Batching with setup time $0$ and batches of at most one job is scheduling on a single machine: the batches are processed one after the other, and a batch with one job takes the processing time of that job, in serial and in parallel batching alike. Conversely, numbering the jobs of a non-preemptive schedule on one machine by start time (ties by completion time and then by number) and putting each job in a batch of its own gives such a batch schedule with the same start and completion times.

In serial batching the capacity may also be unbounded. With setup time $0$, splitting a batch into its jobs, processed one after the other in the order of their numbers from the start of the batch, completes no job later and starts none earlier than the batch. For a regular objective, one that does not decrease when a completion time grows (every objective here except the just-in-time ones), a batch schedule therefore gives a schedule on one machine of no larger cost.

Hence a problem of the family with operations that has one operation per job, one machine, no interruption, no machine-set restriction, processing times unit, equal or arbitrary, a precedence class other than $A_m$-orders, and none of the server, forbidden times, no-idle constraints and batch setups, reduces to the batching problem with the same number of jobs, processing times, release dates, deadlines, precedence constraints and objective: with bounded capacity, in either mode, and with unbounded capacity in serial mode when the objective is regular.

A specification $s$ of the family with operations reduces to a specification $t$ of the batching family if $s$ is of this kind and the batching specification obtained from it, in a mode and a capacity allowed for it, is field by field a special case of $t$.

Theorem. A specification that reduces to a batching specification describes a problem whose decision version reduces to that of the batching problem.