DealMakerz

Complete British News World

Builds machines that will solve our most complex problems

Builds machines that will solve our most complex problems

Our world is full of intractable and time-consuming problems, called combinatorial optimization problems. Wallenberg scientist Johan Ackermann will build machines that can solve complex microsecond problems quickly and energy efficiently. It paves the way for more flexible technologies and tremendous opportunities to use society's resources more efficiently.

Wallenberg's world

institution:
University of Gothenburg

Research field:
Applied spin electronics

Combinatorial optimization problems are all around us. We may not think about it, but in everyday life we ​​often deal with challenges that can be solved in a wide variety of ways.

It is common for combinatorial optimization problems to quickly become unsolvable as the number of parts to be optimized increases, leading to an explosion of possibilities. A simple example is when you are traveling and want to pack your luggage in the car as efficiently as possible. Bags have different shapes and sizes that must fit into a limited space, leading to a huge number of possible combinations in the package, says Johan Ackermann, a professor of experimental physics at the University of Gothenburg.

There are many combinatorial optimization problems in the community, related to areas such as coding, resource allocation, motion planning, or scheduling.

-Which patients should be placed in the hospital? How can we improve the electricity grid and transmission capacity between countries? How can we make elevator movements more efficient in a hotel? Combinatorial optimization problems occur on smaller and larger scales, and we solve them well—probably 80 percent of the time. But if we wanted to solve it too well and too quickly, we wouldn't succeed today, he says.

See also  A systematic overview of the effects of hormonal therapy for gender dysphoria in young men

Effective calculations with oscillators

Johan Ackermann sets his sights on solving harmonic optimization problems with large networks of oscillators. Oscillators can be described as interacting electronic oscillating circuits that can perform advanced calculations – much like our neurons.

Over the past ten years, Ackerman has built world-leading research in this field.

Funding from the Wallenberg Foundations provides security and stability which is absolutely invaluable to research work.

– If we can solve combinatorial optimization problems seamlessly, society can get more out of its existing resources, and reduce all the “waste” that happens because we don't do things the right way. Getting nearly 20 percent additional resources would be hugely important in many areas, but even smaller increases would make a big difference. Think, for example, if we could improve traffic flow by five percent and remove a lot of traffic congestion, or if we could increase train traffic by ten percent on the rail network and reduce delays, he says.

Networks strive to achieve an ideal state

Today, a lot of research is being done on quantum computers to be able to perform advanced and ultra-fast calculations. But building useful quantum computers is still difficult and expensive, and requires a lot of energy to use.

Johan Ackermann's research focuses on building so-called icing machines. These consist of large networks of interacting nano-oscillators that can perform enormous amounts of calculations very quickly and energy efficiently.

Oscillators in networks are connected to each other and constantly interact with their neighbors. By influencing the strength of different bonds, researchers can program many compatibility issues. The oscillators respond by changing their phase states—being in phase or out of phase—and collectively try to reach the ideal state, where all the oscillators are optimally connected to each other.

See also  Prognosis in Nasopharyngeal Cancer | Karolinska Institute News

– The oscillators oscillate and their phase fluctuates until the system finds its best common state – thus solving the harmonic problem. The challenge is to create and program really large networks so that oscillators can solve important problems — for example, the distance between cities to find the shortest travel route, he says and continues:

– If we can solve combinatorial optimization problems quickly and energy efficiently, this could open up possibilities we have not yet thought of.

Paving the way for more efficient technology

Within the next few years, Johan Ackermann hopes to create a fully functional, commercially relevant Ising machine. He is already on his way. So far, the research team has been able to connect a network of 100,000 synchronous oscillators onto a very small surface.

– The oscillators are only ten nanometers in size, and are placed 24 nanometers apart in a grid. This means that it is currently possible to fit 1.7 billion oscillators onto a chip that is one square millimeter in size, he says.

Ultra-small formats open many possibilities for more efficient technology. Especially in small applications such as drones or mobile phones.

– He says that small chips with energy-efficient computational capabilities would facilitate many functions in mobile phones.

The world of electronics has long attracted Johan Ackermann. When he was 10 years old, he modified his ABC 80 computer and built his own electric organ. The driving force has always been to understand and explain how the world works, but also to be able to control processes and events.

See also  Brain training using a mobile phone - a list of the best applications

– I'm basically an engineer, and engineers like to make things. I've started several research-based companies, so, in addition to being a researcher, I'm also an entrepreneur, he says.

Text Ulrika Ernström
Portrait of Johan Wingborg