Optimizing Field-of-View for Multi-Agent Path Finding via Reinforcement Learning: A Performance and Communication Overhead Study

This work investigates the impact of Field-of-View (FOV) on the performance of Reinforcement Learning (RL) models in Multi-Agent Path Finding (MAPF) problems. The study measures the effects of different FOV settings on RL performance, communication overhead, and computation time. Results show that the tested smallest FOV (3x3) reduces communication frequency by 28.9% with only a 1.65% reduction in success rate compared to the baseline (9x9). The study also compares computation time for different FOV for efficiency analysis and provides insights into FOV selection considering computation cost.