Production scheduling
Due dates, resources, and setup changes.
Build a plan that can run.
We model orders, process sequence, resource capability, material release, worker skills, and downtime together so planners can review and adjust a feasible schedule candidate. Late and unscheduled work stays visible.
- Orders, operations, resources, workers, and material release times are separated.
- Precedence, downtime, setup changes, and due-date lateness are checked together.
- Resource Gantt, unscheduled reasons, and order completion status.
Production plan / today
1 review item08:0010:0012:0014:0016:00
MC-AA101C330E512
CNC latheB204D410
InspectionA101B204C330
Inputs Orders, operations, resources, workers, and material release times are separated.Calculation Precedence, downtime, setup changes, and due-date lateness are checked together.Output Resource Gantt, unscheduled reasons, and order completion status.
Problem
Filling empty machine time is not the same as meeting due dates.
Production work has sequence, resources have different capabilities, and materials, staffing, setup, and downtime interact. One change can ripple into downstream operations.
Process sequence and due dates collide
A downstream step cannot start before its predecessor finishes, and rush work can delay many following orders.
- Precedence and waiting time
- Due dates and priority by order
Not every resource can process every job
Capability, tooling, size, certification, and worker skill restrict the assignable resources.
- Primary and alternate resources
- Worker skill and parallel load
Materials and setup are easy to miss
Ignoring material arrival, fixtures, cleaning, color changes, or family setup creates a neat but infeasible chart.
- Material release time
- Setup time by sequence
Breakdowns and rush work force replanning
When a machine fails or material is late, the remaining work must be rebuilt while started work stays fixed.
- Keep started work fixed
- Show impact and delay reasons
We turn the judgement planners make every day into explicit conditions and metrics.
Orders×Operations×Resources×Workers×Materials= executable schedule
Model the decision
Design business rules in three layers.
If every rule has the same strength, the model may find no candidate or produce a schedule people cannot use. We separate hard rules, preferences, and improvement metrics.
01
Hard
Rules that must hold
- PrecedenceStart each operation only after its predecessor is complete.
- Resource and skill fitOnly compatible resources and qualified workers are candidates.
- Materials, downtime, and shiftsRespect material release, maintenance stops, and work windows.
02
Soft
Preferences to preserve
- Fewer setup changesGroup similar families when it does not break harder rules.
- Primary resource and continuityPrefer familiar resources and fixed started work.
- Small changesAvoid moving work that has started or already been communicated.
03
Objective
Metrics to improve
- Due-date latenessReduce late orders and total late minutes.
- Completion time and overtimeExpose expected finish time and overtime.
- BottlenecksShow concentrated load and idle capacity by resource.
∴
We do not optimize only for the shortest time.
A good plan is not just 100% machine utilization.
Where it fits
For sites that coordinate many orders, operations, and resources every day.
The key is not the industry label. It is whether operations have sequence, resources are shared, and change is frequent.
High-mix machining
Cutting, turning, grinding, heat treatment, and inspection across different resources by order.
Alternate resourcesJigsSetupOutsourcing
Food packaging
Batch sequence with raw materials, shelf life, allergens, cleaning, heating, cooling, and packing changes.
PartsSkillsInspectionRework
Assembly and inspection
Plan around part arrival, assembly sequence, test benches, skilled workers, and rework.
CleaningAllergensTemperaturePacking
Painting, chemical, and continuous processes
Sequence colors, tanks, ovens, drying time, batch merging, and quality requirements.
Color changeBatchFurnaceQuality
Interactive example
Change the conditions and watch the production plan move.
This is a simplified explanatory calculation. Real projects include calendars, setup matrices, workers, materials, outsourcing, and ERP or MES integration.
Resource Gantt
Resource Gantt chart
This candidate is grouped by resource. Selecting, fixing, and recalculating operations is added in production design.
Review
Review points
Resource plans
Operation sequence by resource
Order status
Expected completion by order
Unscheduled
Unscheduled operations
Rules
Production scheduling separates hard rules from preferences.
The public demo exposes representative rules only. Production systems are designed around your route sheets, standard times, workers, materials, and outsourcing rules.
R-01HardPrecedence
Start each operation after the previous operation completes.
R-02HardResource capability and skill
Assign only compatible resources and qualified workers.
R-03HardMaterial release time
Prevent starts before materials are available.
R-04PreferenceSetup reduction
Group similar product families to reduce changeover time.
R-05PreferenceFreeze started work
Avoid moving started operations or promised dates.
R-06MetricDue-date lateness
Compare late order count and late minutes.
R-07MetricCompletion time
Review makespan and overtime.
R-08MetricResource load
Show bottleneck resources and idle capacity.
R-09MetricUnscheduled reasons
Expose why operations do not fit.
Output
Show the reasons behind the Gantt chart.
A useful schedule cannot be a black box. People need to review candidates, freeze work, compare impact, and recalculate.
Resource order and timing
Show start, setup, processing, inspection, downtime, and workers by resource.
Late and unscheduled reasons
Classify resource shortage, material delay, missing skill, downtime, and impossible due dates.
Candidate comparison
Compare lateness, completion, setup, overtime, utilization, and change count.
Manual fixing and partial replanning
Freeze started operations or promised due dates and recalculate only remaining work.
1Create candidatesCalculate from orders, resources, and materials
→2Planner reviewShop-floor context, quality, outsourcing
→3Fix operationsKeep started and promised work stable
→4RecalculateRebuild only remaining open work
From model to operations
Go beyond calculation and connect orders to actuals.
A scheduler alone does not change operations. We design the screens needed for orders, operations, resources, instructions, progress, quality, and actual results as one system.
- Order and work-order ledgerManage item, quantity, due date, priority, and operation structure
- Planning and Gantt screenSupport assignment, manual edits, freezing, and partial recalculation
- Shop-floor actual inputRecord starts, completions, stops, defects, and material use
- Progress and delay analysisAggregate plan versus actuals, bottlenecks, load, and due-date responses
SHOP FLOORActual input screenStart, finish, stop, quality
PLANNINGPlanning GanttEdit, freeze, due-date response
MATHEMATICAL COREScheduling engineSequence, resource, setup, recalculation
DATAOrders, operations, resources, actualsDB, API, audit log
Data
Check whether scheduling can work before production development.
We can start with the current schedule and business rules, build a calculation prototype, and avoid assuming a large MES or ERP replacement from the start.
INPUT
Data to check first
Order and work-order list
Item, quantity, due date, priority, and material release time.
Operations and standard time
Precedence, processing time, setup, alternate resources, and lot rules.
Resource and worker calendar
Capability, work hours, downtime, skill, and simultaneous limits.
Current plan and actuals
Starts, finishes, stops, defects, actual time, manual changes, and delay reasons.
Related demos
Connect production planning to operational screens.
We publish demos for manufacturing process management, work-order tracking, production history, and receiving inspection. The calculation on this page is explanatory and is not presented as an optimization case study.
View manufacturing system demos →Fit check
When custom development fits and when it does not.
We first check whether operational complexity, change frequency, and improvement room justify a custom system.
✓
Good fit
- Many rules such as precedence, alternates, and setup
- Planning depends on expert knowledge
- Failure, rush work, and material delay cause frequent replanning
- Packaged tools still leave heavy spreadsheet adjustment
- You want to connect orders, shop-floor actuals, and quality
—
Clean up first
- Products and operations are few and fixed
- Resource conflicts and due-date adjustment are rare
- Packaged production management functions are enough
- Standard time, operation, and resource data are not organized
- Actual collection should come before automatic planning
Delivery process
Solve a small case, then embed it in production management.
We separate calculation validation from operational system development so each step can continue, change, or stop deliberately.
- 01DIAGNOSISFree check
Review issue and data
Check current schedules, orders, resources, materials, and change frequency.
- 02MODELINGDesign
Separate three rule types
Organize hard rules, preferences, and metrics, then agree on evaluation.
- 03PROTOTYPEFixed scope
Small calculation validation
Create candidates for a day or week and compare with the current plan.
- 04PRODUCTEstimate
Build the operational system
Connect orders, planning, actuals, quality, due-date response, and analysis.
- 05IMPROVEContinuous
Improve from actuals
Use actual time, stops, manual edits, and delay reasons to refine the model.
FAQ
Frequently asked questions
These are common checks before a first conversation.
Send another question →Can we start if standard times are rough?
Yes. We can begin with standard times, actual medians, or planner estimates and test sensitivity.
Can one operation use alternate resources?
Yes. Processing time, quality conditions, cost, worker skill, and tooling can differ by resource.
Can setup time depend on product sequence?
Yes. Colors, materials, molds, allergens, and other pairwise setup matrices can be modeled.
Can the system recalculate after failure or rush work?
Yes. Started work and promised dates can be fixed while remaining operations are replanned.
Can people manually change the generated schedule?
Yes. The design assumes people can move operations, fix resources, change sequence, and see impact.
Can it always meet every due date?
No. If capacity, material, and processing time make all due dates impossible, the system should show the cause and relaxation options.
Is production scheduling created by generative AI?
The core is mathematical optimization, constraint satisfaction, and search. AI may help organize rule candidates, but feasibility conditions remain explicit.
Can it connect to an existing ERP, production system, or MES?
Yes. API, CSV, database, and file integration can be considered, including adding only the planning engine and Gantt view.
How many orders and operations can it calculate?
Difficulty depends on alternates, setup matrices, workers, response time, and required solution quality, so we measure with representative sample data.
Next step
Your production schedule may be possible to model.
With anonymized schedule and order data, we can check whether the conditions can be represented and whether a mathematical prototype fits.
- Anonymized data is enough
- We say when packaged tools fit better
- No forced sales process