A well-studied abstract model for navigating agents in a distributed environment without collisions is known as Multi-Agent Path Finding (MAPF). MAPF has two interrelated aspects: one is to find the paths for the agents without collisions, and the other is to optimize the conflict resolutions between the paths. We focus on the first aspect, and we investigate three options: map reduction with waypoints, prioritized search and one-way restrictions. To our knowledge, our map reduction technique is a novelty in MAPF. We evaluate the three speed-up options on the scenarios of the League of Robot Runners competition. The results show that the improvement greatly depends on the size, style and layout of the maps. Prioritized search does not have big impact. The usage of one-way restriction rules is only worthwhile in specific maps. Map reduction on larger, complex maps drastically improves performance.