Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz
AKPIK 6: AI Methods for Physics and Materials Science
AKPIK 6.4: Vortrag
Donnerstag, 12. März 2026, 17:30–17:45, BEY/0127
Probabilistic greedy algorithm solver using magnetic tunneling junctions for traveling salesman problem — •Ran Zhang1,2,3, Xiaohan Li2, Caihua Wan2,3,4, Raik Hoffmann5, Meike Hindenberg5, Yingqian Xu2, Shiqiang Liu2, Dehao Kong2, Shilong Xiong2, Shikun He6, Alptekin Vardar5, Qiang Dai6, Junlu Gong6, Yihui Sun6, Zejie Zheng6, Thomas Kämpfe5,7, Guoqiang Yu2,3,4, and Xiufeng Han2,3,4 — 1Present address: Max Planck Institute of Microstructure Physics, Halle (Saale), Germany — 2Institute of Physics, Chinese Academy of Sciences, Beijing, China — 3University of Chinese Academy of Sciences, Beijing, China — 4Songshan Lake Materials Laboratory, Dongguan, China — 5Fraunhofer IPMS, Dresden, Germany — 6Zhejiang Hikstor Technology Co. Ltd, Hangzhou, China — 7TU Braunschweig, Braunschweig, Germany
Combinatorial optimization is central to AI, logistics, and network design, yet classical methods often trade efficiency for solution quality. We introduce a probabilistic greedy solver that integrates spin-transfer-torque MTJ true random number generators with tunable switching statistics. A temperature parameter controls the balance between deterministic and stochastic choices. Applied to the traveling salesman problem, the framework achieves high-quality tours and surpasses simulated annealing and genetic algorithms in convergence speed, scalability, and computational cost.
Keywords: magnetic tunneling junctions; probabilistic computing; traveling salesman problem; true random number generator; spintronics