunities. We need to rethink how we design these systems to make them more sustainable, resilient, with the goal of delivering better value. This can be done by exploiting the concept of flexibility, supported by more advanced computational tools.”

Expecting the unexpected

Professor Ruth Misener in the Department of Computing discussed some of the optimisation algorithms her group has developed with corporate partners such as BASF, the Royal Mail, and Schlumberger. These are designed to enable operations like postal deliveries and geothermal drilling not only to take place efficiently but to be recovered at minimal cost when unexpected events occur.

My job is to develop long-term visibility and sometimes this is particularly difficult because of the many things that are always going on in the world. What is always extremely important, is to remain agile and decisive, Maria Angeles Diaz Agilent Technologies

Describing the group’s work on scheduling with the Royal Mail, she said: “What we thought about is how can we plan the night before in a way that accounts for the unfortunate things [like breakdowns or staff sickness] that happen on the day, so that when the day comes, we can react very quickly.”

Maria Angeles Diaz, EMEA and India Sales Vice President & General Manager at Agilent Technologies, offered a business point of view. “Coming from an industry perspective, looking more specifically into business and organisations and how we deal with uncertainty, because this is impacting our business and our customers.…my job is to develop long term visibility and sometimes this is particularly difficult because of the many things that are always going on in the world. What is always extremely important, is to remain agile and decisive,” she said.

The human element

All three panellists emphasised that in addition to using tools like algorithms, we need to consider the human element.

One of the challenges of decision-making under uncertainty is communicating uncertainty with humans. Professor Ruth Misener

Ms Diaz said: “We are looking into how artificial intelligence, virtual engagement tools and many others… are helping us to make these decisions. We are still, today, learning, about how technology can help us to make some of these long-term decisions to support business strategies and business plans. Of course, I don’t want to leave out the human touch we believe is extremely important. At the end of the day many of these data [tools] needs to be validated by people who have the expertise and the experience, and I think that we should be keeping in mind that relationships are made amongst humans.”

Professor Misener said: “One of the challenges of decision-making under uncertainty is communicating uncertainty with humans. ‘Maybe this will happen or maybe this will happen and we’re not quite sure’, is very difficult to communicate… I would want decision-makers to actively engage with the uncertainty and explain the uncertainty to stakeholders.”

Partnership opportunities

Imperial works regularly with industry partners from a wide range of sectors to improve decision-making under uncertainty.

Professor Misener said: “What we do – and the Imperial Enterprise team is extremely good at this, and within our research groups we are good at this as well – is that when a new problem comes in we figure out what components the stakeholder needs, and then redirect the problem … for instance in the Computational Optimisation Group, we have me - I’m more of a mixed integer optimisation expert – and then there’s [Dr] Panos Parpas, who is a large scale nonlinear optimisation expert, we have [Dr] Calvin Tsay who is a control expert… and what we do is try to deploy the individual problem to the person who is best suited to it.”

Imperial also has a thriving ecosystem of startups founded by staff, students and alumni. Two startups with technologies relevant to decision-making under uncertainty, Quaisr and MosaiQ Labs, pitched at the event earlier this month.

Further reading

• For an accessible primer on the theory and applications of tools Imperial researchers are developing for better decision-making, view the Imperial Stories feature, Decision-making under uncertainty.
• For a perspective from Imperial Tech Foresight - Managing uncertainties: a foresight perspective
• Discover support for businesses available from Imperial Business Partners and elsewhere in Imperial's Enterprise Division.