Research of the Methods on Multi-Agent Path Finding
Ruining Zhou · Highlights in Science Engineering and Technology · 2023
Multi-Agent Path Finding is an essential real-life application of multi-intelligent systems for which scholars have solved many classical solutions. And with the evolution of theories related to machine learning, intelligent solutions are emerging daily. However, these algorithms have not been well summarized and concluded, and are comparatively difficult to consult. This essay's objective is to offer a thorough analysis of these issues. The paper will start with Path Finding and describe its three categories of classical algorithms. And then it will focus on the centralized planning algorithms based on the classical algorithms, and the distributed execution algorithms based on machine learning, and provide a theoretical explanation and comparison of their effectiveness. Finally, a prospection and outlook on the currently unresolved issues in the Multi-Agent Path Finding research area, including unification standards and directions for improvement, will be presented.