Physics-Inspired Optimization

Harnessing analog neuromorphic computing for society’s hardest problems

We explore novel approaches to tackle optimization problems, utilizing neuromorphic and physics-inspired principles to solve them faster and more efficiently than traditional CMOS in von Neumann architectures.

Optimization problems are ubiquitous across modern society, needed in training artificial neural networks, constructing optimal schedules (e.g., airlines), path planning (shipping packages, VLSI wire routing), cryptography, drug discovery, and graph analytics (social networks, internet search). These problems are extremely challenging, requiring compute resources that scale exponentially with the problem size (i.e., NP-complete or NP-hard complexity). The research community is even exploring how using quantum computers or quantum-inspired approaches might help us tackle these problems better.

Mathematically, in a combinatorial optimization problem, one has a pre-defined cost function, c(χ) that maps from a discrete domain X(nodes, vectors, graph objects) to ℝ, the real number space, and the goal is to find the χopt that achieves the globally optimum cost value cmin( χopt) . We explore novel approaches to tackle optimization problems, utilizing neuromorphic and physics-inspired principles to solve them faster and more efficiently than traditional CMOS in von Neumann architectures.

We take a heuristic, rather than exact, approach. There are many powerful heuristic algorithms that have been developed, including simulated annealing, evolutionary algorithms, and Boltzmann machines that apply to many types of optimization problems. A physics-based approach is to map these problems to a Hamiltonian (such as an Ising system), and then find the ground state (lowest energy) of this system. This has inspired hardware development of physics-inspired Ising solvers, and our work is a part of this class.

Heuristic algorithms typically operate by starting with a candidate solution (a guess) and sensing the gradient around the current solution to move toward better ones. The injection of noise is typically required to allow hopping over barriers and avoid getting trapped in local optima. The system then iterates until convergence to a good solution.

The image shows a diagram with two main sections: on the left, a grid of interconnected nodes labeled as neurons with arrows indicating connections, and on the right, a graph with a downward curve and dots representing a descent process. (Mistral: Pixtral Large 2411, 2026-03-04)
A Hopfield Network with its dynamic update using Gradient Descent to find the minimum (optimal solution)

Since the landscape of potential solutions is exponentially large, searching through this space is very costly. Our approach is to 1) offer massive parallelism through many such units operating at once and conducting the search, 2) use in-memory computing to speed-up and reduce the energy of each gradient computation, 3) use analog computing to simultaneously reduce energy, compute faster, and provide a tunable source of very cheap noise. We have built proto-types to showcase all of these ingredients and shown quantitatively the great potential of this approach.

The image shows a collage with a detailed view of a chip, a colorful heatmap, a geometric diagram with lines and dots, and a graph with multiple lines representing data trends. (Mistral: Pixtral Large 2411, 2026-03-04)
Integrated Memristor-CMOS chip performing analog computation (with Hopfield Network) to solve an NP-Hard problem (Max-Cut) with injection of analog noise.

Mitglieder

Mohammad HizzaniBuilding TZA-Aachen / Room C3.25+49 241/92-780921
Dmitri DobryninBuilding 04.8 / Room 264+49 2461/61-3137
Alexandru CiobanuDoctoral ResearcherBuilding 15.22
Mathis MakarskiBuilding TZA-Aachen / Room C3.03+49 241/92-780421
Rishabh MallikBuilding TZA-Aachen / Room C3.17+49 241/92-780421

Last Modified: 18.11.2026