What is APS, and what is it for?
Known as advanced planning and scheduling, finite-capacity scheduling or detailed production scheduling, this class of software aims to make the production plan workable on the shop floor, not just on paper. An APS calculates on which work centre, in what order, with which mould and operator, and on which day and shift each order will be processed.
The calculation is not a one-off. An urgent order arrives, a machine stops, material is late; APS rebuilds the plan around these changes and shows in advance which delivery dates the change affects. The final decision still rests with the planning team: APS proposes, people set the priorities and the locks.
- Can we deliver this order on the date the customer wants?
- On which resource, and when, will the bottleneck form?
- Which jobs, run one after another, shorten set-up time?
- If a machine stops, which orders are affected, and is there an alternative resource?
- If material is late, how does the plan change?
Where APS sits: ERP, APS and MES
ISA-95 (its international counterpart is IEC 62264) divides the systems of a manufacturing company into levels: ERP sits at the business planning and logistics level, MES at the manufacturing operations level. Detailed scheduling is one of the activities the standard defines for manufacturing operations management.
In practice, APS works as a bridge between these two levels: it takes the order, the routing and the material from ERP, passes a sequenced job list to the shop floor and to MES, and updates the plan with what actually happens on the shop floor. The diagram below shows this relationship; the position of APS is not a definition from the standard but MindDX’s explanation.
- ERP · business planning and logisticsOrders, bill of materials, routing, stock, purchasing and cost: what to produce and how much.
- APS · advanced planning and schedulingWhich job, on which resource, in what order and when: a schedule built on finite capacity.
- MES · manufacturing operationsRunning the work order on the shop floor; recording actual quantities, times and downtime.
- Control and shop floorPLC, SCADA, machines and sensors: where the data is born.
Collecting shop-floor data: the MES and IIoT guide
The difference between APS, MRP, ERP, MES and Excel
These systems are not rivals but consecutive steps. The confusion comes from all of them using the word «planning». The table below sets out, side by side, which question each one answers and what it does not do.
| System | Question it answers | Capacity assumption | Main output | What it does not do |
|---|---|---|---|---|
| MRP (material requirements planning) | Which material is needed, how much and when? | Infinite: capacity limits are ignored | Purchasing and production proposals | Does not balance machine and staff load |
| ERP | What was sold, what is to be produced, what does it cost? | Mostly infinite or rough-cut capacity | Records of orders, work orders, stock and finance | Does not sequence jobs to the minute and the shift |
| APS | Which job, on which resource, in what order, when? | Finite: resource capacity is never exceeded | A sequenced schedule and a delivery date that can be met | Is not a system of record; keeps no accounting or stock records |
| MES | What happened on the shop floor, and what is happening now? | Executes the plan, measures what actually happens | Actual quantities, times, downtime and scrap | Does not build the schedule itself, or builds it only to a limited extent |
| Spreadsheet and Gantt chart | How do we visualise the plan? | The assumption in the planner’s head | A plan updated by hand | Does not recalculate on a change; does not check constraints automatically |
What is finite and infinite capacity planning?
In infinite capacity planning, a job is placed on the requested date without checking whether the resource is already busy with other work that day. The result is a plan that is consistent on paper but overloaded on the shop floor; the slippage is noticed only when production jams. Classic capacity requirements planning works on this assumption: it shows the load but does not resolve it.
In finite capacity planning, capacity is a hard limit. The plan is built forward from the start, and at every step the system checks whether the resource is free. Of two orders that fall on the same machine on the same day, one is either moved to the next free slot or transferred to a suitable alternative resource, and the effect on its delivery date is visible at once. The delivery date rests on calculated load, not on a guess.
Capacity planning: glossary entry
How does APS work?
An APS is a loop with four links. Inputs come from ERP and from the planning team; the engine turns them into a schedule with rules, heuristics or optimisation; the output goes down to the shop floor as a job list; and what actually happens on the shop floor becomes the input to the next plan.
Closing the loop matters as much as the quality of the engine. An APS that gets no feedback from the shop floor rebuilds the plan without knowing what actually happened and soon loses touch with reality.
- InputsOrders and delivery dates, routings, run and set-up times, shift calendar, moulds and operators, material status.
- Scheduling engineRules and priorities, heuristics or optimisation; comparing scenarios.
- OutputsA sequenced job list per resource, a delivery date that can be met, a view of bottlenecks and load.
- Feedback from the shop floorJobs started, finished and stopped flow back into the plan through MES or the job list.
Which constraints does APS take into account?
Whether a plan is realistic depends on how faithfully the modelled constraints reflect reality. The table below lists the kinds of constraint met most often in production and why they matter; not every factory needs all of them.
| Constraint | Example | Why it matters |
|---|---|---|
| Work centre capacity | A CNC machine processes one job at a time | Prevents overloading and hidden queues |
| Alternative resource | A part can be machined on either of two machines | Relieves the bottleneck and protects the delivery date |
| Sequence-dependent set-up | Changing from a light to a dark colour, a size or a mould | Set-up time is shorter when the job sequence is right |
| Moulds and fixtures | Two presses that use the same mould | With only one mould, two jobs cannot be planned at the same time |
| Operator skills | Only a trained operator can run a certain line | A machine that looks free on paper may not run in practice |
| Material status | Raw material still to arrive from the supplier | A job without its material is not started; the plan shifts with the material |
| Shift and maintenance calendar | A line that does not run at weekends, a planned maintenance window | Takes into account that capacity changes over time |
| Subcontracted operation | Coating or heat treatment done outside | The outside time is also a step in the plan |
| Splitting and merging batches | A large order split across two machines | Adds flexibility, but can create extra set-up |
Planning levels: strategic, tactical and operational
Planning is not a single step. Over the long horizon, the question is whether capacity is sufficient; over the medium horizon, which product is produced in which period; over the short horizon, in what order the jobs run. APS is strongest at the operational level, in detailed scheduling; some software also reaches the tactical level.
| Level | Horizon | Question it answers | Typical tool |
|---|---|---|---|
| Strategic | Months and years | Is there the capacity and investment to meet demand? | Sales and operations planning |
| Tactical | Weeks and months | Which product is produced in which period, and how much? | Master production schedule and MRP |
| Operational | Days and shifts | Which job, on which resource, in what order? | Detailed scheduling with APS |
Do you need APS?
If several of the signs below apply to you, planning may have outgrown what can be carried by hand:
- you produce many products and variants on the same machines,
- set-ups and mould changes are long and costly, so the job sequence matters,
- the delivery dates you give often slip, and customer representatives keep asking production «will it make it?»,
- the plan lives in the spreadsheet of one or two people and cannot be built without them,
- squeezing in an urgent order means postponing other jobs without knowing the effect,
- the bottleneck becomes visible only when production jams.
Measure your own situation: Assess
Does every factory need APS?
No. In a business with few products, a single line, long runs of the same job or capacity far above demand, a well-built ERP plan and a simple Gantt view are often enough. The value of APS grows with the number of constraints and the frequency of change; if both are low, the return on the investment stays limited.
The right question is not «should we buy APS?» but «does our planning problem come from capacity, from data or from the process?». If the problem is a missing routing or a wrong stock record, that is fixed first.
What data does APS need?
APS builds a plan only as accurate as the data it is fed. Before it goes live, the following data must exist, be kept current and have a clear owner:
- open orders and requested delivery dates,
- bills of materials and each product’s routing: which operation, on which work centre, in what order,
- run and set-up times: values measured on the shop floor, not estimates,
- work centres, alternative resources, the shift calendar and holidays,
- additional resources such as moulds, fixtures and operator skills,
- stock and open purchase orders: when the material will arrive,
- the owner of each of these data: who updates it, and how often?
What to look for when choosing APS software
Feature lists look alike; the difference shows in how faithfully your own constraints are modelled and how the plan reacts to change. Ask the vendor to run a scenario with your own orders and routings, and ask these questions:
- Constraint model: can alternative resources, sequence-dependent set-ups, moulds, operators and subcontracting be defined?
- Replanning: on a change, is the whole plan rebuilt or only the affected part; are locked jobs kept?
- Planner control: can the proposal be compared with the current plan, and can the reasoning behind a decision be seen?
- Scenarios: can «what happens if we take this order?» be tried without disturbing the live plan?
- ERP integration: is there a documented API, webhook or file transfer; which data flows in which direction?
- Feedback from the shop floor: do actuals from MES or an operator screen flow back into the plan?
- Deployment and data: cloud or your own servers; in which region is the data held, and how is it assessed under the Turkish data protection law (KVKK)?
- Language and support: is there a Turkish interface, Turkish help and local implementation support?
How is an APS project put into use?
Successful APS projects start not with the software but with a clear planning question and clear data. The typical stages are: defining the problem and the scope, preparing the data, a pilot with a single line or product family, ERP and shop-floor integration, setting up the feedback loop, and then rollout to the whole factory.
The most common mistakes are well known too: starting with estimated times, trying to optimise a single goal such as machine utilisation alone, leaving shop-floor feedback for later, taking the planner out of the process, and not defining the measure of success before go-live.
At MindDX this path starts with Assess: the planning problem, the data and the process are reviewed, and whether APS is really needed, and with what scope, is established with evidence. Implementation runs in Implement, continuous improvement in Transform.
What does the cost of APS depend on?
There is no single price; the complexity of the factory and the scope of the project determine the cost. When comparing proposals, ask about these items separately:
- the number of resources, lines and users to be planned,
- the complexity of the constraints to be modelled,
- the route and scope of ERP and shop-floor integration,
- the effort needed to prepare and clean the data,
- the deployment model: subscription, or licence and your own servers,
- training, support after go-live and version updates.
Total cost of ownership: glossary entry
What does AI do in APS?
At its core, scheduling is an optimisation problem, and APS engines have long solved it with mathematical methods; this is not a conversational AI. AI contributes to the loop mainly in two places: estimating times more accurately from past actuals, and letting the planner read the plan faster with an assistant that answers questions.
Whatever the case, the planning team must stay in control: people set the priorities and the locks, and a proposal must be visible together with its reasoning.
MindDX and SkyPlanner
MindDX implements SkyPlanner at manufacturers whose need for APS has been proven. SkyPlanner is a cloud-based APS solution that turns orders, capacity and material into a finite-capacity production schedule; it has a Turkish interface and connects to ERP through an API, webhooks or file transfer. ERP remains the system of record.
MindDX’s role is not limited to installing the software: it runs the planning question, data preparation, ERP and shop-floor integration and the go-live within one scope, and in abas ERP and Odoo projects it connects APS to the same data model.