“The main objective of QuART is to find more efficient approaches to solve the complex equations present in atmospheric simulations by using the properties of a quantum computer”
Quantum Approaches to Ray Tracing
Quantum computing is no longer a futuristic project! New technologies and computer prototypes are being developed at a rapid pace and algorithms are constantly being designed to exploit them. SURF, QuSoft and the WUR are jointly investigating how ray tracing can be used to model and simulate solar radiation in the atmosphere.
SURF, QuSoft (UvA/CWI) and Wageningen University (WUR) are joining forces and expertise to explore their applicability in the modelling and simulation of solar radiation in the atmosphere.
What is ray tracing?
When passing through different optical media, light can scatter, be absorbed or emitted, and reflect or refract from surfaces. These processes are governed by the “rendering equation”, more broadly known as the “radiative transfer equation”.
Ray tracing is a method for modelling light transport in media containing reflecting and refracting surfaces and in essence, it gives a numerical approximation scheme for the solution of the radiative transfer equation.
Read more to learn about this project!
How does it work?
After identifying a point of interest, which could be a pixel on a screen or a point in a three dimensional medium, rays of light are shot in different directions. These rays are then propagated by a small amount and based on the radiative transfer equation it is determined whether they should undergo scattering, absorption, reflection or refraction. The events determine the radiance- flux of radiation emitted per unit solid angle in a given direction by a unit area of a source - on the screen and the media (radiance as a function of space). After a scattering event a new direction is generated for the ray and the process is repeated. In the limit of infinite repetitions an exact solution for the radiance is generated, but of course in practice computational power determines the maximum number of iterations.
Where is it relevant?
Ray tracing is not only used extensively in computer graphics to render realistic images, but also in scientific applications where the interactions of the physical world with light plays an important role. For example, Wageningen University (WUR) is implementing ray tracing in atmospheric simulations to compute solar radiative fluxes. Cloud formation is driven by incident solar radiation and the heat and moisture released from the Earth’s surface. Clouds also affect the distribution of solar radiation on the Earth’s surface and in the atmosphere by partially reflecting and absorbing the incoming light. Atmospheric models need to solve the radiative transfer equation. Solving the radiative transfer equation is far from trivial and in general, a one dimensional atmosphere model is assumed (vertical one dimension). This assumption gives a qualitative but not very realistic approximation. Other alternatives are for example to use Monte Carlo integration to provide a more accurate three dimensional solution.
The (parallel) nature of the problem (following several beams of light as they propagate) suggests it can potentially benefit from quantum computers.
Ray tracing, a quantum problem?
Quantum computers are not meant to solve any type of problem or all problems. Quantum and classical computing technologies have complementary strengths and it is in their combination that we find benefits for a given application.
SURF and QuSoft (UvA/CWI) are looking for a hybrid quantum-classical approach for the ray tracing problem. The solution will then be integrated in the larger classical workflow for atmospheric dynamics simulations. Thus, the full application will be partly executed in HPC infrastructure and partly on a real quantum processing unit (QPU) hosted in a public cloud.
"Finding and testing applications using quantum algorithms also allows us to better understand what quantum computers can be good at, the type of problems they are suited for and what the necessary efforts are to map classical problems into quantum problems to take advantage of this new technology"
"QuSoft is a leading institute for the development of quantum algorithms in the right tasks that promise to take over existing classical algorithms. Ray tracing is a rich subject inviting applications of several distinct quantum algorithms, such as sum estimation and differential equation solving."
Mert is working on this project as a postdoctoral researcher. In the end, SURF and QuSoft hope to disseminate the methodology and the method itself so that the quantum, meteorology and -more generally- the wider scientific communities can benefit from the results and learnings obtained by this project.
Disseminate knowledge
For all parties, the partnership will mean expanding their areas of research.
The WUR will assist with domain knowledge and help guide possible integration into the final atmospheric simulation code.