Load Leveling Across a Planning Week

Load leveling is the practice of spreading work across the days of a planning week so no single day is overloaded while others sit half-empty. Orders tend to arrive in bursts, and the naive plan piles them all onto the first available day, creating a brutal Monday and an idle Friday. Leveling smooths that profile so each day carries a load the resources can actually finish. A finite scheduler does most of this for you automatically: when a day is full, it searches forward for the next day with room, so peaks spill into valleys as a byproduct of refusing to overbook. EDGEBIC by User Solutions levels load this way and gives you the controls to smooth what remains.

Total capacity is not the same as a smooth week

A week can have enough total capacity and still be miserable to run. If Monday carries 60 hours of load against 40 hours of capacity while Friday carries 10, the week's total might balance perfectly and Monday still blows up. That is the distinction between capacity planning and load leveling: capacity planning asks whether there are enough hours overall, and load leveling asks whether those hours are spread sensibly across the days.

The reason the distinction matters is that an overloaded day does not just delay that day's work; it pushes the overflow downstream and can cascade lateness through the whole week. Meanwhile the idle day represents capacity you paid for and never used. Leveling recovers both: it moves the overflow into the empty space so the peak comes down and the valley fills.

How a finite engine levels load automatically

The good news is that finite scheduling levels load as a natural consequence of how it places work. The engine places each operation against real available capacity, and when a day is full, it does not cram more in. It searches forward for the earliest future day that has room, described in why a scheduler searches forward in time for capacity, and lands the operation there.

That forward search is load leveling in action. Work that will not fit Monday lands on the earliest day that can take it, so Monday's peak spills into Tuesday and Wednesday until each day carries only what it can finish. You do not hand-balance the week; the finite constraint does it by refusing to overbook any day. This is also why an infinite-capacity plan never levels: it just stacks everything on Monday and reports an impossible profile.

A worked example

A work center has 40 hours of capacity per day. Monday morning, five jobs arrive, each needing 16 hours, so 80 hours of demand hitting a 40-hour day.

Infinite view: all 80 hours land Monday. The plan shows Monday at 200 percent load, which the floor cannot run, and the dates are fiction.

Finite view with forward search: the engine fills Monday to 40 hours (two and a half jobs' worth), then spills the rest forward. The remaining 40 hours land Tuesday. The load profile becomes 40 hours Monday, 40 Tuesday, and the week is level at capacity instead of spiked at 200 percent Monday and empty after. Every date is real, and no day is asked to do more than 40 hours.

Now suppose Wednesday also gets a 50-hour demand while Tuesday still has 40 hours free after the spill. The engine can only put 40 on Wednesday and pushes 10 to Thursday, but if you want Wednesday smoother, you set the start date on one flexible job to Tuesday so it becomes eligible a day earlier and fills Tuesday's remaining room. The peak on Wednesday drops from 50 to 40, and Tuesday goes from 40 to 40 plus the moved job, spread across the two days.

The controls you have on top of the automatic leveling

The forward search prevents overbooking, but you can shape the profile further with three levers.

Start dates decide when a job becomes eligible. Staggering start dates across the week spreads eligibility so work does not all compete for Monday, which is the single most effective way to pre-level a lumpy week.

Priority decides which eligible jobs claim a full day's capacity first. When Monday can only take some of the demand, priority chooses which jobs get Monday and which spill forward.

Alternate machines and shifts add capacity to a peak day. Routing a flexible job to an alternate work center or adding a shift on the peak day lets you carry more on the day that is spiking, instead of only spilling it forward. Multi-shift capacity and how the engine fills it is covered in how multi-shift allocation fills capacity.

Reading the load profile

The habit that makes leveling work is looking at the daily load profile, not just the weekly total. A weekly total of 200 hours against 200 hours of capacity looks perfect and can still hide a 60-hour Monday. The two views answer different questions, which is the point of how a scheduler thinks about a week vs a day: the near term is executable and the far term is directional. The utilization view by day shows you where the peaks are, and once you can see them, the levers above bring them down.

Level loading is not about making every day identical to the decimal. It is about making sure no day is asked to do more than it can, so the plan the engine produces is one the floor can actually run, with no hidden cliffs. The finite engine gives you a level plan by default; your job is to spot the residual peaks and smooth them with start dates, priority, and capacity. See the daily load profile on your own work centers in EDGEBIC, and read how the engine places work across time in the scheduling engine guide.