DE102021203400A1 - Computer-implemented method and computer program for assembly quantity planning of assembly parts for production optimization of a production system, assembly quantity planning system and production planning and control system - Google Patents

Abstract

Computer-implemented method for planning the number of assembly parts (m) for production optimization of a production system (F) comprising the steps: Obtaining input data comprising optimality criteria (P1) comprising deviations of the production system (F) from specified range targets and/or production specifications for the assembly parts (m) assembleable products (O1); Optimizing the deviations from the specified range targets in terms of quantities of assembly parts (m), based on the input data while complying with restrictions (C) of the production system (F) (O2); Outputting information about which specifications cannot be met if the optimization cannot be met (O3); otherwise, providing planning-relevant outputs including the number of assembly parts (m), which optimize the deviations from the specified range targets (O4).

Description

The invention relates to a computer-implemented method and a computer program for planning the number of assembly parts for production optimization of a production system. Furthermore, the invention relates to an assembly quantity planning system and production planning and control system.

During production, products, including material goods and services, are created based on production factors, including materials and equipment. For example, gears are produced. Additional material goods are created within the gearbox production, for example output shafts. Production planning and control, including assembly quantity planning, also called assembly requirements planning, and sequence planning and control, also called sequencing, optimizes the entire production system, for example a factory.

Currently, a human controller plans the production processes of a specific product, workpiece or semi-finished part, such as the production of the output shaft, with or without the support of known optimization methods. The production consists of several production stages that the part has to go through one after the other. To do this, the controller must take a large number of input variables into account. The planning, for example which parts are to be produced on which line/part line at what time, should be optimal with regard to a large number of optimality criteria. To make matters worse, on the one hand, the parameters that determine the existing production processes frequently change over time, as do the optimality criteria. This results in the need for frequent replanning, which must be done as quickly as possible, so that production does not stand still or produce suboptimal.

The German patent applications with the file numbers 10 2020 203 716.7 and 10 2020 203 718.3 each disclose an adaptive production optimization through which production sequences, worker assignments and supplier orders are created from given requirements and production sequences are evaluated and optimized using given criteria. This adaptive production optimization supports the controller or completely takes over the control of the production processes.

Currently, a human assembly planner has the complex task of planning the assembly quantity planning and assembly sequence of a product consisting of variant sequences and number over several days. This so-called assembly requirements planning precedes the control of the production processes. This planning should be optimal in terms of a variety of optimality criteria. Thus, for this planning, a variety of conditions and constraints must be considered, such as coverage to the customer, factory capacity, and material flow, that is, the flow of parts needed to produce the assembly sequences. The latter is made more difficult by the fact that there is both a factory-internal material flow from delivered semi-finished parts to factory-internal production steps to the end product and an external material flow. There are different restrictions for both types of material flow, such as maximum total delivery quantity, maximum quantity of a single delivery variant and/or a mixture of these.

For example, more than 10,000 gearboxes are assembled in a factory every day. There are over 1000 possible hardware variants. There are transmission components that are manufactured in the factory, for example mechatronic components that come from suppliers, for example converters, and a mixture of these, for example output shaft. In order to keep this material flow stable, the number of gearboxes at the variant level must be fixed for a period of time, depending on the delivery lead, which can be several days, for example. At the same time, the assembly planner who plans the assembly of gearboxes must be able to react quickly to events such as late deliveries, stop messages and/or customer adjustments.

The assembly requirements, in the above example the number of gearboxes that are built in one day, is determined from the technical capacity and the real capacity derived from it. The real capacity is usually lower than the technical capacity due to the availability of employees and possible disturbances. The assembly requirement is divided among the customers/goods recipients based on a range of coverage strategy. The range strategy specifies the target range of gearbox inventory to be kept in the warehouse. For example, a range of 5 days means that the customer requirements for the next 5 days can be covered by the available stock.

Currently, the assembly requirements at assembly level, on which several variants are summarized by generation, size and/or customer, are planned for the next 4 weeks under the known material bottlenecks. However, this plan is too imprecise. There are customers for whom an assembly consists of many different variants, which also have very different required quantities. The calculated range via the assembly is therefore not very meaningful: It can happen that the range on assemblies level corresponds exactly to the goal, but individual variants are not covered. The variants within an assembly also differ in bottleneck parts or critical resources, i.e. parts for which there is a restriction in terms of delivery, which is why an exact bottleneck part analysis is not possible only at assembly level.

The object of the invention was to improve the planning of assembly requirements with regard to the problems mentioned above and to achieve production optimization.

The objects of claims 1, 12, 13 and 15 each solve this problem.

• • Erhalten von Eingabedaten umfassend Optimalitätskriterien umfassend Abweichungen des Produktionssystems zu vorgegebenen Reichweitenzielen und/oder Produktionsvorgaben für aus den Montageteilen montierbare Produkte;
• • Optimieren der Abweichungen zu den vorgegebenen Reichweitenziele hinsichtlich Stückzahlen der Montageteile, basierend auf den Eingabedaten unter Einhaltung von Einschränkungen des Produktionssystems;
• • Ausgeben einer Information, welche Vorgaben nicht erfüllbar sind, falls die Optimierung nicht erfüllbar ist;
• • anderenfalls Bereitstellen von planungsrelevanten Ausgaben umfassend die Stückzahlen der Montageteile, welche die Abweichungen zu den vorgegebenen Reichweitenzielen optimieren.

According to one aspect, the invention provides a computer-implemented method for assembly quantity planning of assembly parts for production optimization of a production system. The procedure includes the steps

• • Obtaining input data including optimality criteria including deviations of the production system from specified range targets and/or production specifications for products that can be assembled from the assembly parts;
• • Optimizing the deviations from the specified range targets in terms of quantities of the assembly parts, based on the input data while adhering to the limitations of the production system;
• • Outputting information about which specifications cannot be met if the optimization cannot be met;
• • Otherwise, providing planning-relevant outputs, including the quantities of the assembly parts, which optimize the deviations from the specified range targets.

According to a further aspect, the invention provides an assembly quantity planning system which is designed to carry out a method according to the invention.

According to a further aspect, the invention provides a production planning and control system which comprises an assembly quantity planning system according to the invention and is designed to carry out a method according to the invention.

According to a further aspect, the invention provides a computer program for planning the assembly quantity of assembly parts for production optimization of a production system. The computer program includes instructions that cause a computer to carry out a method according to the invention when the computer program runs on the computer.

Advantageous refinements of the invention result from the definitions, the dependent claims, the drawings and the description of preferred exemplary embodiments.

The assembly quantity planning is also called Assembly Quantity Planning, abbreviated AQP.

The input data is the data relevant for assembly requirements planning. The data is entered manually, for example via an input module, or received automatically, for example by means of NFC, RFID or Bluetooth radio technologies. According to one aspect of the invention, the input data are obtained each time the method is carried out, ie each time the optimization is run.

According to one aspect of the invention, an optimization run is triggered each time the production parameters have changed. The invention thus provides adaptive production optimization.

Another cause is, for example, a change in the weighting of the different optimization criteria by the assembly planner. But every other change in the input conditions usually leads to a new start of assembly requirement planning. This enables optimal planning of the assembly sequences to be implemented.

The method according to the invention supports the assembly planner at least in the assembly requirements planning and subsequent sequencing to determine the sequence. According to one aspect of the invention, the method according to the invention carries out the assembly requirements planning and the sequencing in an automated or autonomous manner.

According to one aspect of the invention, an overall cost function of the production system is minimized in order to optimize the deviations from the specified range targets. According to one aspect of the invention, the total cost function includes costs for deviation from the specified range per goods recipient and material. The weighting of the deviation, i.e. how much the deviation is included in the costs, can be specified in intervals, for example cost factor 1 for deviation up to 1 day and cost factor 10 for deviation of more than 1 day. According to a further aspect of the invention, the total cost function includes costs for deviations from the production specification. According to a further aspect of the invention, the total cost function includes costs for Deviations from previous planning. For a smooth production process, it is important to keep the sum of the bottleneck parts and critical resources constant over the short term, since internal and external suppliers cannot implement short-term changes. The invention proposes minimizing medium-term deviations. According to another aspect of the invention, the total cost function includes costs from other production quality functions, such as setup costs, lot sizes, and leveling. According to a further aspect of the invention, the total cost function comprises all of the aforementioned components or a combination of a selection of specific ones of these components.

Since there can be multiple optimal cost solutions, one aspect of the invention returns all of those solutions. According to another aspect of the invention, the N best solutions are returned for a parameter N, for example N=10.

• • Menge von zu bauendem Material pro Warenempfänger für alle in der Eingabe festgelegten Zeiträume, beispielsweise Stunde, Schicht, Tag;
• • Menge und Umfang von Lieferungen für Teile, die zur Produktion verwendet werden, umfassend interne und externe Lieferungen;
• • Zuordnung von fertig gebautem Bestand zu Warenempfänger.

The planning-relevant outputs provided include the data that the production system needs to regulate and/or control the production process and that are further processed in the sequencing. According to one aspect of the invention, the planning-relevant output

• • Amount of material to be built per ship-to party for all time periods defined in the input, eg hour, shift, day;
• • Quantity and scope of supplies for parts used in production, including internal and external supplies;
• • Allocation of completed stock to ship-to party.

If there is no solution, according to the invention, a reason as to which specifications cannot be met together is returned to the assembly planner. The latter can then rectify the problem and restart the method according to the invention, the planning system for the number of assembly parts according to the invention or the production planning and control system according to the invention with changed parameters.

As a result, the invention provides all planning-relevant information, such as optimal assembly planning with regard to the given total cost function, delivery quantities and assignment of stock to goods recipient, and regulation and/or control signals, which the assembly planner and/or the assembly quantity planning system or the production planning and - control system regulate and/or control the production process. For example, the assembly planner uses this result to change the entries, or he releases the planning for implementation in the factory.

The invention enables production planning and production regulation and/or control under complex production conditions that change over a horizon based on the assembly quantity planning according to the invention. For example, according to the invention, long-term stable material flows can be planned with a flexible and rapid reaction to production changes at the same time. The advantage of the inventive output of reasons for negative satisfiability checks is a significant improvement over known optimization methods based on constraint optimization, since incorrect or inconsistent inputs are rejected with a detailed reason.

The assembly quantity planning system and the production planning and control system include hardware and/or software modules for controlling the production system, for example a factory. The hardware modules include multiprocessor systems-on-chip, central processors and/or graphics processor-based hardware accelerators, memory units, and connectivity elements. According to one aspect of the invention, a production system, for example a factory, includes a private cloud infrastructure, a public cloud infrastructure or a hybrid cloud infrastructure. The assembly quantity planning system and the production planning and control system are each executed as functional software in one of these cloud infrastructures. The machines, production sections and production lines of the production system include sensors that are communicatively connected to and exchange data with the assembly quantity planning system and the production planning and control system via Internet of Things technology. This allows automated or autonomous operation of the production system to be implemented.

The instructions of the computer program according to the invention include machine instructions, source text or object code written in assembly language, an object-oriented programming language, for example C++, or in a procedural programming language, for example C. According to one aspect of the invention, the computer program is a hardware-independent application program which, for example, has a data carrier or a data carrier signal is provided using Software Over The Air technology.

According to one aspect of the invention, the method according to the invention, the assembly quantity planning system, are based on production planning and control system and/or the computer program based on artificial intelligence.

The invention has the advantage that assembly sequences no longer have to be planned manually, for example using Excel, but can be created automatically, in particular with machine learning algorithms, even with several 100 replannings per day. For example, 10,000 gearboxes are scheduled in the assembly sequence every day, with a variance of 300 different hardware parts lists and 100 goods recipients all over the world. The invention creates the necessary assembly requirements, also in compliance with a large number of secondary conditions and restrictions, for example due to short-term material shortages, requirements for flexibility and variance.

• • Randbedingungen umfassend fixierte Stückzahlen der Montageteile und/oder Planungshorizonte;
• • Produktionsparameter umfassend Arbeitersituation, Maschinenfähigkeiten, Materialverfügbarkeiten, Lager-, Pufferbestände und/oder Lieferantenkapazitäten;
• • Bedarfe umfassend zu vorgegebenen Zeitpunkten bereitzustellende Produktvarianten, Gewichtungen und/oder Priorisierungen zu produzierender Teile;
• • Daten des Produktionssystems umfassend Anzahl von Produktionsabschnitten, Produktionslinien und/oder Puffer;

und/oder die Optimalitätskriterien umfassen Auslastungen von Produktionskapazitäten und/oder die Abweichungen werden über Zeitintervalle unterschiedlich gewichtet. Durch diese weiteren Eingabedaten wird die Montagestückzahlplanung verbessert.According to one aspect of the invention, the input data includes

• • Boundary conditions including fixed numbers of assembly parts and/or planning horizons;
• • Production parameters including worker situation, machine skills, material availability, stock, buffer stocks and/or supplier capacities;
• • Requirements for product variants, weightings and/or prioritization of parts to be produced that are to be provided comprehensively at specified times;
• • Production system data including number of production sections, production lines and/or buffers;

and/or the optimality criteria include utilization of production capacities and/or the deviations are weighted differently over time intervals. This additional input data improves assembly quantity planning.

The boundary conditions must be adhered to for an optimization run to be started. According to one aspect of the invention, the boundary conditions can be changed again by the assembly planner.

The planning horizon specifies, for example, over how many hours, shifts and/or days the assembly requirements planning should be preplanned.

The worker situation includes the worker occupancy, that is, how many workers are or are needed on which line for which shift.

The requirements include product variants that must be available for shipping at a given point in time.

The weightings that differ over time can, for example, weight the range strategy differently in one or more points in time/periods of time. The total cost function GK has the following structure, for example: GK = α_1*range deviation+β_1*production specification deviation+γ_1*setup times+...+α_n*range deviation+β_n*production specification deviation +γ_n*setup times, where α_i≥0,β_i≥0,γ_i≥0 one weighting of the various partial terms over the periods 1,...,n and is variable.

According to a further aspect of the invention, a model for the constraints of the production system and a model for an optimization function are created based on the input data. The model for the constraints is formulated into a decision problem based on SMT - Satisfiability Modulo Theories - and an SMT solver checks the satisfiability of the constraints. If the constraints are satisfiable, the model for the constraints and the model for the optimization function are formulated into an optimization problem based on Linear Programming involving MILP - Mixed Integer Linear Programming - or on CP - Constraint Programming - and a MILP or CP solver determines a Solution that satisfies the constraints and has a globally optimal value in the optimization function.

The model for the restrictions, also known as the constraint system, first finds out whether the restrictions can be fulfilled. In addition, possible solutions may be of interest. The constraints model has a solution if and only if there is an assembly need that satisfies the constraints. The questions of whether restrictions can be met and concrete solutions are closely linked. If a constraint can be satisfied, then there is at least one solution. However, the calculation of a solution is often more complicated than the determination of the feasibility.

For example, it can be determined relatively quickly whether a restriction cannot be met and, in this case, the assembly planner can quickly be given a reason as to which specifications cannot be met together. For example, the assembly planner fixes the assembly quantity of 100 parts of a material. However, due to delivery restrictions, only 50 parts can be built. In this case, the reason is the combination of the fixed assembly quantity and the delivery restriction. To calculate this declaration, the number of assembly planning according to the invention uses a solution method based on the satisfiability modulo theories.

Formulating the model for the constraints in a decision problem or an optimization problem means using logical operators from the constraint system for an application of the target method, here Satisfiability Modulo Theories, abbreviated SMT, or Mixed Integer Linear Programming, abbreviated MILP, or Constraint Programming CP, to be rephrased.

Satisfiability modulo theories are well known in the fields of computer science and mathematical logic. The SMT solution method enables an advantageous determination of constraints that cannot be fulfilled, also known as unsatisfiable cores or unsatisfiable assumptions. Here, non-changeable restrictions such as line capacity, fixed assembly quantities and/or material deliveries are passed to the SMT solver as assumptions. If the constraint system is unsatisfiable, the SMT solver returns a subset of those assumptions that led to the unsatisfiable assumptions. In a subsequent step, these assumptions are used to create justifications for the assembly planner and, if necessary, recommendations for action. In principle, an alternative solution method can also be used for this step, as long as it offers the possibility of generating unsatisfiable cores or unsatisfiable assumptions. SMT solvers, also known as SMT solvers, are available as open source software or as proprietary software. The SMT solver performs a satisfiability check. According to one aspect of the invention, the SMT solver is a Z3 Theorem Prover SMT solver executable on Linux, Mac OS and Windows operating systems and is available as open source software in source code under the MIT License.

MILP/CP methods have the advantage that they find a global minimum relatively quickly. The MILP/CP solver, also known as the MILP/CP solver, is then able to use the given constraint system and the optimization function to determine the solution that satisfies the constraint system and has a globally optimal value in the optimization function. The solution contains all planning-relevant information. The optimal solution specifies the assembly requirements. For example, in the commercial software Matlab Mixed-Integer Linear Programming is implemented by the program command intlinprog(f,intcon,A,b).

According to one aspect of the invention, the constraint system, the satisfiability check based on SMT solution methods and the optimization based on MILP/CP methods form independent software modules or an overall software module of the assembly quantity planning system. According to a further aspect of the invention, the constraint system and the satisfiability test are carried out in parallel, for example on a hardware architecture for parallelized computing, for example graphics processors.

According to a further aspect of the invention, the limitations of the production system include compliance with capacities of production lines of the production system and the number of assembly parts on the production lines are optimized for specified products while maintaining the respective line capacities.

mit vorgegebener Linienkapazität capa(l). Ferner umfasst das Constraint System als Optimierungsfunktion beispielsweise eine Minimierung der Abweichungen zur geplanten Kundenreichweite. Dieses Constraint System wird dann in die Eingabe für das SMT-Lösungsverfahren und MILP/CP-Verfahren kodiert oder transformiert.For example, the constraint system includes the variable x _m,l,c , that is, the number of materials m on production line I for customer c, where m, I and c are variable. For example, the constraint is that for all production lines I: $$\forall l\in P\mathrm{right}O\mathrm{i.e}\mathrm{and}ktiOnlinien{\displaystyle \sum _{c\in K\mathrm{and}n\mathrm{i.e}e}{\displaystyle \sum _{m\in Mate\mathrm{right}ial}{x}_{m,l,c}\le \text{capa}\left(l\right)}}$$

with given line capacity capa(l). Furthermore, as an optimization function, the constraint system includes, for example, minimizing the deviations from the planned customer range. This constraint system is then encoded or transformed into the input for the SMT solution method and MILP/CP method.

According to a further aspect of the invention, ranges, stocks, proportions of components of an optimization function in an overall result and/or costs of the deviations are provided as outputs or ranges, stocks, a chronological progression of the ranges and/or stocks, and/or proportions of components of a Optimization function on an overall result and costs of the deviations over time intervals are provided as outputs. These outputs increase the ability to explain the result, for example why what quantity was planned and how. Furthermore, these outputs can improve the assessment of the optimization result or the optimization parameters for the assembly planner.

• • Materialbedarfe in Produktionsabschnitten des Produktionssystems optimiert geordnet werden;
• • einer der optimierten Materialbedarfe ausgewählt und auf Materialien in vorausgehenden Produktionsabschnitten, die für die Produktion in dem Produktionsabschnitt des ausgewählten Materialbedarfs benötigt werden, projiziert wird;
• • zumindest eine Bedarfsmenge und/oder ein Bedarfszeitpunkt der jeweiligen Materialien angepasst werden;
• • Produktionslinien des Produktionssystems basierend auf zumindest der angepassten Bedarfsmenge und/oder des Bedarfszeitpunktes geregelt und/oder gesteuert werden;
• • in einer Überprüfung überprüft wird, ob die angepasste Bedarfsmenge der jeweiligen Materialien für den ausgewählten Materialbedarf ausreichend ist, wobei bei positiver Überprüfung die jeweiligen Materialien für das Produktionssystem reserviert werden, und/oder der ausgewählte Materialbedarf ausgeführt wird.

According to a further aspect of the invention, the planning-relevant outputs are sequenced in order to optimize assembly sequences and/or production sequences

• • Material requirements in production sections of the production system are arranged in an optimized manner;
• • one of the optimized material requirements is selected and projected onto materials in previous production stages that are required for production in the production stage of the selected material requirement;
• • at least one required quantity and/or a required time of the respective materials are adjusted;
• • Production lines of the production system are regulated and/or controlled based at least on the adjusted required quantity and/or the required time;
• • A check is carried out to determine whether the adjusted required quantity of the respective materials is sufficient for the selected material requirements, with the respective materials being reserved for the production system if the check is positive, and/or the selected material requirements being executed.

The material requirements include material types or categories. Material types include raw materials, such as iron, auxiliary materials, such as screws, supplies, such as energy, work in progress, such as pre-assembled components that still need to be assembled, finished goods, such as finished products ready for shipment, and merchandise. Optimized material requirements optimize an overall cost function of the production system. The total cost function determines the cost-minimum production processes from the technically efficient production processes. The cost function represents the total cost of a production process resulting from the inputs used multiplied by their respective market prices or weights. For example, the total cost function is defined as: Demand Fulfillment: Time of delay, weighted per demand; Production utilization: times with production downtime and/or production ancillary conditions: transport between lines, set-up times.

In production planning, sequencing or sequence planning, also known as sequencing and scheduling, includes the formation of a production sequence for production orders.

• • Optimierte Produktionssequenz: Welches Material wird auf welcher Linie zu welcher Zeit benötigt?
• • Arbeiterbelegung: Wie viele Arbeiter sind oder werden benötigt auf welcher Linie zu welcher Schicht?
• • Lieferantenaufträge: Welches und wie viel zugeliefertes Material ist zu welcher Zeit verfügbar?

After running through this sequencing, the method delivers signals relevant to regulation and/or control, including signals for production sequences, worker assignments and supplier orders, for example

• • Optimized production sequence: Which material is needed on which line and when?
• • Worker occupancy: How many workers are or are needed on which line for which shift?
• • Supplier orders: Which and how much delivered material is available at what time?

The sequencing for an adapted production optimization is in the German patent application with the file number 10 2020 203 716.7 disclosed, for example claims 1 to 4, associated description pages and 2 and 3 and associated description pages. These cited disclosures are incorporated herein by reference.

• • die Produktionslinien in Abhängigkeit zumindest der angepassten Bedarfsmenge und/oder des Bedarfszeitpunktes geregelt und/oder gesteuert;
• • die Materialbedarfs ordnungsgemäß ausgewählt;
• • in einer Überprüfung überprüft, ob die angepasste Bedarfsmenge der jeweiligen Materialien für den jeweils ausgewählten Materialbedarf ausreichend ist, wobei bei positiver Überprüfung die jeweiligen Materialien für das Produktionssystem reserviert werden, und/oder der ausgewählte Materialbedarf ausgeführt wird, und ein weiterer Materialbedarf ordnungsgemäß ausgewählt wird.

According to a further aspect of the invention, the assembly quantity planning, the production system, production planning and control are simulated and in the simulation

• • regulated and/or controlled the production lines depending at least on the adjusted required quantity and/or the required time;
• • the material needs properly selected;
• • Checks in a check whether the adjusted requirement quantity of the respective materials is sufficient for the respectively selected material requirement, whereby if the check is positive, the respective materials are reserved for the production system and/or the selected material requirement is executed and another material requirement is properly selected .

According to one aspect of the invention, the simulation is a detailed image of the entire production. In the simulation, existing problems in the implementation of the assembly requirements become apparent, for example due to the influence of factors that are not taken into account in the quantity planning. According to a further aspect of the invention, internal deliveries are represented as a simulation of the corresponding production sections. This makes it possible to determine the feasibility of the production plan with all the influences and restrictions of the entire factory. Through the simulation, the optimization achieved and thus the entire production system is advantageously adapted to production changes.

• • Durchführen eines ersten Subverfahrens und eines zweiten Subverfahrens, wobei
• • das erste Subverfahren die Schritte
  • ◯ Priorisieren von Materialbedarfen in den Produktionsabschnitten in Abhängigkeit einer Einwirkung auf eine Optimierung einer Kostenfunktion des Produktionssystems;
  • ◯ Auswählen eines der Materialbedarfe in Reihenfolge der Priorisierung, Anpassen von zumindest einer Bedarfsmenge und/oder eines Bedarfszeitpunktes von Materialien in vorausgehenden Produktionsabschnitten zur Ausführung des Materialbedarfs und Reservieren der Materialien und der jeweils angepassten Bedarfsmenge und/oder des Bedarfszeitpunktes;
  • ◯ Auswählen eines weiteren der Materialbedarfe, Wiederholen der vorangehenden Schritte, bis für alle der priorisierten Materialbedarfe die Materialien und die jeweils angepassten Bedarfsmengen und/oder Bedarfszeitpunkte reserviert sind, und Erhalten der Produktionssequenz; und
• • das zweite Subverfahren die Schritte
  • ◯ Fixieren eines ersten Produktionszeitraums in der Produktionssequenz;
  • ◯ Optimieren der Produktionssequenz außerhalb des fixierten ersten Produktionszeitraums zur weiteren Optimierung der Kostenfunktion
  umfasst, wobei
• • das Produktionssystem gemäß der in dem zweiten Subverfahren erhaltenen optimierten Produktionssequenz geregelt und/oder gesteuert wird.

According to a further aspect of the invention, the simulation comprises the following steps:

• • performing a first sub-method and a second sub-method, where
• • the first sub-procedure the steps
  • ◯ prioritization of material requirements in the production sections depending on an impact on an optimization of a cost function of the production system;
  • ◯ Selecting one of the material requirements in order of prioritization, adjusting at least one requirement quantity and/or one Requirement time of materials in previous production sections for the execution of the material requirements and reserving the materials and the respectively adjusted required quantity and/or the required time;
  • ◯ Selecting another of the material requirements, repeating the previous steps until all of the prioritized material requirements have the materials and the respective adjusted requirement quantities and/or requirement times reserved, and obtain the production sequence; and
• • the second sub-procedure the steps
  • ◯ fixing a first production period in the production sequence;
  • ◯ Optimizing the production sequence outside of the fixed first production period to further optimize the cost function
  includes, where
• • the production system is regulated and/or controlled according to the optimized production sequence obtained in the second sub-method.

The first sub-method corresponds to a fast optimization method, which delivers an initial production sequence as the first result within a few seconds. A first result after a very short time is relevant, since production must never come to a standstill after an incident. The optimization goal is to cover demand in combination with maximizing production utilization, i.e. as little production downtime as possible. The second sub-procedure corresponds to a more thorough optimization procedure.

The initial production sequence is also given for implementation in the real production system or the real factory. As soon as a result from one of the downstream thorough optimizations that is better in terms of the total cost function is available, this is output and implemented directly by the system in real production or output to the latter to support a human controller. This ensures that the better result matches the production sequence that has already started. This is ensured by the fact that every plan just put into production by a previous optimizer for the current time is a constraint of the downstream optimizer.

The fix ensures that the output of the second sub-procedure can also be implemented. Due to the fixation, a part of the production sequence determined by the fixation can no longer be changed in the second sub-process. According to one aspect of the invention, all input parameters are fixed in time for a certain period of time. The fixation is implemented, for example, by a prefix in the production sequence, worker situation and/or in the deliveries obtained from the first sub-process. Due to the fixation, the second sub-procedure can take as much time as is covered by the fixation. For example, the time until the end of the current shift is taken as the fixation time. The fast optimizer optimizes over all production periods of production. The first production period optimized by the fast optimizer is then executed in the real factory and can no longer be changed. Therefore, the slower but more thorough optimizer optimizes further production periods outside of the fixed period.

According to a further aspect of the invention, an evolutionary algorithm is executed to carry out the second sub-method, which is initialized with the production sequence obtained in the first sub-method or with a mutation of this.

• • Initialisierung: Die erste Generation von Lösungskandidaten wird erzeugt. Erfindungsgemäß ist die erste Generation die initiale Produktionssequenz. Die initiale Produktionssequenz wird durch das erfindungsgemäße Verfahren, das heißt den schnellen Optimierer, erzeugt.
• • Evaluation: Jedem Lösungskandidaten der Generation wird entsprechend seiner Güte ein Wert einer Fitnessfunktion zugewiesen. Die Fitnessfunktion ist die Zielfunktion des evolutionären Algorithmus. Vorbild für die Fitnessfunktion ist die biologische Fitness, die den Grad der Anpassung eines Organismus an seine Umgebung angibt. Bei dem evolutionären Algorithmus beschreibt die Fitness einer Produktionssequenz, wie gut die Produktionssequenz das zugrunde liegende Optimierungsproblem löst.
• • Durchlaufe die folgenden Schritte, bis ein Abbruchkriterium erfüllt ist:
  • ◯ Selektion: Auswahl von Individuen für die Rekombination
  • ◯ Rekombination: Kombination der ausgewählten Individuen
  • ◯ Mutation: Zufällige Veränderung der Nachfahren
  • ◯ Evaluation: Jedem Lösungskandidaten der Generation wird entsprechend seiner Güte ein Wert der Fitnessfunktion zugewiesen.
  • ◯ Selektion: Bestimmung einer neuen Generation.

Evolutionary algorithms are inspired by the functionality of the evolution of natural living beings and are processed according to the following procedure:

• • Initialization: The first generation of solution candidates is created. According to the invention, the first generation is the initial production sequence. The initial production sequence is generated by the method according to the invention, ie the fast optimizer.
• • Evaluation: Each solution candidate of the generation is assigned a value of a fitness function according to its quality. The fitness function is the objective function of the evolutionary algorithm. The model for the fitness function is biological fitness, which indicates the degree of adaptation of an organism to its environment. In the evolutionary algorithm, the fitness of a production sequence describes how well the production sequence solves the underlying optimization problem.
• • Go through the following steps until an abort criterion is met:
  • ◯ Selection: selection of individuals for recombination
  • ◯ Recombination: Combination of the selected individuals
  • ◯ Mutation: Random change in the descendants
  • ◯ Evaluation: Each solution candidate of the generation is assigned a value of the fitness function according to its quality.
  • ◯ Selection: determination of a new generation.

According to one aspect of the invention, the evolutionary algorithms are processed with machine learning models.

Typical termination criteria are listed below.

• • Adaptive Anpassung oder 1/5-Erfolgsregel: Die 1/5-Erfolgsregel besagt, dass der Quotient aus den erfolgreichen Mutationen der initialen Produktionssequenz, also Mutationen, die eine Verbesserung des Produktionsablaufs bewirken, zu allen Mutationen etwa ein Fünftel betragen sollte. Ist der Quotient größer, so sollte die Varianz der Mutationen erhöht werden, bei einem kleineren Quotienten sollte sie verringert werden.
• • Selbstadaptivität: Jedes Individuum hat ein zusätzliches Gen für die Mutationsstärke selbst. Dies ist zwar nicht in der Biologie möglich, aber die Evolution im Rechner findet eine geeignete Varianz auf diese Weise ohne Einschränkung des Menschen. Im Rechner werden dabei Rekombination und Mutation entsprechend der Mutationsstärke entsprechend angepasst.

Evolutionary algorithms include genetic algorithms. With evolutionary algorithms, a solution candidate is sought exclusively via mutation, recombination does not take place. Genetic algorithms take recombination into account. According to one aspect of the invention, the evolutionary algorithm is executed based on one of the following evolution strategies:

• • Adaptive adaptation or 1/5 success rule: The 1/5 success rule states that the quotient of the successful mutations in the initial production sequence, ie mutations that improve the production process, should be around one fifth of all mutations. If the quotient is larger, the variance of the mutations should be increased, if the quotient is smaller, it should be reduced.
• • Self-adaptivity: Each individual has an additional gene for the mutation strength itself. This is not possible in biology, but computer evolution finds a suitable variance in this way without human limitations. In the computer, recombination and mutation are adjusted according to the mutation strength.

For example, the deep optimization genotype consists of a data structure that the fast optimizer uses. The solution of the fast optimizer is used as the initial population and then the order of the material requirements in the data structure is changed by recombination and mutation. In this case, the mutation operator changes the order of a randomly selected material requirement of a randomly selected production area. The recombination operator takes two chromosomes from parents and produces two chromosomes from children. This is achieved, for example, by recombination of permutations. The phenotype is derived from the genotype by running the fast optimizer on the changed data structure.

This further improves the adaptive production optimization.

The first and second sub-procedures and evolutionary algorithms for an adapted production optimization are in the German patent application with the file number 10 2020 203 718.3 discloses, e.g. claim 1, associated description pages and 4 and 5 and associated description pages. These cited disclosures are incorporated herein by reference.

According to a further aspect of the invention, a digital twin of a real production system is generated in the simulation. A planning horizon is determined for the digital twin. The real production system is regulated and/or controlled using the planning horizon. Bottlenecks or critical paths, for example, are simulated using the digital twin. Future states of the production system are simulated using the digital twin. This enables planning horizons that reach as far into the future as you like, for example several weeks.

According to a further aspect of the invention, outputs relevant to regulation and/or control and/or further outputs are provided, the outputs relevant to regulation and/or control comprising production sequences, worker allocation and/or supplier orders and the further outputs covering material requirements, completion dates, utilization, bottlenecks, include critical paths and/or temporal development of the production system. This allows production sequences to be implemented automatically.

According to a further aspect of the invention, the production planning and control system according to the invention includes a cloud infrastructure. The cloud infrastructure includes cloud-based storage. A simulation and a real regulation and/or control of the production system, the production planning, including the assembly quantity planning, and production control take place in the cloud. The cloud infrastructure is, for example, a public, private or hybrid cloud infrastructure.

• 1 ein Ausführungsbeispiel eines erfindungsgemäßen Produktionsplanung undsteuerungssystems,
• 2 ein Ausführungsbeispiel eines erfindungsgemäßen Montagestückzahlplanungssystems,
• 3 ein Ausführungsbeispiel einer Kostenfunktion,
• 4 eine schematische Darstellung eines erfindungsgemäßen Verfahrens und
• 5 eine Ausführungsbeispiel einer erfindungsgemäßen Ergebnisdarstellung der Montageplanung.

The invention is illustrated in the following exemplary embodiments. Show it:

• 1 an embodiment of a production planning and control system according to the invention,
• 2 an embodiment of an assembly quantity planning system according to the invention,
• 3 an embodiment of a cost function,
• 4 a schematic representation of a method according to the invention and
• 5 an exemplary embodiment of a presentation of the results of the assembly planning according to the invention.

In the figures, the same reference symbols denote the same or functionally similar reference parts. For the sake of clarity, only the relevant reference parts are highlighted in the individual figures.

1 shows an inventive production planning and control system APO. The production planning and control system APO includes an assembly quantity planning system AQP according to the invention and a sequence optimizer APO_1. The production planning and control system APO supports, for example, an assembly planner MP in determining assembly requirements and assembly sequences.

Based on optimality criteria P1, boundary conditions P2, production parameters P3, requirements P4 and data P5 of a production system F, which can be partially, combined or completely adapted by the assembly planner to changed situations during the optimization, the assembly quantity planning system AQP determines planning-relevant expenditure comprehensively according to the method according to the invention the assembly needs that optimize the total cost function of the production system F. The parameters P1 to P5 form the input data for the assembly quantity planning system AQP. If certain restrictions C cannot be met, the assembly quantity planning system AQP reports back to the assembly planner MP which specifications cannot be met.

In addition, the AQP assembly quantity planning system provides further information on covering material requirements, utilization of machines and workers and critical paths. The other outputs include visualizations of the components of the optimization function, for example broken down over time periods, forecast stock development and/or detailed range and stock development for each planned material and each goods recipient.

The assembly requirement is entered, for example, in a data structure in the sequence optimizer APO_1. The sequence optimizer APO_1 determines how above and in the German patent applications with the file numbers 10 2020 203 716.7 and 10 2020 203 718.3 discloses the optimal production sequences encompassing a planning horizon. If no optimal production sequences can be determined under the given restrictions C, feedback is sent to the assembly planner MP with an indication of which specifications cannot be met.

The optimized production sequences are released and implemented in the production system F, either by the assembly planner MP or automatically or autonomously. Alternatively, the assembly planner MP changes the input data for the assembly quantity planning system AQP based on the optimized production sequences.

The individual functions of the assembly quantity planning system AQP are in 2 shown. According to one aspect of the invention, the constraint system comprising a model M1 for the restrictions C and a model M2 for the optimization function, the feasibility check using SMT solvers and the optimization of the assembly quantity planning using MILP/CP solvers are individual software modules, according to another aspect the Invention an overall software module.

The constraint system is created from the inputs P1-P5 in a method step O5. This has a solution exactly when there is a need for assembly that meets the specifications. This constraint system includes variables, e.g. integer variables x _m,l,c ∈Z, representing the number of assembly parts m on production lines I for customer c, and constraints C, e.g. compliance with the line capacity capa. The optimal solution of this constraint system corresponds to the optimal assembly requirements according to the specifications in the input. This constraint system is encoded or transformed into the input for the respective target methods SMT and MILP/CP.

First, the constraint system model M1 is encoded as an SMT constraint system. A solution is determined in a method step O6 with the aid of an SMT solver, after which the optimization is continued in a method step O7. If there is no solution, the assembly quantity planning system AQP uses the SMT solver to determine the unsatisfiable cores and converts them into an explanation for the assembly planner MP. An alternative solution method can also be used for this step, as long as it offers the possibility of generating unsatisfiable cores.

The constraint system together with the model M2 is encoded as a MILP/CP constraint system. The MILP/CP solver is then able to use the given constraint system and the optimization function to determine the solution that satisfies the constraint system and has a globally optimal value in the optimization function. The solution contains all planning-relevant information. Other optimization methods, such as heuristic search, for example Tabu Search, genetic and evolutionary optimization can be used in this step, possibly with initialization of the solution from method step O5.

3 shows as a total cost function costs for deviations from the specified range per goods recipient and material. The weighting of the deviation, i.e. how much the deviation is included in the costs, can be specified in intervals, for example cost factor 1 for deviation up to 1 day and cost factor 10 for deviation of more than 1 day.

4 symbolically shows a method according to the invention for planning the number of assembly parts m.

In a method step O1, input data including the optimality criteria P1 including deviations of the production system F from specified range targets and/or production specifications for products that can be assembled from the assembly parts m are obtained.

In a method step O2, the deviations from the specified range targets in terms of quantities of the assembly parts m are optimized based on the input data while complying with the restrictions C of the production system F. The method steps O5-O7 form the method step O2.

In a method step O3, information about which specifications cannot be met is output if the optimization cannot be met.

Otherwise, in a method step O4, the outputs relevant to planning are output, including the number of pieces of assembly parts m, which optimize the deviations from the specified range targets.

5 shows a representation of the result from the assembly quantity planning system AQP. Shown is the quantity of material numbers to be built for one working day obtained according to the invention, broken down into the various production lines L1-L12. According to the invention, the capacity specifications are fully utilized and production is thus optimized. The production planning and control system APO plans the entire horizon, for example 14 days, in intervals, for example days. The result of the assembly quantity planning system AQP overall horizon/interval length has many entries. For each of these entries, the representation of the 5 to be created.

Reference List

AQP

assembly quantity planning system

APO_1

sequence optimizer

APO

Production planning and control system

P1

optimality criteria

p2

boundary conditions

P3

production parameters

P4

needs

P5

data production system

f

production system

MP

assembly planner

M1

Model for Constraints

M2

Model for an optimization function

C

restriction

SMD

SMT solver

MILP/CP

MILP/CP solver

capa

line capacity

I

production line

L1-L12

production line

m

assembly part

c

customer

O1

-O7 process steps

QUOTES INCLUDED IN DESCRIPTION

This list of documents cited by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent Literature Cited

• DE 102020203716 [0004, 0051, 0074]
• DE 102020203718 [0004, 0065, 0074]

Claims (15)

Computer-implemented method for planning the number of assembly parts (m) for production optimization of a production system (F) comprising the steps • Obtaining input data including optimality criteria (P1) including deviations of the production system (F) from specified range targets and/or production specifications for products (O1) that can be assembled from the assembly parts (m); • Optimizing the deviations from the specified range targets in terms of quantities of the assembly parts (m), based on the input data in compliance with the restrictions (C) of the production system (F) (O2); • Outputting information about which specifications cannot be met if the optimization cannot be met (O3); • Otherwise, providing planning-relevant outputs including the number of assembly parts (m) that optimize the deviations from the specified range targets (O4). procedure after claim 1 , wherein the input data • boundary conditions (P2) comprising fixed quantities of the assembly parts (m) and/or planning horizons; • Production parameters (P3) including worker situation, machine skills, material availability, stock, buffer stocks and/or supplier capacities; • Requirements (P4) comprising product variants, weightings and/or prioritization of parts to be produced that are to be provided at specified times; • data (P5) of the production system (F) comprising a number of production sections, production lines and/or buffers; and/or the optimality criteria include utilization of production capacities; and/or the deviations are weighted differently over time intervals. procedure after claim 1 or 2 , where • based on the input data, a model (M1) for the restrictions (C) of the production system (F) and a model (M2) for an optimization function are created (O5); • the model (M1) for the restrictions (C) is formulated in a decision problem based on SMT - Satisfiability Modulo Theories - and an SMT solver (SMT) checks the satisfiability of the restrictions (O6); • if the constraints can be satisfied, the model (M1) for the constraints (C) and the model (M2) for the optimization function are formulated in an optimization problem based on linear programming including MILP - Mixed Integer Linear Programming - or on CP - Constraint Programming and a MILP or CP solver (MILP/CP) determines a solution that satisfies the constraints (C) and has a globally optimal value in the optimization function (O7). Procedure according to one of Claims 1 until 3 , whereby the restrictions (C) of the production system (F) include compliance with capacities (capa) of production lines (I) of the production system (F) and the quantities of assembly parts (m) on the production lines (I) for specified products in compliance with the respective Line capacities (capa) are optimized. Procedure according to one of Claims 1 until 4 , where ranges, stocks, proportions of components of an optimization function in an overall result and/or costs of the deviations are provided as outputs or where ranges, stocks, a chronological progression of the ranges and/or stocks, and/or proportions of components of an optimization function in one The overall result and costs of the deviations over time intervals are provided as outputs. Procedure according to one of Claims 1 until 5 , wherein the planning-relevant outputs are sequenced to optimize assembly sequences and/or production sequences, wherein • material requirements in production sections of the production system are arranged in an optimized manner; • one of the optimized material requirements is selected and projected onto materials in previous production stages that are required for production in the production stage of the selected material requirement; • at least one required quantity and/or a required time of the respective materials are adjusted; • Production lines (I) of the production system (F) are regulated and/or controlled based at least on the adjusted required quantity and/or the required time; • A check is carried out to determine whether the adjusted required quantity of the respective materials is sufficient for the selected material requirements, with the respective materials being reserved for the production system (F) if the check is positive, and/or the selected material requirements being executed. procedure after claim 6 , wherein the assembly quantity planning, the production system (F), production planning and control are simulated and in the simulation • the production lines (I) are regulated and/or controlled depending at least on the adjusted required quantity and/or the required time will; • material requirements are properly selected; • a check is carried out to determine whether the adjusted requirement quantity of the respective materials is sufficient for the selected material requirement, whereby if the check is positive, the respective materials are reserved for the production system (F) and/or the selected material requirement is executed, and another Material requirements are properly selected. procedure after claim 7 , the simulation comprising • carrying out a first sub-procedure and a second sub-procedure, wherein • the first sub-procedure includes the steps o prioritizing material requirements in the production sections depending on an effect on an optimization of a cost function of the production system (F); ◯ Selecting one of the material requirements in the order of prioritization, adjusting at least one required quantity and/or a required time of materials in previous production stages for the execution of the material requirements and reserving the materials and the respectively adjusted required quantity and/or the required time; ◯ Selecting another of the material requirements, repeating the previous steps until all of the prioritized material requirements have the materials and the respective adjusted requirement quantities and/or requirement times reserved, and obtain the production sequence; and • the second sub-method includes the steps ◯ fixing a first production period in the production sequence; ◯ Optimizing the production sequence outside of the fixed first production period for further optimization of the cost function, wherein • the production system (F) is regulated and/or controlled according to the optimized production sequence obtained in the second sub-method. procedure after claim 8 , wherein an evolutionary algorithm is executed to carry out the second sub-process, which is initialized with the production sequence obtained in the first sub-process or a mutation of this. Procedure according to one of Claims 7 until 9 , wherein a digital twin of a real production system (F) is generated in the simulation, a planning horizon is determined for the digital twin and the real production system (F) is regulated and/or controlled using the planning horizon. Procedure according to one of Claims 1 until 10 , wherein regulation and/or control-relevant outputs and/or further outputs are provided, wherein the regulation and/or control-relevant outputs include production sequences, worker occupancy and/or supplier orders and the further outputs cover material requirements, completion dates, utilization, bottlenecks, critical paths and/or or development of the production system (F) over time. Assembly quantity planning system (AQP), executed, a method according to at least one of Claims 1 until 5 to execute. Production planning and control system (APO) including an assembly quantity planning system (AQP). claim 12 and executed a method according to any one of Claims 6 until 11 to execute. system after Claim 13 comprising a cloud infrastructure, the cloud infrastructure comprising a cloud-based memory, with a simulation and a real regulation and/or control of the production system (F), the production planning, comprising the assembly quantity planning, and production control taking place in the cloud. Computer program for assembly quantity planning of assembly parts (m) for production optimization of a production system (F) comprising commands that cause a computer to carry out a method according to one of Claims 1 until 11 executes when the computer program runs on the computer.

Images