Mitsubishi Electric and QunaSys develop algorithms for Quantum CAE

Mitsubishi Electric and the quantum computing startup QunaSys have developed two foundational algorithms designed to advance the implementation of Quantum Computer-Aided Engineering (Quantum CAE). This development aims to enable large-scale physical simulations in manufacturing, such as fluid, thermal, structural, and electromagnetic field analyses, to improve product development efficiency.

The first algorithm addresses the high computational cost of handling large matrix data on quantum computers. By utilizing a method that reorganizes matrix data to reuse common patterns and structures, the team reduced the normalization coefficients that typically lead to excessive quantum gate operations. This allows for more efficient input formats for quantum circuits.

The second algorithm improves calculation accuracy by stabilizing the parameter calculations required for quantum circuit design. Specifically, it addresses instabilities in quantum signal processing where rounding errors and precision loss occurred during polynomial approximation. By reversing the calculation direction and then restoring the order, the new method mitigates these errors, demonstrating higher stability in tasks such as matrix inversion and Hamiltonian simulation.