Skip to content
SCSupply Chain Ideas
MODULE 04 · 20 ESTIMATED HOURS

Disruption, freeze, and stability

Repair from actual events and govern changes to communicated work.

LESSON 1 · 8-HOUR ACTIVITY PLANRepair a disrupted schedule

Rebuild remaining work from a factual event without rewriting completed history.

LESSON 2 · 8-HOUR ACTIVITY PLANBalance service with schedule stability

Use an explicit change measure and exception process to compare revisions.

Watch, listen or read

Repair from actual events and govern changes to communicated work.

5:01 · Original narrated lesson slides. AI-generated narration through Runway; factual diagrams assembled for this course. Video is optional. The complete transcript is directly below.

Read the complete narration transcript

A schedule is a set of promises about future work. A breakdown changes the facts, so some of those promises may need to change. Our task is to repair the plan without losing the truth about what has already happened. Start at the event clock. Keep done work done. State what is active, what is blocked, and when the resource is expected to return. Then plan the work that remains. A fresh drawing that erases history is not a useful repair. Northstar has A from zero to sixty, B from sixty to one hundred, and C from one hundred to one hundred fifty. At sixty, A is complete and the line fails. It will be back at one hundred. This case puts the event between jobs, so there is no half made batch to scrap or restart. We can preserve the B then C order. B now ends at one hundred forty and C at one hundred ninety. Each start moves forty minutes. B is due at one hundred ten, so it is thirty minutes late. C is due at one hundred seventy, so it is twenty minutes late. Total tardiness is fifty. Total absolute start movement is eighty. Those two totals answer different questions. One records missed due time. The other records how much the communicated start plan changed. Neither can be replaced by a claim that the new plan is simply better or worse. We need to see the tradeoff. Try C before B after the line returns. C runs from one hundred to one hundred fifty. B then ends at one hundred ninety. C is on time but B is eighty minutes late. That is worse total tardiness in this case. Its movement score is also different. C stays at its old start while B moves by ninety minutes. The final finish is still one hundred ninety. A swap does not make the remaining run time disappear. Northstar also uses a frozen horizon. In our stated policy, starts strictly before one hundred twenty are protected from routine changes. That is a control rule, not a force that can keep a broken machine running. An outage needs an exception record. Name the old start, the new start, the cause, and the owner of the revised instruction. Do not treat a freeze as absolute physics, and do not treat a small score gain as automatic permission to break it. A separate case shows how stability can change a choice. The old order is A, B, C, with starts zero, fifty, and ninety. The new order is A, C, B, with starts zero, fifty for C, and eighty for B. C moves forty minutes earlier and B moves thirty later. Absolute movement is seventy. The signed sum would be minus ten, but that hides the two changes. People still need to act on both new instructions. The old plan has twenty tardy minutes and no movement. The new plan has no tardiness and seventy movement minutes. If we add one tenth of movement to tardiness, the new score is seven and wins. If we use three tenths, its score is twenty one and the old score of twenty wins. The weight is a judgment about the cost of change. Show the raw measures beside that score so the judgment stays visible to the reviewer. For the lab, keep a baseline copy, a repair copy, and a short list of changed jobs. Test the given restoration times one at a time. State which plans are physically legal and which need a policy exception. A new job has no old start in our measure, so report it on its own. End with one clear instruction and a trigger for review. Good control does not mean no change. It means that a change has a reason, an owner, and a traceable effect. The release note should also say when to check again. If the line is not back at the assumed time, the new starts are no longer supported by the facts. A useful trigger names that event in plain words. It does not need to predict every possible failure. This gives the operator a clear next step while keeping the owner of the plan in view. Save the old version, mark the new one, and make sure the team knows which set of instructions is now active.

Video teaching plan
  1. Start at the event

    Original static calculation card for start at the event, using the Northstar numbers stated in the narration; show labeled units and no decorative charts.

    • A done at60
    • Outage60-100
    • Do not erase history
  2. Keep done work fixed

    Original static timeline for keep done work fixed, using the Northstar numbers stated in the narration; show labeled units and no decorative charts.

    • A stays complete
    • B and C remain
    • No interrupted batch here
  3. Repair future work

    Original static calculation card for repair future work, using the Northstar numbers stated in the narration; show labeled units and no decorative charts.

    • B100-140
    • C140-190
    • Total tardiness50
  4. Check the freeze

    Original static timeline for check the freeze, using the Northstar numbers stated in the narration; show labeled units and no decorative charts.

    • Frozen starts need rules
    • Outage needs exception
    • Owner releases revision
  5. Measure movement

    Original static timeline for measure movement, using the Northstar numbers stated in the narration; show labeled units and no decorative charts.

    • Use absolute differences
    • Do not cancel moves
    • Show changed-job count
  6. Release one version

    Original static calculation card for release one version, using the Northstar numbers stated in the narration; show labeled units and no decorative charts.

    • Keep baseline copy
    • Name new instructions
    • Set a review trigger
PUT IT TOGETHER · 3 HOURS

Module practice & recap

At time0 the released baseline is A0-50, B50-90, C90-120; dues A60,B200,C100. All ready0, one machine0-240, no setups. Freeze protects starts strictly before60. At minute50 A finishes and an outage closes the machine50-80. Only unavoidable outage exceptions may change protected starts.

  1. Repair in baseline order B-C from the actual event and calculate tardiness plus movement from the original starts.
  2. Evaluate C-B as an alternative, identifying that B is protected and cannot move for a routine score improvement.
  3. Compare scores tardiness +0.1×movement and tardiness +0.5×movement for both physically feasible repairs.
  4. Write a versioned repair instruction preserving A history and separate the unavoidable B delay from an optional resequencing request.

What to produce

Ordinary practice answer checks
  • B-C repair B80-120,C120-150: C tardiness50, movement30+30=60; scores56 and80.
  • C-B repair C80-110,B110-150: C tardiness10, movement|80-90|+|110-50|=70; scores17 and45.
  • Both are physically feasible, but B-C is the authorized preserve-order outage repair. C-B adds a discretionary B change and needs a separately approved exception under the stated policy.
  • A stays0-50 and contributes zero movement. A good answer separates the score winner from the policy-authorized release.
CHECK YOUR UNDERSTANDING

10 quick questions

Choose an answer, then check it. Retry any question without penalty. The one-hour budget includes checking, revisiting examples, and explaining your choices.

0 of 10 answered · 0 correct

1. A completes at 60 before an outage 60-100. In repair, A should be?
2. B original start 60, repaired 100. Absolute movement?
3. B movement+30 and C movement-40 produce total absolute movement?
4. Freeze protects starts strictly before 60. A job at exactly 60 is?
5. An outage makes a frozen start impossible. Appropriate response?
6. Old plan tardiness 20 movement 0; new tardiness 0 movement 70. At alpha 0.1, new score?
7. Same plans at alpha 0.3: which score is lower?
8. A newly inserted job has no old start. In the stated stability policy it is?
9. Which action distinguishes authorized repair from best numeric score?
10. Restoration time changes from 100 to 130. What should the planner do?
Next module →