Multi-agent path planning in known dynamic environments