We turn the production plan into a sequence at machine and line level: tooling and set-up times, bottleneck capacity, the date material will be ready and priority rules are calculated together. The result is a schedule that can genuinely be applied on the shop floor and a defensible due date that can be given to the customer.
In most factories the production plan lives in two separate places: an order list in the ERP and a spreadsheet on the planner's desk. The list says which order will be delivered when; the spreadsheet holds which machine that order will be made on, in which sequence and with which tooling. The sequence is usually set by talking it through in the morning meeting, is broken by an urgent order that comes in during the day, and by the evening nobody remembers what the first version of the schedule was. This arrangement works in a small operation. As the product range widens, changeovers grow longer and several jobs land on the same machine, it stops working, because the knowledge that sets the sequence is in a single person's head, and when that person goes on leave the factory's planning speed drops visibly.
What is missing here is usually not planning software but which question planning answers. Material requirements planning (MRP) solves the question of 'how much and when do we need it': it calculates the requirement from the bill of materials, counts back over the lead time and produces order proposals. In its classic form, however, it assumes capacity is infinite; it sees nothing wrong in writing three jobs onto the same machine on the same day. Advanced planning and scheduling (APS) comes in at exactly this point and answers a different question: which job, on which machine, in which sequence and at what time. MRP gives the quantity and the date; APS places that date on the machine calendar. The two are not rivals; one uses the other's output as its input.
When APS builds a schedule it takes the constraints into account at the same time: the availability of the machine and the tooling, the set-up time lost in changing from one product to another, the operator and shift pattern, the date the material will enter the warehouse, the windows reserved for maintenance and the priority order of the orders. The critical point is that set-up time is not fixed; going from a light colour to a dark one does not take the same time as going from dark to light, and when two jobs in the same tooling group are run back to back the changeover almost disappears. Getting the sequence right is, in most plants, the source of the capacity gained without buying a new machine. When the schedule is built, two things are obtained at once: a sequence the shop floor can follow, and a due date that sales can give the customer and that is not imaginary.
The condition for setting this work up honestly is to discuss up front how solid the input is. APS cannot produce a better schedule than the routing, standard time and set-up time data it is given; where that data is missing, the system produces plans that look good but do not hold on the shop floor. APS also does not create capacity, does not remove the bottleneck and does not replace the planner; what it does is move the decision from intuition to calculation and make the alternatives comparable. The most common mistake in practice is rebuilding the schedule at every small change and continually sending a new sequence to the floor; the result of this is that the team stops trusting the schedule. That is why the freeze window and the rescheduling rules are defined at the start of the work.
Businesses where many different products are produced on the same machine and where the set-up, tooling change and cleaning time between products is significant. In these plants the sequence itself means capacity directly; a single badly built day produces unnecessary changeover time measured in hours, and that time never comes back.
Machine builders producing to incoming orders rather than to stock, subcontract plants and project-based manufacturers. For companies that have to give the customer a date, and that face both a penalty clause and a loss of reputation when they cannot meet the date they gave, the schedule is the only solid backing behind the sales promise. Being able to answer the due-date question with a calculation makes sales and production look at the same picture.
Plants where capacity is determined by a single press, oven, dye house or assembly line. In these structures the whole plan has to be built around that bottleneck; the bottleneck standing idle or being occupied with the wrong job is lost production directly and cannot be made up afterwards. Timing the steps before and after it accordingly is the most concrete area of gain.
Factories where only an experienced planner can build the schedule. When this knowledge is not written down, planning weakens seriously in the event of leave, illness or resignation. The first concrete benefit of the system is not speed but that priorities and rules become recorded in the company and can be handed over. The planner's job does not disappear; it is freed from the repetitive calculation load.
The schedule is built according to the fact that each machine can do only one job at a time. Machine, tooling, equipment, operator and shift availability are assessed together; planned maintenance and downtime windows take their place in the calendar. Where a job can be done on more than one machine, the alternative machines and their different speeds are defined. The result, unlike a plan list that assumes capacity is infinite, is a real sequence that can be applied on the shop floor.
How long the change from one product to another will take is not a single fixed number. Changeover times are defined as a matrix at colour, thickness, tooling or recipe-group level, and the sequence is built taking this matrix into account. Similar jobs are thereby deliberately brought together. The first version of the matrix is filled in with the knowledge of the supervisors on the floor, then corrected with actual data.
The machine or line that determines capacity is marked and the schedule is built around it. Buffer stock and timing rules are defined so that the bottleneck does not stand idle, the job in front of it is ready on time and the steps after it do not wait for it. If the location of the bottleneck changes with the product mix, the system recalculates this from period to period and warns the planner.
When a new order arrives it is placed on top of the current schedule and a realistic delivery date is calculated; the sales team gives the date by calculation rather than by estimate. If the order has to be brought forward, which jobs will move back is seen in advance. In this way the price of a promise given to one customer does not emerge later as a delay for another customer. If preferred, only a proposed date appears on the sales screen.
More than one schedule scenario can be produced for the same order pool and compared side by side: what happens if we bring the urgent order forward, what do we gain if we add a weekend shift, what does reducing tooling changes do to the delivery dates. Scenarios are saved and the reason why one was chosen is recorded. The decision is your planner's; the system only shows the result in advance.
A job cannot enter the schedule before its material and its tooling are ready. By connecting to stock, purchasing and warehouse data, the start condition of every job is checked; a job whose material will be late is flagged and an alternative sequence is proposed. That the same tooling cannot be used on two jobs at once is also defined as a constraint. A schedule produced without this check usually falls apart on the shop floor within the first half day.
The schedule does not stay on the planner's screen; it goes down to the terminal at the machine or to your existing production tracking system as a job sequence. When actual start, finish and downtime data comes back, the remaining part of the schedule is updated. The difference between plan and actual is measured; whether the source of the deviation was set-up time, a breakdown or material is seen separately.
We map the production flow on site: which job can be done on which machines, how many tools and fixtures there are, what the shift pattern is, and who sets the priority today and on what basis. How an urgent order gets slotted in is also recorded at this step. The constraints the schedule must comply with thereby stop being spoken knowledge and become written.
Product routings, standard times and the changeover time matrix are compiled; missing data is brought to a first version with the knowledge of the supervisors on the floor. This step is usually the longest part of the project and cannot be skipped. A schedule built without data will not hold on the shop floor, however good the algorithm used. How the data will be corrected later is also planned from the start.
A model is built and run for a single line or machine group rather than the whole factory. The schedule produced is compared with the schedule the planner builds by hand and the differences are examined one by one. The points where the model does not understand the shop floor correctly come out in this comparison. The scope is not widened until the differences are explained and corrected.
The system runs alongside the existing method for a while; the priority rules, the freeze window and the rescheduling frequency are tuned during this period. The aim is to find an applicable rhythm without sending a constantly changing sequence to the floor during the day. This tuning period usually takes a few weeks and is a natural part of the work.
The validated model is spread to the other lines in stages; handover is made to the planner and the production team. How the rules will be updated and how a new machine, tool or product will be introduced to the system is left in writing. After go-live, monitoring, maintenance and development support continue; as the process changes, the rules are updated with it.
The two answer different questions. Material requirements planning (MRP) works from the bill of materials to calculate how much material is needed and when, and produces order proposals. APS lines up the jobs you have according to machine, tooling and shift constraints; it says which job will start on which machine and at what time. In most set-ups MRP assumes capacity is infinite; APS does not. The right set-up runs the two as a chain: MRP gives the quantity and the date, APS places that date on the machine calendar.
It does not increase it directly, and we do not promise that. The quantity that can be produced without buying a new machine rises only as lost time falls, and that is where APS contributes: sequencing that reduces changeover times, keeping the bottleneck from standing idle, and choosing the right job from among those waiting. It is not a method for taking you above your physical capacity. The size of the gain depends on how much time you lose today in set-ups and in waiting; we measure that in discovery and do not state it as a percentage in advance.
It can be, but the largest part of the work then becomes creating this data. In practice the first version starts with a time table filled in from the estimates of the supervisors on the floor; then, as actual production data accumulates, the table is corrected. This cycle has to be planned from the start. Projects that assume the data is ready and go straight to optimisation are usually abandoned quickly, because the schedule does not hold on the shop floor. For that reason we write data preparation as a visible line item in the scope.
The purpose of the schedule is not to forbid change but to make the price of change visible. A freeze window is defined for the near future; the first shift is fixed, for example, and what comes after stays flexible. When an urgent order arrives, the system shows which orders slotting that job in pushes back and which due dates it puts at risk. The decision is still yours; the difference is that the consequence of the decision is seen in advance rather than afterwards.
No. The APS layer works alongside your existing ERP (Logo, Mikro, Netsis, Canias or another system); it reads order, bill of materials and stock data over an API or the database, and writes back the schedule it produces and the actual production. We do not try to change the ERP's own planning screen; we continue from where it leaves off. The source code, the data model and the rule set we build belong to you.
A schedule built with finite capacity at machine and line level; a changeover time matrix and written priority rules; due dates calculated for new orders; comparable what-if scenarios; a job sequence that reaches the shop floor and a plan updated with actual production; deviation reports between plan and actual. Everything, including the source code, the data model and the planning data that accumulates, is yours.
In a 30-minute discovery call we listen to what you need and tell you honestly whether custom development or an off-the-shelf product is the better answer.
We bring production, quality, maintenance and inventory operations together in a shared data model. The same stoppage, the same lot and the same shift meet in one record instead of four separate places; shift handover is put in writing, and management looks at a single set of indicators.
Details →We keep machines, plant, vehicles and equipment on a single record from the moment they are bought to the moment they are disposed of. Warranty, criticality, spare parts, maintenance cost and depreciation accumulate on the same asset card; the decision to replace stops being a guess and starts resting on the record that has built up.
Details →In companies that sell machinery and equipment, service work usually starts on the telephone and ends in a notebook. We build the whole chain in a single system — from the service request to technician assignment, from the mobile work order to the parts used and the customer's signature — and track warranty cover and SLA times automatically.
Details →We establish from one place which version of a technical drawing, CAD file, recipe or sample record is the valid one. PDM manages the file and its revisions; PLM extends that across the product's whole life, from idea to serial production, together with change approvals. Whether the drawing that reaches the shop floor is the right drawing stops being a matter for debate.
Details →We take buildings, floors, rooms and desks into a single inventory, and run lease and subscription contract dates, planned building maintenance and the inspection schedule for fire and safety equipment in the same system. The cost of running the facility thus becomes visible by building and by area.
Details →We build a human resources layer that brings employee records, documents, leave and absence, the organisation chart and employee self-service together in a single record. It does not replace payroll; it prepares the data that goes to payroll accurately and traceably, and it connects to the timekeeping and PDKS (time and attendance) side through the same employee record.
Details →
