Optimierung mit Quantencomputing

Optimization is one of the core requirements for many companies. Everywhere business actions are optimized, be it the calculation of an optimal timetable, the ideal use of the means of production, or when to carry out maintenance work best without disturbing the rest of the operation. Often the optimization problems are so complex that they are difficult to solve classically and are accordingly time-consuming. This is exactly where the potential of quantum optimization lies.

The basis for this is, among other things, the tunnel effect, which can drastically improve the solution quality of the optimization process. Where a classic optimizer gets stuck in a local minimum, the tunnel effect helps to magically leave this and jump into a deeper (global) minimum. As a result, the global minimum of a cost function can be found with significantly higher efficiency. However, the optimization problems have to be formulated in the language of quantum mechanics. This is exactly what we are already investigating. For this we build graph models, which are the basis of the quadratic binary optimization approach, whose minimization function is as follows:

min ∑_i,j a_i,j x_i x_j + ∑_i b_i x_i mit x_i ∈ [0, 1]

These models can now be solved on both adiabatic and gate-based quantum computers. For this the hybrid Quantum Approximate Optimization Algorithm (QAOA) can be used. We are also using bridging technologies such as the D'Wave Leap2 Quantum Annealer and the Fujitsu Digital Annealer to deliver value for our partners today.

In the future, thanks to the high computing power of quantum computers, problems can not only be solved for all of Germany, but also much faster, so that e.g. dispatchers can be supported automatically in almost real time.