日本フィジカルAI新聞

世界のフィジカルAIを、日本語で。

週刊ニュースレター購読
量子マルチエージェント制御arXiv:2207.14404

もつれが導く待ち合わせ:ネットワーク上の移動エージェントに対するベル非局所性の応用可能性

Entangled Rendezvous: A Possible Application of Bell Non-Locality For Mobile Agents on Networks

シェア:XThreadsFacebookLINEはてブBluesky

位置情報を持つ2体のエージェントが限られたステップ数でネットワーク上で待ち合わせる問題を、量子もつれと古典計算の両方で解析し、三次グラフやサイクルで量子もつれが優位性をもたらすことを示した。

著者: Piotr Mironowicz

分類: quant-ph, cs.MA, cs.RO

原文アブストラクト

Rendezvous is an old problem of assuring that two or more parties, initially separated, not knowing the position of each other, and not allowed to communicate, meet without pre-agreement on the meeting point. This problem has been extensively studied in classical computer science and has vivid importance to modern applications like coordinating a fleet of drones in an enemy's territory. Quantum non-locality, like Bell inequality violation, has shown that in many cases quantum entanglement allows for improved coordination of two separated parties compared to classical sources. The non-signaling correlations in many cases even strengthened such phenomena. In this work, we analyze, how Bell non-locality can be used by asymmetric location-aware agents trying to rendezvous on a finite network with a limited number of steps. We provide the optimal solution to this problem for both agents using quantum resources, and agents with only ``classical'' computing power. Our results show that for cubic graphs and cycles it is possible to gain an advantage by allowing the agents to use assistance of entangled quantum states.

PR本紙発行元 EmplifAI