Skip to content

Operation time: queue, setup, production and move

Kinetic + Classic

Each operation on a job takes up a span of the schedule. Only part of that span is work on a resource; the rest is waiting and moving. Understanding the four components explains why an operation that takes two hours of machine time can occupy most of a day on the schedule.

operation time = queue time + setup time + production time + move time
Component What it represents Defined on Loads the resource?
Queue Time parts wait at the resource before work begins Resource group or resource (Queue Hours) No
Setup Preparing the machine or station The operation in the method Yes
Production Making the quantity, from the production standard The operation in the method Yes
Move Time to get finished parts to the next operation (transport, cooling, curing) Resource group or resource (Move Hours) No

Production time is the quantity divided by the production standard. For example, 100 pieces at 10 per hour is 10 hours. Add a 30-minute setup, one hour of queue and 15 minutes of move, and the operation spans 11 hours 45 minutes of schedule while loading the resource for only 10.5.

Resource Group Maintenance showing the Queue Hours and Move Hours fields in the Scheduling group

  • Queue sits in front of the operation, move sits after it. The schedule shows them as part of the operation’s overall span, not as separate bars, so it isn’t obvious from a Gantt view how much of an operation is padding.
  • Because they don’t consume capacity, queue and move are the right tool for time that passes without work: waiting for an inspector, paint drying, a trip to another building.
  • Adding some queue or move time also builds slack into the schedule for the unexpected. Too much makes every lead time look longer than it is.
  • By default, queue and move are counted in elapsed hours. Select Use Calendar for Queue Time or Use Calendar for Move Time on the resource group to count only working hours on its calendar. Only select these when the matching hours are greater than zero; the combination of calendar-based queue/move with zero hours is a known cause of scheduling loops.
  • Queue and move apply to every operation that uses the resource group. If only some operations need the delay, create a separate resource group for them.

Engineering, programming or tooling often has to happen weeks before production but takes few actual hours. Modelling it as a normal operation schedules it just before machining starts, which is too late.

A simple fix is a planning resource group with a long queue time:

  1. Create a resource group for the planning work (for example Engineering).
  2. Set its Queue Hours to the lead time you need. For a 30-day lead time counted in elapsed hours, that’s 720.
  3. Put the planning operation first in the routing, assigned to that group.

Backward scheduling then reserves the lead time in front of production without loading the engineers’ calendar with 720 hours of work. For other ways to model this work, see Tooling, engineering and programming work.

Parts also have a Receive Time (days), set in part planning. The scheduler subtracts it from the job’s required date before back-scheduling, to leave time for putting finished parts away or moving them to the next job. The same idea works for purchased parts: receive time can cover inspection and put-away after a PO arrives. See Purchase orders.