What Is Longest Processing Time (LPT) Scheduling?

Longest processing time, abbreviated LPT, is a priority rule that runs the job with the most work content first and the shortest job last. It is the scheduling equivalent of loading the biggest boxes onto the truck first so the small ones fill the gaps around them. LPT is used to balance load across parallel machines and to keep a large job from being stranded at the end of a shift, and it is one of the ordering heuristics EDGEBIC can apply when sequencing a queue.

This entry is part of the EDGEBIC by User Solutions glossary series; for the broader vocabulary of planning, see the manufacturing glossary.

How Longest Processing Time Works

LPT sorts the waiting jobs by total work content, largest first, and runs them in that order. It is the mirror image of shortest processing time. Where SPT minimizes average flow time by clearing quick jobs, LPT targets a different problem: how to spread heavy work so the resources finish at roughly the same time.

The classic use is parallel machines. When several interchangeable machines can each take the next job, assigning the biggest jobs first tends to even out the total finish time, the makespan. Small jobs are flexible and can slot into whichever machine has room, but a giant job needs a clear block, and finding that block is far easier at the start of the horizon than after the schedule is already crowded. Start with small jobs and you risk painting yourself into a corner where the big one no longer fits cleanly.

The cost is average wait. Running the longest job first makes every job behind it wait the full long duration, so LPT inflates average flow time and work in process, exactly the opposite of SPT. LPT also ignores due dates. It is a load-balancing and anchoring rule, not a queue-clearing one.

A Concrete Example

Three identical mills share five jobs with these work contents:

LPT assigns the largest jobs first, one per mill: A to Mill 1, B to Mill 2, C to Mill 3. The remaining jobs fill the lightest-loaded mill: Mill 3, at 4 hours, takes Job D to reach 7, then Mill 2 and Mill 3 balance the last small job. The mills finish within a couple of hours of each other. Had the schedule started with the two-hour jobs, the nine-hour Job A could have landed on an already-busy mill and pushed the overall finish out. Placing the heavy jobs first is what keeps the machines even.

How EDGEBIC Uses It

EDGEBIC exposes longest processing time as one of the ordering heuristics inside its multi-run schedule optimizer, described as longest total work content first. It sits alongside its opposite, shortest processing time, plus earliest due date, critical ratio, priority-then-due-date, and seeded shuffles.

When the optimizer runs, each candidate ordering is fed through the unmodified scheduling engine to produce a full finite-capacity schedule. Every schedule is scored against your goals and the best is kept, with a never-worse clamp guaranteeing the result is never inferior to the engine's default greedy plan.

LPT therefore competes rather than commands. The engine adopts the LPT sequence only when the complete schedule it produces actually scores best, which is most likely when you have interchangeable machines to balance or a large job to anchor early. Because ordering and capacity interact, the comparison of full schedules is what decides, not a preference for big jobs. When your work centers are pooled, load balancing across members becomes especially relevant; see the work center groups guide and the optimizer goals and presets guide.

Related Reading

LPT is the opposite of shortest processing time, and it is often weighed against earliest due date when deadlines matter. The generic priority rules glossary covers the wider dispatching family, and job shop scheduling challenges shows why sequencing choices matter under real capacity.