IBM develops quantum algorithms for differential equations.
IBM researchers have developed quantum algorithms that solve complex differential equations with exponential speedups, bringing the industry closer to practical quantum circuit simulation.
IBM senior researcher Hari Krovi and his team are designing quantum algorithms to solve differential equations, targeting areas where classical computing struggles. Their recent work focuses on differential algebraic equations, which combine differential equations with algebraic constraints. Specifically, the team applied these methods to RLC circuits containing resistors, inductors, and capacitors. This research represents an initial step toward building a quantum circuit simulation tool comparable to the industry-standard SPICE software.
The mathematical foundation of this approach relies on mapping equations to linear systems. Krovi's team utilizes the HHL algorithm, named after its creators Harrow, Hassidim, and Lloyd, which solves linear systems of the form Ax = b. While classical methods scale linearly with the dimension N of the matrix, the quantum approach scales logarithmically as log N. For RLC circuit simulations, the quantum algorithm runtime scales logarithmically with the number of components, whereas classical SPICE simulations typically scale linearly.
Beyond electrical circuits, the researchers are exploring integral differential equations, which incorporate historical data to model systems with memory, such as those in epidemiology. The team is also looking at fluid dynamics, specifically the Navier-Stokes equation. Currently, quantum methods can only resolve Navier-Stokes under mild, laminar flow conditions, such as the Viscous Burgers equation, but they fail in highly turbulent regimes characterized by a high Reynolds number.
Despite these promising speedups, significant bottlenecks remain. Krovi identified data loading and the extraction of classical information as primary bottlenecks. To address these limitations, Krovi is collaborating with Andrew Childs at the University of Maryland. Future research will focus on expanding the quantum SPICE framework to include nonlinear elements, as well as exploring applications in biology and materials science.
This is our own summary of reporting by IBM Research AI
