A Gantt chart can make a production day look reassuringly orderly. Each job occupies a colored bar. Nothing overlaps on the machine. The planned completion times appear to fit. Yet a schedule can be perfectly drawn and still be a poor plan, because the chart does not automatically reveal all the consequences of the sequence it represents.
To see the difference, consider a single production line with three orders waiting at nine in the morning. Two orders belong to product family A, while the third belongs to family B. Changing between families requires twenty minutes of preparation. The orders are ready to run and cannot be interrupted.
| Order | Processing time | Promise |
|---|---|---|
| A1 | 60 minutes | Finish by 10:00 a.m. |
| A2 | 40 minutes | Flexible |
| B1 | 50 minutes | Finish by 11:15 a.m. |
There is no hidden trick in the numbers. They illustrate a familiar factory tension: grouping similar work saves setup time, but an urgent order may need to interrupt the group.
One changeover or two?
Start with A1, which must finish by ten. If we continue with A2, the line can process both A orders without switching families. A1 finishes at 10:00, A2 at 10:40, and the line spends twenty minutes changing to B before producing B1 from 11:00 to 11:50. The full workload finishes at 11:50, with one changeover. But B1 is thirty-five minutes late.
Now try A1, B1, A2. After A1 finishes at ten, the change to B takes twenty minutes. B1 runs from 10:20 to 11:10 and meets its deadline. Another twenty-minute change brings the line back to A, and A2 finishes at 12:10. There are two changeovers, and the workload finishes twenty minutes later, but both committed orders are on time.
The schedules have the same three processing jobs, on the same machine, with the same available hours. The difference comes entirely from the order and what we value.
There are two changeovers, and the workload finishes twenty minutes later, but both committed orders are on time.
| Sequence | Family changes | Last completion | B1's delivery |
|---|---|---|---|
| A1 → A2 → B1 | 1 | 11:50 a.m. | 35 minutes late |
| A1 → B1 → A2 | 2 | 12:10 p.m. | On time, 5 minutes early |
It would be misleading to declare one universally better. If meeting B1's commitment is mandatory, the first plan is unacceptable. If lateness is negotiable and setup time is unusually expensive, the decision may require an explicit tradeoff. The important thing is to make that difference visible before production begins.
The bars don't contain the whole factory
This little example assumes one machine, known processing times and a single type of setup. Actual production often involves shared technicians, tool availability, material arrivals and dependencies between operations. A job that fits into an empty machine slot may still be impossible to run because its components have not arrived. Two jobs on separate machines may compete for the same operator or inspection station.
Adding those facts changes the set of feasible schedules. So does the way changeovers behave. Some factories have sequence-dependent setups: moving from product A to B takes a different amount of time from moving B to A. In others, setup can begin before the preceding job finishes, or only certain people may perform it. The right formulation depends on the operation rather than a standard chart template.
Often the revised objective should value stability too, because changing many unaffected assignments can create operational costs that the original schedule did not need to consider.
Optimization methods help explore sequences, assignments and start times together. In a job-shop model, a job may pass through several machines in a required order, while each machine handles only one operation at a time. Constraints express those relationships; an objective determines which feasible plans are preferred. Depending on the application, an algorithm might emphasize lateness, total completion time, changeovers or another measure.
A schedule begins to change as soon as it is used
By ten-thirty, new information may have arrived. A machine fault delays B1. An urgent order is inserted. Material due at noon turns up early. The planner needs more than a fresh picture; they need a new decision about what can move and what should stay fixed.
This is where reoptimization enters. The model takes account of work already completed or started, new resource availability and commitments made to people and customers. It then searches for an updated plan. Often the revised objective should value stability too, because changing many unaffected assignments can create operational costs that the original schedule did not need to consider.
A Gantt chart remains an excellent way to communicate the result. It lets operators and planners understand the proposed sequence, question assumptions and make adjustments. But the chart is the surface of a deeper decision: how to use scarce resources while satisfying connected requirements.
The goal is not simply to color every empty hour. It is to find a plan that can be run, and to understand what will happen when the plan needs to change.