From a naive perspective it might. It doesn’t work like that in practice though, as it wouldn’t be cost effective since it’d require lots of extra overhead. Mos

From a naive perspective it might. It doesn’t work like that in practice though, as it wouldn’t be cost effective since it’d require lots of extra overhead. Most any complex system will face a trade off between latency vs throughout. Shipping isn’t much different mathematically from a system of queues much like a database or network. There’s plenty of math for it: http://www.cs.binghamton.edu/~dima/ep01.pdf http://www.cs.binghamton.edu/~dima/ep01.pdf Optimizing costs requires keeping work queues as full as possible which requires increasing latency.