Planning & APS
MES + APS: planning on the same reality as production
A schedule is only as good as the data it starts from. Why advanced planning works best when it shares orders, resources, materials and progress with execution.
Most schedules are wrong by mid-morning. Not because the algorithm is poor, but because the plan was computed on yesterday’s picture of the factory.
The gap between plan and floor
In many plants, planning and execution live in different systems. The planning tool imports orders and routings, computes a sequence and exports it. The shop floor works, declares, stops and restarts in another system. Every exchange is a delay, and every delay widens the gap:
- progress declared on the floor reaches the plan hours later
- a machine stop is known to the operator, not to the planner
- material arrivals and quality holds are invisible to the schedule
- the plan’s assumptions — times, setups, calendars — drift from reality
One model, two jobs
When planning runs on the same model as execution, the gap closes:
- Same orders and work stages. The plan schedules the very stages operators start and declare.
- Same resources and calendars. Machine availability, shifts and maintenance windows are shared.
- Same materials. Work is released only when stock or incoming receipts cover it.
- Same events. Progress, completions and disruptions update the planning context as they happen.
React — but decide deliberately
Closing the gap does not mean replanning automatically every minute. Shop floors need stability; constant resequencing creates its own chaos. A better pattern:
- Production events update the planning context.
- An impact analysis shows which orders are at risk.
- A draft scenario proposes a recovery: resequencing, alternative machines, different priorities.
- The planner compares it with the current plan and releases it — or not.
That is how Metalya APS works: integrated with execution, explicit about impact, and never publishing a new plan without a person deciding.