Jul. 23 – 29, 2022 Vienna IJCAI-22 Offline Time-Independent Multi-Agent Path Planning Keisuke Okumura, François Bonnet, Yasumasa Tamura & Xavier Defago Tokyo Institute of Technology, Japan ౦ژ޻ۀେֶ 5PLZP*OTUJUVUFPG5FDIOPMPHZ (OTIMAPP)

/9 2 goal (left robot) goal (right robot) https://kei18.github.io/otimapp/

/9 4 given start goal graph solution path Problem Def. – OTIMAPP s.t. all agents eventually reach goals regardless of action orders

/9 5 Solution Analysis no reachable* terminal deadlock no reachable* cyclic deadlock two conditions are necessary & sufficient *non-reachable deadlocks exist

/9 6 Computational Complexity 1. finding solutions is NP-hard 2. verification is co-NP-complete main observations the proofs are reductions from 3-SAT OTIMAPP is computationally intractable

/9 7 Solvers both solvers can solve large OTIMAPP instances to some extent agents 0 20 40 60 80 100 0 20 40 60 80 100 random-32-32-10 32x32 0 40 80 120 160 200 0 20 40 60 80 100 random-64-64-10 64x64 0 40 80 120 160 200 0 20 40 60 80 100 den520d 257x256 success rate (%) ≤ 5 min MAPF avoids collisions OTIMAPP avoids deadlocks propose two algorithms based on Multi-Agent Path Finding (MAPF) algorithms prioritized planning deadlock-based search extending conflict-based search [Sharon+ AIJ-15] extending conventional PP [Erdmann+ Algorithmica-87]

/9 8 Execution Demo no synchronization only local interactions centralized style with toio robots decentralized style with AFADA [Kameyama+ ICRA-21] all robots are guaranteed to reach their goals

/9 9 https://kei18.github.io/otimapp/ Takeaways OTIMAPP: novel multi-agent path planning problem that agents cannot share holding resources With OTIMAPP solutions, agents are guaranteed to reach their destinations without any timing assumptions as long as they avoid collisions locally OTIMAPP solution MAPF solution 2 1 4 3 0 0 1 2