This isn't as "CPU-based" as I had hoped, just metaphorically taking the idea of Performance and Efficiency cores from the Core i5. The simulated warehouse has a Performance zone, with ground-level storage for items that are needed frequently, an Efficiency Zone, with high-density racks for low-frequency items, and a Shared Zone for mid-frequency items and overflow. The simulation showed that this idea worked best, but the paper doesn't investigate why.

Many years ago I operated the largest nationwide online DVD rental business in my country. Our warehousing evolved from alphabetical to numerical, with layout changes along the way. Eventually we handed the rearchiving of DVDs to the database itself and allowed it to allocate blocks of scanned in DVDs that were being returned. New content also was allocated by this system.

It was incredibly fun seeing things like cached new releases going straight to dispatch rather than going into the warehouse. Blocks of returning stock would be allocated to warehouse sections that had gaps that had the physical size of the block of DVDs that were being checked in.

Obviously, those empty areas were most commonly rotating stock, so over time the warehouse would become more active in areas, requiring a reallocation of stock to allow walking lanes for picking staff to not become congested. Fun times and all done on early Dell AXIMs with upgraded batteries, red laser scanners, and this new thing called wifi.

This is not particularly novel. Essentially any large-ish retail warehouse I have been in uses some kind of layout similar to their "CPU-based layout". And I have seen warehouses that were not explicitly designed like that, but this kind of layout came up somewhat organically.