Seminar by Stephanie Riedmüller (Zuse Instite Berlin)
Date: 13 Apr. 15:00—16:30
Venue: 502A (5F), Science Bldg No.2.  理学部２号館502A号室

Title: Approximating Multi-Objective Max-Cut with QAOA

Abstract: We explore the use of quantum computing for multi-objective combinatorial optimization and demonstrate its potential to outperform classical approaches. Multi-objective optimization is a compelling problem class for quantum algorithms, since it can be classically challenging in both theory and practice, even if the
corresponding single-objective optimization problems are efficiently solvable.
We combine a low-depth Quantum Approximate Optimization Algorithm (QAOA) with the classical weighted sum method and apply it to multi-objective weighted Max-Cut problems. We demonstrate its performance on an IBM Quantum computer and via matrix product state numerical simulation, with benchmarking against classical multi-objective integer optimization methods.
Our algorithm not only finds unsupported solutions that cannot be recovered by classical weighted sum approaches but also shows an unexpectedly successful application of parameter transfer.