When Overtime Beats Resequencing (and When It Does Not)

Overtime beats resequencing when the total work exceeds the total capacity in the window, and resequencing beats overtime when the capacity is there but the job order is wasting it. The two moves solve different problems. Resequencing changes which jobs are late; it cannot change whether the work fits. Overtime creates hours you do not have; it is wasted money if the hours already exist and are just poorly ordered. EDGEBIC by User Solutions shows you both the demanded hours and the available hours, so you can tell a sequencing problem from a capacity problem before you authorize a single Saturday shift.

The rule of thumb is simple: resequencing is free, so try it first, and reach for overtime only for the shortfall that survives the best possible sequence.

The one comparison that decides it

Every late-job crisis is one of two things, and one number tells you which. Add up the run and setup hours of the jobs competing for a resource in the window. Compare that to the hours the resource actually offers in the same window, at your planned schedule-at-utilization. If demanded hours fit inside available hours, you have a sequencing problem. If demanded hours exceed available hours, you have a capacity problem.

Sequencing problems are free to fix: reorder the jobs and the dates recover. Capacity problems cannot be fixed by reordering at all; reordering just moves the lateness from one job to another. That is the entire decision, and a finite capacity schedule hands you both numbers instantly.

A worked example: same three jobs, two diagnoses

One work center, one instance, an 8-hour day shift planned at 80 percent, so 6.4 hours a day. Three jobs, all due end of Friday, released Monday:

Total demand is 18 hours. Monday to Friday at 6.4 hours a day is 32 hours available. The work fits, so any lateness here is a sequencing artifact. If the schedule shows a slip, reorder (put the shortest or most-at-risk job first) and the dates recover with zero overtime. This is a sequencing problem.

Now change MO-A to 20 hours and MO-C to 14 hours. Total demand is 38 hours, but you only have until Friday. If the deadline is Wednesday end of day, that is 3 days times 6.4, or 19.2 hours available, against 38 demanded. No sequence fits 38 hours into 19.2. This is a capacity problem, and the only real fixes are overtime, a second shift, or offloading work to an alternate work center. Reordering would just choose which jobs miss.

Try the free move first

Because resequencing costs nothing, it is always the first move. The manual version is reordering by priority or start date and re-running. The powered version is the optimizer, which searches many job orderings, scores each complete schedule against your goal, and proposes the best one. The multi-run layer is guaranteed never worse than your current plan, so running it is risk-free: at worst it hands back the schedule you already had. If the best sequence it finds still leaves a job late, that is a genuinely useful result, because it proves the problem is capacity, not order. How that search works is covered in how mathematical optimization improves a schedule.

Resequencing also includes smarter moves than raw reordering. Grouping jobs with similar setups cuts changeover time and can free enough hours to close a small shortfall without overtime, which is the whole point of campaign sequencing to minimize changeovers. Sometimes the "missing" hours were being burned in avoidable setups.

When overtime is the honest answer

Once the best sequence still shows a shortfall, overtime stops being a firefight and becomes a clean decision. You know exactly how many hours you are short and exactly which resource is short, because the schedule points at it. Add the hours there, whether that is a Saturday shift or an extended day, and re-run to confirm the late job now lands on time. This is also how you recover from a mid-shift disruption where a breakdown eats capacity you had already committed, covered in how a scheduler recovers from a mid-shift disruption.

Overtime is capacity, and adding it means editing the resource's available time (an extra shift, a longer day, or a working weekend) so the finite engine has more hours to place work into. The mechanics of adding and removing available time sit alongside how a schedule accounts for holidays and downtime, since a working Saturday is just the inverse of a holiday.

The order of operations

Diagnose, then resequence, then buy hours for what remains. Compare demanded hours to available hours to name the problem. If it is sequencing, reorder or run the optimizer, for free. If a shortfall survives the best sequence, it is real capacity, and overtime spent on the resource the schedule identifies is money well spent. Doing it in that order means you never pay for hours you already had, and you never resequence forever against a wall that only more hours can move.

This diagnostic discipline is the same one behind decades of User Solutions schedules, from job shops quoting realistic dates to GE Railcar's climb from 30 to 90 percent on-time delivery: know whether you are short of order or short of hours before you act. The full engine is laid out in the scheduling engine guide, and you can compare demanded against available on your own load in EDGEBIC.