A minimal POS is the cheapest product of sums for a function: fewest sum terms first, then fewest literals. It is the mirror image of a minimal sop, and everything works the same way on the 0s:
- Start from the 0-rows, the ΠM list.
- Combine adjacent maxterms with the dual rule = X. In cube notation, the differing bit becomes a dash.
- Read each pattern back with maxterm polarity: a 0 bit gives the plain variable, a 1 bit gives the barred one.
10-is . - Drop redundant sums, including consensus terms.
A second route: find the minimal SOP of the complement F′ (group the 0s), then apply de morgans laws to it.
Why bother? The minimal SOP and minimal POS of a function can have different costs. Compare both and build the cheaper one: NAND-NAND for the SOP or NOR-NOR for the POS.