One model, from the 24-month horizon down to the detailed schedule

The situation

A top 20 biopharmaceutical manufacturer runs a production process with long product cycles, dense interdependencies, and so much process specificity that the plant is, in effect, built around the process. Planning an operation like that is usually broken into bands by horizon: long-range capacity planning, master production scheduling, and detailed scheduling, each a separate process, each handing off to the next.

The handoffs are where the trouble lives. Dependencies that run the length of the horizon get cut at the scope boundaries, and a question asked at one level cannot be answered honestly at another. Yet the decision is the same at every level: given everything competing for the same constrained capacity, what should be made, and when. Splitting that decision across three disconnected processes does not make it easier. It makes it inconsistent.

The work

Kinetica is building a single optimisation model that spans the whole range, from a 24-month capacity view, through the master schedule, down to the detail, without relying on handoffs or scope boundaries by horizon. Because the dependencies are modelled the whole way through, the same engine answers a detailed sequencing question and a long-range what-if, and answers them consistently. Scenarios run quickly: a change at the front of the horizon flows through to capacity two years out, and a constraint two years out is visible in next month’s schedule. A proof of concept is already in place.

The result

The proof of concept has shown that one optimisation core can carry planning from long-range capacity down to execution detail, so the organisation can ask a what-if at any horizon and get an answer the rest of the plan agrees with. The work is ongoing.

This is the heart of how Kinetica approaches scheduling. Not a stack of separate planning tools joined by handoffs, but one model of the real dependencies, queried across whatever horizon the decision happens to need.