An optimization variant of a problem of path planning for multiple robots is addressed in this work. A proof of the claim that optimal path planning for multiple robots is NP-complete is sketched in the paper.