Quantum Optimization and the Future of Government

Quantum optimization is a direct practical application for quantum computing. Moreover, actors can already use it, even with the nascent and imperfect quantum computers currently available. The Volkswagen Group, Daimler, Ericsson, Total, Airbus (including with the Airbus Quantum Computing Challenge – AQCC)), Boeing, EDF, are examples of companies with ongoing research projects involving quantum optimization. Quantum software start-ups such as QCWare and Zapata Computing, and mammoth IT companies such as Google similarly highlight quantum optimization as one category for their use-cases.

• Fifth Year of Advanced Training in Early Warning Systems & Indicators – ESFSI of Tunisia
• Towards a U.S. Nuclear Renaissance?
• AI at War (3) – Hyperwar in the Middle east
• AI at War (2) – Preparing for the US-China War?
• Niger: a New Severe Threat for the Future of France’s Nuclear Energy?
• Revisiting Uranium Supply Security (1)
• The Future of Uranium Demand – China’s Surge
• Uranium and the Renewal of Nuclear Energy

Furthermore, in February 2019, the U.S. Defense Advanced Research Projects Agency (DARPA) created a whole program focused on quantum optimization: Optimization with Noisy Intermediate-Scale Quantum devices (ONISQ). Meanwhile, the Dubai Electricity and Water Authority (DEWA) also seeks to use quantum computing to address energy “and other” optimization and management (DEWA News, July 2018).

As far as quantum optimization is concerned, the future quantum world is therefore already almost here. Its impacts may take place tomorrow, but it is now the future is created.

And here we face a first hurdle. To foster interest and action in the quantum field, actors must first be able to imagine the benefit of their investment. They thus need to be able first to foresee the quantum world. Yet, this is particularly difficult (see Helene Lavoix, Foreseeing the Future Quantum-Artificial Intelligence World and Geopolitics, The Red (Team) Analysis, 28 October 2019). As a result, because it is challenging to understand quantum information science, hardly anyone outside quantum scientists and engineers consider current and future usages, as well as impacts of quantum technologies. This lack of awareness – with the exception of cryptography, takes place even in areas as crucial as security, defense, politics and geopolitics.

Interest in and discussions about QIS remain the preserve of an extremely small circle of scientists and engineers. Yet, those who have to consider the impacts of quantum technologies, take decisions about usage and funding, envision responses and strategies that need to include quantum technologies, are, most of the time, neither quantum scientists nor quantum engineers.

This series on strategic foresight and quantum technologies seeks thus first to foster imagination around the future emerging quantum world. It aims to do so in a way that is understandable to people who are neither quantum scientists nor engineers. Hence, it also seeks to contribute to bridging the gap between various communities, with different backgrounds, knowledge and interests.

This article starts practically imagining the future quantum world. It focuses on a first way quantum computing is likely to impact the future, namely through quantum optimization.

We first explain what are algorithms, quantum algorithms and quantum optimization algorithms, aiming for a “good enough understanding”.

Then, we use a concrete case – a research project involving quantum optimization that the Volkswagen group started with D-Wave in 2017 – to improve our comprehension of quantum optimization’s application. We therefore provide our imagination with concrete elements that will act as building blocks for foresight.

Finally, we imagine ways governments will use quantum optimization n the future, and even, actually, could already start using them, in the present. From solving the problem of “AI and the future of work” to possible quantum optimised resource management, we give examples of the way quantum optimization could revolutionise government. We then turn to possible applications for defence, armies and security. Finally, we look at what that may imply in terms of international influence and global power distribution.

A good enough understanding of quantum optimization algorithms

This part is aimed at readers who are neither quantum scientists nor engineers. It is thus for all those who will increasingly take decisions regarding quantum computing and quantum information science, use these technologies, and interact in a world where quantum technologies operate. Interested readers will find in the bibliography a couple of references for technically-focused (and advanced) approaches.

Algorithms and quantum algorithms

In the next video, David Gosset, IBM quantum computing research scientist, gives us clear explanations of an algorithm and a quantum algorithm. He points out why they are different.

Quantum Optimization algorithms

Optimization algorithms are algorithms that aim at finding the best solution to a problem out of a set of solutions, given some constraints.

When the problem involves many variables, it becomes impossible to run optimization algorithms on classical computers, even supercomputers, because too much computing power is needed. Quantum computers thus become the computing machine of choice. They are faster and use less resources (Ehsan Zahedinejad, Arman Zaribafiyan, “Combinatorial Optimization on Gate Model Quantum Computers: A Survey”, 16 August 2017, arXiv:1708.05294).

Currently, two main types of quantum computers are available. We can use adiabatic computers, such as those D-Wave develops, or gate-based quantum computers (for a detailed explanation on the types of quantum computing, e.g, National Academies of Sciences, Engineering, and Medicine, Quantum Computing: Progress and Prospects, Chapter 2, 2019).

Most types of current quantum computing efforts are gate-based. We have, for example, IBM and its quantum cloud offer, IBM-Q, with a maximum of 53-qubits microprocessor and Google and its 54 Qubit microprocessor, Sycamore (IBM’s new 53-qubit quantum computer is the most powerful machine you can use, MIT Technology Review, 18 September 2019; Elizabeth Gibney, “Hello quantum world! Google publishes landmark quantum supremacy claim“, Nature, 23 October 2019).

D-Wave and IBM machines are currently available for commercial use; Google’s machine is not. D-Wave’s computers, because if the chosen approach, are especially well suited to quantum optimization (see D-Wave’s explanation). For optimization algorithms, D-Wave currently, offers higher computing power.

Considering the so-far small number of qubits available and the high level of noise (for gate-based computers), “Quantum Optimization Approximation Algorithm” (QAOA) is the favoured current approach. Edward Farhi, Jeffrey Goldstone, Sam Gutmann developed it (“A Quantum Approximate Optimization Algorithm”, 14 November 2014, arXiv:1411.4028). The algorithm’s aim is to find an approximate or “good enough” solution for the optimisation problem and not the best solution (Ibid.). It is thus a compromise. It allows for using the new power of quantum computing even though the number of qubits is still small and the rate of errors or noise this small amount of qubits produces is still high. The results obtained are nonetheless better than what could be done with classical computing.

Unpacking Volkswagen and D-Wave quantum traffic flow optimization

The Volkswagen (VW) group started as early as 2017 a research project for traffic flow optimization with D-Wave. Computer scientists at Volkswagen sought to find a way to prevent traffic-jam in mega-cities, such as Beijing. They used taxi traffic data to optimize the taxis’ route and movements. They sought to be able to apply those findings in quantum optimization algorithms to other cases.

One year later, the VW group further developed the project with D-Wave, while starting new ones. Martin Hofmann, Chief Information Officer of VW, explains their research projects in the video below:

The VW Group and D-Wave work to

• Optimize traffic routes for a fleet of taxi (the initial project).
• Find out the perfect speed to the millisecond a self-driving car should use; send in real-time the signal allowing the car to use this speed. The aim is to avoid all stops and slow downs. Meanwhile, reliance on traffic lights stops.
• Optimize when and where taxis are needed. Here both quantum optimisation and deep learning are used. The latter seeks to predict taxis’ demand according to time and place. The final prototype succeeds in sending predictions to taxi drivers up to one hour in advance, which also reduces unproductive times and related costs.
• Optimize routes and types of vehicles in a city, in jam circumstances.
• The final objective would be to build a quantum-artificial intelligence “augmented mobility system” for a city, made of various predictive and optimization algorithms permanently interacting with objects, and controlled.

First, this case study shows us that optimization may also need to be coupled with the latest progress in artificial intelligence (AI), i.e. deep learning. This confirms what we expected when we started our deep dive in the future quantum world (e.g. The Coming Quantum Computing Disruption, Artificial Intelligence and Geopolitics – 1, 2018). Indeed, the 2019 consensus report Quantum Computing: Progress and Prospects of the U.S. National Academies of Sciences, Engineering and Medicine also links both in terms of potential applications (see p. 86). Coupling both quantum optimization and deep learning makes imagining applications easier.

Second, “time criticality” appears to be an ideal issue for quantum optimization (Tobias Strobl “Solving real-world problems with quantum computing“, BMI, nd). In other words, quantum optimization is particularly interesting when a problem involves “time-components”.

Finally, actors researching quantum optimization applications change. This point will most probably also be true for all quantum computing types of use. Here, we see the VW Group not only developing new possibilities for their traditional core industrial production. Volkswagen also sees new possible activities emerging (Strobl makes a similar point with regard to new business models, ibid).

Actors will thus see their expertise build up with research and as they construct upon achievements. Meanwhile, they will also see entirely new fields open up, they will be able to enter because of the new expertise developed. As a result, their activity can evolve, even substantially.

We thus witness the twin emergence of completely new usages and fields, and changing actors.

Imagining a world with quantum optimization

Bearing in mind the VW Group and D-Wave case study on the one hand, major problems and issues for political authorities on the other, we can now imagine ways to apply quantum optimization to government.

We take here a leap of faith with the capabilities and creativity of quantum algorithms researchers and with the ability of actors to create multidisciplinary teams including them.

Towards smart 3.0 polity planning?

Solving the AI and future of work problem

The impact artificial intelligence will have on work is a current, major worry that keeps many awake at night. Indeed, beyond excessive fear and ill-placed reassurances,

“…there is consensus in academic literature that AI will have a considerable disruptive effect on work, with some jobs being lost, others being created, and others changing.”

Consensus report, The British Academy for the Humanities and Social Sciences and The Royal Society, “The impact of artificial intelligence on work: An evidence synthesis on implications for individuals, communities, and societies”, September 2018.

As large parts of the world are already suffering of long-term unemployment while working poverty and inequalities are globally on the rise, further pressure on work and subsistance could trigger rising feelings of injustice and outrage, with, in turn a whole range of negative impacts (ibid. pp.34-37; IMF World Economic Outlook, October 2019, chapter 2; Richard Partington, “Inequality: is it rising, and can we reverse it?“, The Guardian, 9 Sept 2019; Durukal Gun et al. “The elephant in the room“, Barclays, 2 June 2017; Barrington Moore, Injustice). These negative effects could then snowball, converge and escalate, up to civil war and international conflict.

However, AI is also considered as beneficial. Furthermore, considering its drivers, AI will almost certainly continue to develop and spread (see ★ Artificial Intelligence – Forces, Drivers and Stakes and specific articles on each drivers). The key question, considering the possible impact on work thus becomes: how do we handle the disruption?

If we use the British Academy consensus report, then we find that future pressure on work results not only from AI but also from other factors. Moreover, one of the challenges is to manage a “time lag between the adoption of technology and its benefits appearing” (pp. 28-31).

We are thus actually faced with a problem of optimization, including many factors, compounded with “prediction” and including time-critical components.

Thus, we may imagine that quantum optimization and deep learning will greatly contribute – to remain cautious – to solve the transition to a world where various types of narrow AI will increasingly carry out many tasks (see, for more details, our series on AI).

Considering the vast amount of detailed data on citizens available to political authorities, those could be put to good use to optimize capabilities, training and education, and future changing work needs. To alleviate fears about choice and freedom – but honestly, which freedom is there in unemployment and living below poverty line – the necessity to offer (real) choices to citizens may be integrated from the start into the design of the new quantum-AI designed job disruption mitigation planning. Throughout their lives, the new planning platform will present citizens with series of choices for training and new guaranteed possible jobs. The quantum training possibilities will consider the citizens innate and acquired specificities, as well as their tastes. They will prepare them, ahead of time, to jobs that, for some, do not yet exist.

We shall thus become able to optimize dynamically and over the long period citizens’ skills, tastes and historically constructed socialisation, education and training, AI production of AI workers, as well as job markets and need for talents.

Quantum optimization and AI algorithms for government

Other types of quantum optimization and AI algorithms can be created with, as objective, to better handle the problem of resources. That issue is likely to become increasingly crucial and difficult to solve considering decades of unsustainable development and climate change. An early example of such a case, at the level of a city, is the strategic partnership between the Dubai Electricity and Water Authority and Microsoft for energy optimization (Press Release, Microsoft, 28 June 2018).

Emergency situations, with evacuation of large flows of people, are also candidates for the use of quantum optimization. They are a direct application of the VW Group and D-Wave research (Strobl, Ibid.). This application is even more interesting in the case of earthquakes. Indeed, we still do not know how to foresee earthquakes, thus evacuation under duress is crucial. Seismologic prediction, may also progress, thanks to quantum simulation, quantum sensing and metrology (e.g. University of Waterloo event, “The potential applications of quantum computation in exploration geophysics“, Feb 2019; Vladimir Kuznetsov, “Geophysical field disturbances and quantum mechanics“, 2017).

Industrial and trade policies, infrastructures, public services can also similarly benefit from the use of such quantum optimization algorithms.

Actually, this reminds us very much of central planning at state-level, as developed notably since World War I (e.g. Michael DiNoto, “Centrally Planned Economies: …” 1994; Andrew Gilg, Planning in Britain: Understanding and Evaluating the Post-War System, 2005). However, this new planning would be done with means undreamt of previously.

Towards a new type of government?

Compared with past central planning, we may wonder about the ideal type of unit for the new “quantum planning”. Could we, for example have to consider different scales according to different types of quantum optimization and AI algorithms? In other words, some quantum optimization problems could best be solved at city level, some at state level, others at region levels, others again at “specific areas” levels, etc.

Meanwhile, new types of staff and units will have to be included within states’ ministries and agencies, as well as at other levels of government (cities, regions, etc.). These will need to include multidisciplinary teams allowing for the creation of the new quantum optimization and AI algorithms. All necessary expertise will have to be included, not only of quantum algorithms researchers. Indeed, the aim will be to avoid a dangerous “over-technicisation” and to avoid losing accumulated understanding and expertise. On the contrary, we need to create teams that benefit from thousands of years of accumulated knowledge across disciplines.

As research proceeds to develop the best possible quantum optimization and AI algorithms, then new knowledge and skills, sometimes completely unexpected, will develop, alongside new ways to govern. As we saw in the case of the VW group, the various actor(s) involved will thus change. We shall progressively see emerging a novel form of political authorities, as expected from the ongoing paradigmatic transition.

Defence, armies and power

Defence and armies are clients of choice for the use of quantum optimization and AI algorithms. The DARPA (ibid.) already singled out “scheduling, routing, and supply chain management in austere locations that lack the infrastructure on which commercial logistics companies depend” as likely benefiting from quantum optimization.

Quantum optimization for extreme environments

We could most probably go further, first, with optimisation taking place not only in “austere locations”, but also in extreme environments.

By extreme environments we mean: cold (Arctic and Antarctica), hot (operations under intense heat waves for example), deep sea, space, and underground (see our series on Extreme Environment Security).

The future quantum computing power and optimization algorithms could handle the supplementary variables and factors related to the extreme characteristics of those environment. Furthermore, they could also factor in their changes according to climate change and extreme weather events.

Towards the quantum-AI battlefield

Second, we could also imagine going further than optimising current existing logistics, as well as deployment.

Quantum mules

For example, quantum optimization and AI algorithms could handle the coupling of advanced autonomous vehicles (e.g. drones) with soldiers to deliver in real time new necessary ammunitions, or other weapons better adapted to the enemy or the terrain or a change of action.

This would be a quantum variation and improvement on even the most advanced army mules (e.g. Matthew Cox, “Robotic Mules Could Deploy with Army Advisers to Afghanistan“, Military.com, 18 July 2019).

Quantum optimized cyber defense… and attack

Meanwhile, always thanks to optimization, cyber attacks could be carried out to disarm the enemy, open this or that defence, interdict reinforcement, etc. Here we should bear in mind all the new technological capabilities endowing the enemy (see Artificial Intelligence, Computing Power and Geopolitics – 2).

The need for new concepts and doctrine

Needless to say, being able to benefit from usable quantum computers and proper algorithms will fully be part of the new armament and capabilities of the army of the future. New concepts, doctrines and training would probably be necessary to create the soldiers and armies best able to take advantage of the new possibilities the quantum-AI algorithms create.

The quantum geopolitical disruption – The die is not cast!

If we go on being optimistic and imagine all these quantum and AI algorithms deliver on their promises, then the countries being able to create them, deploy them, then use not only each system of algorithms but also all systems together, will first be much stronger. Indeed, their political authorities will thence fully ensure the security of the ruled. They will thus be strengthened into their legitimacy.

Meanwhile, countries benefiting from a quantum-adapted government will also be richer, while the resources of the state, notably through an optimised industrial-scientific ecosystem and through taxes will increase.

As a whole, the use of a successful quantum optimization for government will renew and strengthen the social contract. It is not only that the political authorities will succeed in adapting the social contract to the new paradigm. They will also succeed in making the new paradigm serve the social contract.

By the same token, such a country will also be more powerful. Having been able to create, design and organise the novel tools of government necessary for tomorrow’s world, the political authorities will have developed the corresponding skills and knowledge. Those, in turn will boost the country and its political authorities’ influence abroad, including in symbolic terms.

Inversely, being unable to create and develop such new government is likely to rapidly drag a country to the bottom of the relative distribution of power.

Quantum technologies, as we saw here with the advances that quantum optimization will allow, usher a new very disruptive international game. Some states are already very advanced in terms of investments and developments of conducive ecosystems. Yet, the die is not cast. The very novelty of the change of paradigm, the capacity to think out of the box and, strategically, to seize and create opportunities will probably even the playing field, for those who want to play the game.

Featured image by Gerd Altmann from Pixabay – Public Domain.

Bibliography

For a technical approach to quantum optimization algorithms

Ashley Montanaro (mathematician), “Quantum algorithms: an overview”, Nature, npj Quantum Information, volume 2, article number: 15023 (2016), https://doi.org/10.1038/npjqi.2015.23

National Academies of Sciences, Engineering, and Medicine; Emily Grumbling and Mark Horowitz, Editors; “Chapter 3: Quantum Algorithms and Applications“, in Quantum Computing: Progress and Prospects; a Consensus Study Report, Washington, DC: The National Academy Press (2019), pp.57-94.

Patrick J. Coles et al. (for computer scientists) “Quantum Algorithm Implementations for Beginners”, 10 April 2018, arXiv:1804.03719v1

Olivier Ezratty (engineer), 504 pages report, Comprendre l’informatique quantique, septembre 2019 (in French).

References

DiNoto, Michael; “Centrally Planned Economies: The Soviets at Peace, the United States at War”; The American Journal of Economics and Sociology, Vol. 53, No. 4 (Oct., 1994), pp. 415-432.

Gilg, Andrew, Planning in Britain: Understanding and Evaluating the Post-War System, SAGE, 2005.

Gun, Durukal, Christian Keller, Sree Kochugovindan, Tomasz Wieladek, “The elephant in the room“, Barclays, 2 June 2017.

Kuznetsov, Vladimir, “Geophysical field disturbances and quantum mechanics”, E3S Web of Conferences 20, 02005 (2017) DOI: 10.1051/e3sconf/20172002005.

Moore, B., Injustice: Social bases of Obedience and Revolt, (London: Macmillan, 1978)

The British Academy for the Humanities and Social Sciences and the Royal Society; “The impact of artificial intelligence on work: An evidence synthesis on implications for individuals, communities, and societies”; September 2018.