Chocolate Factory in the Rhythm of the Season | Module 4 of 9

The Changeover Sequence:
In which order the factory stands still the least

Every switch costs: cleaning from milk to dark chocolate, full clean-down after nut-containing items, start-up scrap. Today the shift lead sets the order from experience. An optimisation model finds the sequence with the fewest changeover minutes while honouring every allergen rule.

Model calculation - fictitious sample company

This module is part of a nine-part model calculation on a fictitious mid-sized chocolate manufacturer (€95 million net revenue, 140 items, 4 lines). All euro amounts are derived from the documented assumptions of the sample company. They are not results of real customers and not a performance promise. Which values are achievable on your data is shown only by a pilot.

01 The problem - many items, few lines, expensive switches

The number of possible sequences is astronomical, the number of good ones small

The sample company makes 140 items on 4 lines. Switching from light to dark needs a clean, switching from nut-containing to nut-free needs a full clean-down with release. Whoever puts nut items in the middle of the shift cleans twice.

With 900 changeovers per year of 40 minutes each, the lines stand 600 hours. Transition scrap comes on top. Choosing the order is a combinatorial problem no person solves in their head, but a solver solves in seconds - honouring delivery dates, shift limits and allergen order.

600 changeover hours per year - the quietest capacity of the factory

Plus 90 tonnes of transition scrap, partly reusable, partly lost material.

Weekly orders
Changeover matrix from MES
Sequence optimisation
Weekly plan
Fewer changeover minutes

02 The model - sequence optimisation with hard rules

Changeover matrix from actual times per product pair, allergen order as a hard constraint, delivery dates and shift limits as the frame

▸ Example values - illustration of the method, not an analysis carried out
Changeover time week: plan today 820 min -> optimised 650 min
Allergen switches with full clean-down: 6 -> 4
Changeover minutes per weekday of one line - plan today against optimised sequence (example values)
↳ The allergen rule is not a cost item

Nut-containing items belong at the end of the shift, followed by the full clean-down. The model treats this rule as a hard constraint, not as a cost term to be weighed against changeover minutes. Result: fewer switches and no clean-down missing from the daily plan. The changeover matrix comes from measured MES times, not from standard values.

03 Business impact - line hours, scrap, overtime

Three levers from one sequence

€162,000
Savings / year
−20 %
Less changeover time
+120 h
Line hours gained
Savings by lever - line hours, transition scrap, overtime
Model calculation · How the amount arises - derived in the open

Every assumption comes from the sample company and is stored centrally. With your figures the input changes, not the method.

PositionValue
Changeovers / year × mean changeover time900 × 40 min = 600 h
Lever: Less changeover time (20 %) × value per line hour120 h × €500 = €60,000
Transition scrap / year × material value90 t × €4,000
Lever: Less transition scrap (20 %)€72,000
Seasonal overtime / year€150,000
Lever: Less overtime through a steadier plan (20 %)€30,000
Result: savings / year€162,000

Assumptions of a sample company - in a real project your data replaces these values.

↳ What a gained line hour is worth

A gained line hour is only worth what can be produced and sold in it. That is why the sample company uses a value per line hour of about one third of the theoretical contribution margin per line hour - an assumption, not a calculation from your factory. In the season the hour is worth more, in summer less.

04 Next steps in your factory

From the changeover matrix to the Monday plan

① Measure the changeover matrix

Six months of MES data deliver the actual changeover times per product pair. Standard values are not enough; the deviation is often the real finding.

② Weekly plan in hindsight

For one line the last weeks are recalculated with the optimised sequence. You see the changeover minutes it would have taken against those it took.

③ Plan on Monday morning

The sequence per line from the ERP orders, ready to execute; a short-notice order is rescheduled in seconds.

All 9 modules: AI in the chocolate factory