Faculty of Computer Science AGH and IBM Software Laboratory in Krakow invite to Krakow Quantum Informatics Seminar (KQIS)
Objectives:
• understand and discuss current problems in quantum informatics,
• discuss new quantum computing technologies,
• exchange ideas and research results,
• integrate information across different research teams,
• build a community around quantum informatics.
Program:
Tuesday, 22 October 2024, 9.35-11.00 via Teams
Arkadiusz Wołk, Faculty of Computer Science, AGH Krakow, PL
Abstract
Quantum Approximate Optimization Algorithm (QAOA) [1] in its standard form allows for solving unbounded combinatorial problems. As many real-life problems require incorporating constraints, the most common method for using QAOA is to combine cost function and the constraints using penalty weights. In this talk, we will present alternative RU2023 [2] method, which can be used to handle inequality constraints in QAOA mixer without explicitly including them in the objective function. In this work, we show how to apply Hamiltonian simulation algorithm from [3] to obtain fully functional RU2023 circuit implementation. We will discuss the precision and the complexity of Hamiltonian simulation, the results of applying RU2023 method for sample problems, its limitations, and practical assessment.
References
[1] Edward Farhi, Jeffrey Goldstone, and Sam Gutmann. A Quantum Approximate Optimization Algorithm. 2014.
[2] Yue Ruan et al. “Quantum approximate optimization for combinatorial problems with constraints”. In: Information Sciences 619 (2023), pp. 98–125.
[3] Graeme Robert Ahokas. Improved algorithms for approximate quantum Fourier transforms and sparse Hamiltonian simulations. en. 2004.
Bio Arkadiusz Wołk is currently a PhD student at the Faculty of Computer Science AGH Krakow. His interests include physics, and everything related to computer science.