Finite capacity planning: why reality always wins

An infinite-capacity plan answers one question: when should each order be produced if every machine were always free? The answer is easy to compute and rarely true.

A finite-capacity schedule answers the question that matters: given the machines, people, materials and time we actually have, when will each order really be finished?

What finite capacity considers

• Resources — each operation runs on a machine that can actually do it, one job at a time (or N at a time where parallel capacity exists).
• Calendars — shifts, holidays and planned maintenance remove time from availability.
• Setups — changing colour, tooling or material costs time, and the sequence decides how much.
• Precedence — an operation cannot start before the previous one ends.
• Materials — work cannot start before the material is there.
• Constraints — rules such as “this product only on these machines” or “never this sequence”.

The data it needs

Finite capacity is only as good as its inputs: realistic times, correct routings, real calendars, current progress. That is why it works best on top of an MES that records what actually happens — and why planning projects so often turn into data-quality projects.

What you get in return

• delivery dates you can promise with confidence
• bottlenecks visible before they cause lateness
• setups grouped deliberately instead of by habit
• a baseline to measure every disruption against

Reality always wins. Finite-capacity planning simply lets you see it coming.