I am going through Richard Sutton book about Reinforcement Learning and I just encountered the tile coding method. I understood pretty well the principles, however, at the very end of the section, tile hashing is mentionned.
Hashing produces tiles consisting of noncontiguous, disjoint regions ran- domly spread throughout the state space, but that still form an exhaustive partition. [...] Through hashing, memory requirements are often reduced by large factors with little loss of performance. This is possible because high resolution is needed in only a small fraction of the state space.
I understand how hashing can consequently limit memory requirement, but I cannot visualize how hashing would work for tile coding. How can collisions be handled if tiles are spread randomly ? How are the tilings represented then ?
I have been doing some research, however, most of the ressources lead directly to the mentionned book.
