MAPF-EGT, an evolutionary game theory policy learner with weighted automaton rewards, is reported to beat deep RL baselines on large-grid multi-agent pathfinding.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.MA 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Multi-agent Path Finding for Timed Tasks using Evolutionary Games
MAPF-EGT, an evolutionary game theory policy learner with weighted automaton rewards, is reported to beat deep RL baselines on large-grid multi-agent pathfinding.