WO2018050252A1

WO2018050252A1 - Method and device for planning a production process from a specification and basic functions

Info

Publication number: WO2018050252A1
Application number: PCT/EP2016/072176
Authority: WO
Inventors: Sebastian Albrecht; Stephan Grimm; Daniel MEYER-DELIUS DI VASTO; Maja MILICIC BRANDT
Original assignee: Siemens Aktiengesellschaft
Priority date: 2016-09-19
Filing date: 2016-09-19
Publication date: 2018-03-22

Abstract

The invention relates to a method for planning a production process of a product composed from multiple sub-products in a production system (SYS) consisting of multiple technical components (D1, D2, D3). The current state of the production system (SYS) and semantic information (ASM,SKM,PO,PR) are stored in a knowledge base (KB) or a database. The semantic information comprises data on the production system and on basic functions (SK) which can be executed in the form of encapsulated programs by the components (D1, D2, D3) in the production system, on sub-products which can be processed in the production process, and on production steps which can be carried out in the production process and dependencies thereof which can be represented as a directed graph with edges and nodes. The following steps are carried out in the method: a) reading a digital specification (SPEC) for producing a specified product; b) planning a production process for the specified product on the basis of the read specification (SPEC) using a planning program (PE) and an inference program (RE), wherein the planning program (PE) generates the production process in the form of a sequence of basic functions (SK) using information derived by an inference from the knowledge base by means of the inference program (RE); and c) controlling the execution of the sequence of basic functions (SK) by means of an execution program (EE).

Description

METHOD AND DEVICE FOR PLANNING A PRODUCTION PROCESS FROM ONE

SPECIFICATION AND BASIC FUNCTIONS

description

Method and device for the computerized control of a production process for producing a product composed of several partial products

The invention relates to a method and a device for computer-aided control of a production process for producing a product composed of several partial products in a production system comprising a plurality of technical components.

Moreover, the invention relates to a computer program or a computer program product.

Production systems are often used only for the automatic production of a single high-volume product. This is because the establishment of a production ^¬ tion systems for different types of products with a large ^¬ SEN expenditure connected for experts. In particular, are the design and implementation of a producti ^¬ onssystems mostly on the system itself and not to the tasks to be performed of the system aligned (so-called. System-centric programming) today. Consequently, a producti ^¬ onssystem is designed as a single large complex unit and needs to be reconfigured for a new production task essentially and programmed.

To reduce the expense of setting up a production system for a new production task, two different approaches are known. , According to a procedure ^¬ is attempted, the configuration and Programmierauf ^¬ wall to reduce directly. According to another approach, the degree of autonomy and robustness of the system is improved verbes ^¬ .

To directly reduce the configuration and programming effort, there are approaches that make simple and intuitive Programming environments are used. For example, graphical programming can be used instead of the textual creation of application programs. Such solutions are usually proprietary and therefore not universally applicable.

From the prior art, the concept of so-called. "Skills" is also known to those implemented certain basic functions of a robotic system via encapsulated programs ^¬ the. Result, the cost can be reduced in the design and setting up a robotic system.

Furthermore, there are approaches to improve the autonomy and robustness of pro ^¬ duktionssystemen via digital knowledge-based systems. Such knowledge-based systems store semantic knowledge about the production system or the production process to be carried out in a knowledge base. To date, there is no functional implementation of a wissensbasier ^¬ th system for automatic production of a product. The object of the invention is to provide a method for controlling a production process, which can be easily and flexibly adapted to different production tasks. This object is achieved by the independent claims ge ^¬ triggers. Advantageous developments of the invention are defined in the dependent claims.

The inventive method is used for computer-aided control of a production process for producing a composite of several partial products product in a production system of several technical components. In other words, the steps a) to c) of the method explained below are carried out in an automated and computer-aided manner. Under a production system, an automated system is meant, which can carry out by means of the just ge ^¬ called technical components which are involved in the production process, the production process. Here and in the following, the term of producing a product is to be understood broadly. This may be, for example, the manufacture or assembly of a product. It can also be the production of a product via various chemical processes. The technical components involved in the production process are such devices that are required to carry out the production process. In particular, it is actors, such as Ro ^¬ boter or robotic units, and possibly also to sensors that monitor the production process.

In the method according to the invention, the current state of the production system and semantic information are stored in a knowledge base in digital form. The semantic information includes data on the production system, on basic functions that can be executed as encapsulated programs by the components in the production system, on subprocesses that can be processed in the production process and on production steps that can be carried out in the production process and their dependencies. Partial products both starting materials in the production process and intermediates of this pro ^¬ zesses can be. The ^¬ nen-described in the semantic information production steps are not terlegt unlike the Basisfunkti ^¬ tions as encapsulated programs in the production system back ^¬.

In the context of the method according to the invention, the steps a), b) and c) explained below are carried out. In step a) a digital specification for the production of a predetermined product is read. A production process for the predetermined Pro ^¬ domestic product on the basis of the read specification is then scheduled using a planning program, and an inference in step b). The planning program generates the production process in the form of a sequence of the above-mentioned basis functions using information which is derived from the knowledge base by means of the inference program via inference (ie inference). After all In step c), an execution program controls the execution of the sequence of basis functions.

The inventive method is characterized that are per se known concepts suited to each other combinatorial ^¬ ned to deploy a production system flexible and robust for different production tasks. In particular, the concept of basic functions, which are encapsulated programs for specific functionalities, is combined with a semantic knowledge representation in the form of a digital knowledge base. Knowledge can be derived from the knowledge base by means of inference. For this an inference program is used, which can consist of one or more known reasoners (eg FaCT ++). The planning of the production process is reali a planning program ^¬ Siert which accesses the knowledge of the knowledge base and the knowledge derived by the inference engine. The Pla ^¬ calculation program can turn se known computer-aided planner included (eg Fast Downward or Hierarchical Task Network Planner).

In a preferred embodiment of the inventive method, it is checked between step a) and b) whether a production process for producing the predetermined product in accordance with the read-in Spe ^¬ fication can be planned with ^¬ means of the knowledge base. If this is not the case, the procedure is usually already at this time abgebro ^¬ chen, as the conditions do not exist for the production of the corresponding product.

In a further variant of the method the information in the above step b) derived through inference are stored in the knowledge base, so that the Wis ^¬ sen of the knowledge base is further increased continuously.

In a particularly preferred embodiment of the method according to the invention, the semantic information about the production processes that can be carried out in the production process steps and their dependencies represented by one or more process graphs of nodes and edges. These process graphs describe in a simple way the possible production steps in a production system.

In a particularly preferred variant, a process graph is at least designed so that each node represents a production step and the nodes are interconnected via court ^¬ preparing edges. The directed edges describe a partial order such that the

Production step of a node in a directed Kan ^¬ te in running is performed from ^¬ after the production steps of the node from which the directed edge originates and the production step of a node from which a directed edges boils down to is carried out before the production step, in whose node runs in the directional edge.

Depending on the configuration of one or more of the process graph can allow variable partial products and / or one or more of the process graph can define fixed predetermined Teilpro ^¬-products.

In a further, particularly preferred embodiment, the semantic information about a respective basic function comprises the following variables:

- parameters needed to execute the respective base function;

Preconditions which the state of the production system must fulfill to perform the respective basic function; Effects of the execution of the respective base function, which are conditions which the state of the production system fulfills after execution of the respective base function;

Performance characteristics of the respective base function, e.g. the time required to execute it.

In order to enable an efficient access of the execution program to the basic functions, a respective base ^¬ function preferably comprises an interface via which the Execution program can leave the execution of the respective base function and can retrieve the runtime state of execution of the respective base function. In another particularly preferred embodiment, the execution program monitors the execution of the sequence of basis functions and maps the changing states of the production system into the knowledge base, the execution program being aware of the occurrence of disturbances in the execution of the sequence of basis functions (eg, false preconditions or effects ) the re-execution of steps b) and c), ie a re-planning of the production process, causes ^¬ . According to this variant of the invention, the production process can be adapted in real time in the event of unforeseen disturbances.

In a further variant of the method according to the invention, the read-in specification is information entered or edited by a user via a user interface. In other words, a user interface of the ^¬ le is provided, via which to manually set the predetermined produ ^¬ ornamental product or properties of the production process. For example, the process graphs described above for changing the production steps or their dependencies can be edited appropriately.

In addition to the method described above, the invention relates to a device for computer-aided control of a production process for producing a product composed of several sub-products in a production system of several technical components, which are involved in the production process. For example, this device may be realized in the production system with suitable software by ei ^¬ NEN central computer.

In the apparatus, a knowledge base is stored in wel ^¬ cher, the state of production ^¬ system and semantic information are stored in digital form. The Semantic information includes data on the production system, on basic functions that can be executed as encapsulated programs by the components in the production system, on sub-products that can be processed in the production process, and production steps and their dependencies that can be carried out in the production process. The device comprises a read-in unit, eg a user interface

Inference program, a planning program and an execution program, with which the following steps are carried out during operation of the device:

a) reading a digital specification for the production of a predetermined product by the read-in unit;

b) planning a production process for the predetermined product on the basis of the read specification using the planning program and the

Inference program, wherein the scheduling program generates the produc tion process in the form of a sequence of Basisfunktio ^¬ NEN using information which can be derived by means of the inference through inference from Wis ^¬ sensbasis;

c) controlling the execution of the sequence of basis functions by the execution program.

The inventive apparatus is preferably arranged through ^¬ guide one or more preferred variants of the inventive method.

The invention further relates to a computer program with a program code for carrying out the method according to the invention or one or more preferred variants of the method according to the invention, when the program code is executed on a computer.

Furthermore, the invention relates to a computer program product with a program code stored on a machine-readable carrier for carrying out the method according to the invention or one or more preferred variants of the invention. method according to the invention, when the program code is executed on a computer.

An embodiment of the invention will be described below in detail with reference to the accompanying drawings.

Show it:

Figure 1 is a schematic representation of a production system, which is controlled with a variant of erfindungsge ^¬ MAESSEN method. and

Fig. 2 shows the representation of a process graph, in the

1 is based on the knowledge base of the production system and on the basis of which a sequence of

Basic functions is derived.

The method according to the invention will be described below with reference to the production system SYS shown in FIG. On the one hand, this system comprises a central control device CO, which can be stored, for example, in a central computer. With this control device, an embodiment of the method according to the invention is performed. The individual components of the control device CO, which are explained below, are software programs.

The production system SYS further comprises a plurality of technical components or devices, which are reproduced in the block D. By means of these devices is the automated execution of a corresponding Produktionsprozes ^¬ ses. In the embodiment described here, the production system SYS is a production cell for assembling various products, and the assembly of a toy vehicle will be described below. For this assembly, the reproduced in block D technical components Dl to D3 are provided. It is a monocular camera Dl with a corresponding controller DCl pre ^¬ see what the poses of the TEI used in the assembly le recorded. Furthermore, a robot arm D2 is provided with a corresponding controller DC2, via which the components can be moved during assembly. As a further technical component, a gripper D3 with a corresponding controller DC3 is attached to the robot arm, via which gripping of the components takes place during assembly. The production cell further includes a work surface (not shown) for temporarily depositing partial products during assembly. In the embodiment described here has the production ^¬ tion system following basic functions in terms of the claims, said base functions are referred to hereinafter as skills. The skill "pick", which specifies a picking up of a component;

the skill "place", which specifies a placement of a component;

the skill "insert", which specifies the insertion of one component into another component;

the skill "fixate" which specifies the fixation of a component by means of a suitable fastening on the work surface Before the generation according to the invention of a sequence of skills for assembly of the toy vehicle is discussed, the software components required for this purpose are first of all explained in the control device CO controller CO is a digital knowledge base KB rear sets, which contains knowledge on the application domain of the producti ^¬ onssystems SYS as well as the current status of the production ^¬ system. the knowledge about the application domain is ba ^¬ sierend on a semantic standard, in particular in the form of a ontology language (eg OWL) deposited. This knowledge is represented on the models ASM, SCM, POM and PRM. the model ASM (ASM = Asset model) Wis ^¬ sen represented on the production system (ie, its hardware and resources). the model SKM (SKM = Skill Model) Knowledge about the skills of the production system. The POM model (POM = Product Model) represents knowledge about sub-products that can be processed in the production process of the production system (ie raw materials, materials used and intermediate products). The model PRM (PRM = Process Model) be ^¬ writes in the production process executable Produktionsschrit ^¬ te and their dependencies. The model PRM (hereinafter also referred to as process model) is represented by a process graph ^¬ PG, which will be explained later with reference to Fig. 2 in more detail.

The skills described in the above model SKM are shown in FIG. 1 are denoted by SK and provide encapsulated executable program ^¬ me to implement certain functionalities in the production system. In order to execution of the encapsulated skills programs to the appropriate controller DC1 to DC3 of the devices Dl transferred to D3. A respective skill is characterized by the following features, which are stored in the knowledge base KB:

- skill parameters needed to perform the skill;

- Skill preconditions which must meet the condition of the production system ^¬ to perform the skills;

- skill effects, which are conditions that the state of the production system fulfills after execution of the skill;

- Performance characteristics, which define performance characteristics of the skill (eg the duration of its execution). A skill SK further comprising an interface to an in ^¬ to allow to the actual executable program of the skill teraktion. Through this interface, the white ^¬ ter execution unit EE described below can cause execution of the skill or information about the runtime state of skills to retrieve.

The skills explained above are based on the following principles: - Loose coupling: The skills have as little dependency on other skills or technical components as possible;

- Abstraction: The control logic or the encapsulated program of the skills and their internal resources are in front of others

Hidden skills and the rest of the system;

- Reusability: The encapsulated programs can be used more frequently in the production process;

- Autonomy: Skills autonomously control their control logic or their encapsulated program;

- Freedom of State: The execution of each skill is an ^independent version, which has no relation to the previous execution of the same or another skill. Due to the use of semantic standards, more advanced semantic information from the models ASM, SKM, POM, and PRM of the knowledge base KB about inference (ie

Concluding) with one or more reasoners. For this purpose, a CO is in the control device

Inference program RE (RE = Reasoning Engine) provided, which includes several reasoners for inferring from the semantic information of the knowledge base KB.

The control device CO further includes a planning program PE (PE = Planning Engine), wherein the planning program comprises one or more known computer-aided planner (eg Fast-Downward or HTN = Hierarchical Task Network). In addition, a control program CP with a control shell (ie user interface parts) is integrated in the control device CO. Through the control shell, a user enters a specification SPEC specifying which product is to be manufactured using the SYS production system. The control shell can also be used to change the models in the knowledge base KB. For example, the process graph PG described below may be adapted. The specification SPEC is processed by the engineering program to calculate, based on this specification, a sequence of skills with which the desired product can be manufactured by the production system according to the specification. Essential to the invention is that the planning program PE interacts with the inference program RE. The

RE inference leads on inference semantic information from the knowledge base KB from the advertising required by the Planungspro ^¬ gram PE to determine the sequence of the Skills. The planning program PE and the

Inference program RE also interact iteratively with each other. For example, the planning program PE may first one un ^¬ complete macroscopic sequence of skills without all he ^¬ ford variable skill parameters based on from the

Derive information obtained in the inference program RE. Further information from the knowledge base KB can then then again by the inference engine RE are abgelei ^¬ tet, to finally obtain a detailed sequence of skill with all the necessary skill parameters from the scheduling program PE. This sequence can then be executed by the individual components Dl to D3 of the production system SYS.

When determining the sequence of skills, the planning program PE also uses the process model PRM, among other things. As already mentioned, this model is described by a process graph PG (see FIG. 2). A process graph in the embodiment described here is represented by a plurality of nodes and directed edges between the nodes. Each node represents a producti ^¬ onsschritt, which can be performed with the corresponding components of the production system. About the directed edges a partial order is described. In other words, a production step of a node into which a directed edge enters is to be carried out after the production step of the node from which the directed edge originates. Further, the production step of a node is one of directed edge out, before the production step from ^¬ feed, in whose node the directed edge runs. Depending on the embodiment, the process graph can be a general process graph containing only essential loading restrictions with regard to the execution order of the production steps and also allows variable intermediate production ^¬ te under production. Similarly, the graph can be a special assembly tree that unlike ei ^¬ nem general process graph precisely specified intermediates sets during production. A special exporting ^¬ approximate shape of the assembly Trees is the process sequence, which specifies the exact order of all production steps and entspre ^¬ sponding intermediates. If an assembly tree or a process sequence is processed by the planning program PE, the skill sequence may be calculated by the planning program PE directly by means of a classical AI planner (eg fast-downward) (AI = Artificial Intelligence). In other words, only initial information must be derived by the inference program RE for use by the planning program PE, so that in the context of the planning does not refer again to the

Inference program must be resorted to. In the case that the process graph is a general process graph, an HTN scheduler is typically used to determine a macroscopic skill sequence. Subsequently ^¬ again loading the right by means of the inference RE further information in the form of lack of skill parameters, whereupon the scheduler PE then derives the detailed executable skill sequence.

After the generation of the skill sequence by the planning program PE, the production process based thereon is carried out under the control of an execution program EE (EE = Execution Engine). The exemplary program causes the off ^¬ guide the respective skills of the sequence by the individual components Dl to D3. In addition, the Implementing Food Program this execution and updates the entspre ^¬ sponding state of the production system in the knowledge base KB. If faults occur in the production process, the execution program EE may optionally replanning as manufacturing processes based on the current state of the production system ^¬ tion through the scheduler PE in combination with the inference engine RE cause.

The steps performed by the programs of the control device CO are explained again below. First, a specification SPEC is read in the control shell of the control program CP. In a next step, the knowledge base KB is checked as to whether the producti ^¬ onssystem is even able to produce the product according to the specification SPEC. It verifies whether the knowledge relevant to the production of the product exists in the knowledge base. If this is not the case, the procedure is aborted. If the above verification is positive, one or more combined inference and planning phases are performed. In each of these phases, the relevant information required by the planning program PE in addition to the already existing information in the knowledge base KB is derived from the knowledge base via inference (inference) by the inference program RE. Then a sequence of skills is calculated by the planning program PE. If the sequence of skills to execute them does not yet contain all the required information, the required information will be retrieved by the

Inference program RE derived from the knowledge base KB. It should be noted that all information newly derived via the inference program is also repeatedly stored in the KB knowledge base. At the end of this process you just described will contain an executable sequence of skills that can be ^¬ directly under control of the executive program EE. Upon receipt of the executable skill sequence, it is executed, as already mentioned, using the execution program EE. The execution of the encapsulated programs of the skills takes place via the corresponding controllers DC1 to DC3 of the technical components D1 to D3 of the production system SYS.

The assembly of a toy vehicle via the generation of a corresponding sequence of skills by means of the production system of FIG. 1 will now be explained. In this case, the process graph PG of FIG. 2 is used as the process model. The process graph PG comprises a plurality of nodes NO to N9, which represent different production steps. The node NO represents the start of production and node N9 the end of production. These nodes are also to be understood as production steps according to which the production is initialized or terminated. The nodes are connected to one another via directed edges ED, whereby the partial order already described above is determined. In other words, a production step in which a directed edge ED in running always be performed according to the Pro ^¬ duktionsschritt from which the directed edge ED is derived. Also, a production step from which a directed edge amounts to carry out always in front of the production step in which this directed edge ^¬ runs into it. According to the process graph PG all producti ^¬ onsschritte Nl to N6 after the production step and before the NO production step N7, N8 of the production step to the production step N7 and the production step N9 after the production step N8 are thus carried out. It is by the

Not fixed graph in which order the production ^¬ tion steps Nl to N6 must be carried out.

The nodes N1 to N8 of FIG. 2 specify the following production steps for the assembly of a toy vehicle:

Nl: insert the front axle into the chassis;

N2: inserting the rear axle into the chassis; N3: inserting the left front wheel into the front axle;

N4: inserting the right front wheel into the front axle;

N5: inserting the left rear wheel into the rear axle;

N6: inserting the right rear wheel into the rear axle;

N7: inserting the seats into the chassis;

N8: putting the body on the chassis.

By means of this process graph PG, a corresponding sequence of skills is then determined based on the methods explained above. First of all, the control shell of the control program CP of FIG. 1 receives the specification SPEC for the production of a toy vehicle. Then, if the robot arm, the gripper and the camera are ready for use and whether the process graph and the PG weite- ren necessary semantic information for the production of the toy vehicle in the knowledge base KB are stored is checked via ^¬.

After these checks, the inference planning phase already explained above is performed. In this case, the planning program PE interacts with the inference program RE. For example, it can be derived by inference which fastening ^¬ tion is to be used for which component. At the end of this phase you will get a sequence of skills, below is a part of this sequence that shows the attachment of the axles with the corresponding wheels to the chassis:

1 pick (axis)

2 fixates (axis, fixturel, left)

3 pick (wheel)

4 insert (wheel, axis, left)

5 pick (pin)

6 insert (pin, wheel, top)

7 pick (axis)

8 fixates (axis, fixturel, right)

9 pick (wheel)

10 insert (wheel, axis, right) 11 pick (pin)

12 insert (pin, wheel, top)

13 pick (chassis)

14 fixates (chassis, fixture2, bottom)

15 pick (axis)

16 insert (axis, chassis, bottom)

The above skills related to picking up (pick) of a component on the fixation (fixate) of a component in a respective mounting (fixturel or fixture2) on the working surface as well as to the onset (insert) of a construction ^¬ (partly in a particular orientation top, bottom, right, left) into another component. The above syntax is to be interpreted as follows:

Let X or X 'be a component, which may be an axle, a chassis, a pin, or a wheel. Further, let Y be a fixture for a component on the work surface, with the two fixtures fixturel and fixture2. In addition, Z refers to a direction from ^¬ selected from left, right, up and down (left, right, bottom, top).

Accordingly, pick (X) denotes the gripping of the component X, Fixate (X, Y, Z) denotes the fixation of the component X in the attachment Y on the corresponding side Z of the component X. Insert (X, X ', Z) denotes the Insertion of the component X into the component X 'on the corresponding side Z of the component X'.

The sequence of skills described above can now under control of the executive program to be performed by the D3 EE ^¬ production system by means of the three components Dl.

The embodiments of the invention described above have a number of advantages. In particular, for a production system, automatically based on a In order to ^accomplish this task, a skill-based program for producing a product must be planned. Compared to conventional systems, the production system determines computer-aided its own behavior for a specific production task, without having to deposit in advance fixed In ^¬ struktionen to perform the task in the system.

In the event of an error in the execution of the production process, the system may self-recover due to monitoring by the execution program. In addition, skill-based programs are very robust and flexible and can handle variations and errors, e.g. by feedback of the state of the system via sensors.

Furthermore, the effort of the design (so-called engineering) of a production system is reduced by the invention. Insbesonde ^¬ re can be resorted to several times usable skills, which cost in the design of the production system can be saved in planning the production process.

Through the representation of knowledge about semantic information, an inference for the derivation of new knowledge can be dynamically performed and knowledge queried dynamically. With the use of known semantic standards also special Wis ^¬ sen can from external sources easily integrated into the knowledge base ^¬ to.

Claims

claims

1. A method for computer-aided control of a production process for producing a composite of a plurality of partial products the product in a production system (SYS) of a plurality of technical components (Dl, D2, D3) which are involved in the production process, wherein in a Wissensba ^¬ sis (KB) in digital form the current state of the producti ^¬ onssystems (SYS) as well as semantic information (ASM, SKM, POM, PRM) are hiterlegt, wherein the semantic Informatio ^¬ NEN (ASM, SKM, POM, PRM) data on the production system (SYS), via basis functions (SK), which as encapsulated Pro ^¬ programs by the components (Dl, D2, D3) in Produktionssys ^¬ system (SYS) are executable on processed in the production process beitbare partial products and the production process execute ^¬ bare production steps and their dependencies comprising the following steps:

a) reading a digital specification (SPEC) for the production of a predetermined product;

b) scheduling a production process for the predetermined product on the basis of the read-in specification (SPEC) using a planning program (PE) and an inference program (RE), the planning program (PE) defining the production process in the form of a sequence of basis functions (SK) using information generated by the inference program (RE) via inference from the knowledge base (KB); c) controlling the execution of the sequence of basic functions (SK) by an execution program (EE).

2. The method according to claim 1, characterized in that between step a) and step b) is checked whether by means of the knowledge base (KB), a production process for the production of the predetermined product according to the read Specification (SPEC) is plannable.

3. The method according to claim 1 or 2, characterized in that in step b) derived via inference information ^¬ (DI) in the knowledge base (KB) are stored.

4. The method according to any one of the preceding claims, characterized in that the semantic information (PRM) on the executable in the production process production ^¬ steps and their dependencies by one or more process graphs (PG) of nodes (NO, Nl, ..., N9 ) and edges (ED).

5. The method according to claim 4, characterized in that at least one process graph (PG) is designed such that each node (NO, Nl, ..., N9) represents a production step and the nodes (NO, Nl, .. ., N9) through directed Kan ^¬ th (ED) are connected to each other, wherein the directed edges (ED) describe a partial order such that the production step of a node (NO, Nl, ..., N9), in a directional Edge (ED), according to the production step of the node (NO, N1, ..., N9) from which the directional edge (ED) originates, and the production step of a node (NO, N1, ... , N9), from which a directed edge (ED) results, is to be executed before the production step in whose node (NO, N1, ..., N9) the directional edge (ED) runs.

6. The method according to claim 4 or 5, characterized in that one or more of the process graphs (PG) allow variable partial products and / or define one or more of the process graphs (PG) predetermined sub-products.

7. The method according to any one of the preceding claims, characterized in that the semantic information (SKM) via a respective basic function (SK) include:

Parameters required for execution of the respective basic function (SK); Preconditions which the state of the production system (SYS) must fulfill for the execution of the respective base function (SK);

- Effects of the execution of the respective basic function (SK), which are conditions that satisfies the condition of the production ^¬ system (SYS) according to the execution of the respective base function (SK);

- Performance characteristics of the respective basic function

(SK).

8. The method according to any one of the preceding claims, characterized in that a respective base function (SK) comprises an interface over which the execution program (EE) can make the execution of the respective base function (BS) and the running time state of the execution of the respective base function (SK) can retrieve.

9. The method according to any one of the preceding claims, characterized in that the execution program (EE) monitors the execution of the sequence of basic functions (SK) and stores the changing states of the production system (SYS) in the knowledge base (KB), wherein the Ausführungspro ^¬ grams (EE) in the event of disturbances in the execution of the sequence of basis functions (SK) causes the rerun of the steps b) and c)

10. The method according to any one of the preceding claims, characterized in that the read-in specification (SPEC) comprises an input or edited by a user via a user interface information.

11. A device for computer-aided control of a production process for producing a product composed of several partial products in a production system (SYS) of several technical components (Dl, D2, D3), which are involved in the production process, wherein in the pre ^¬ direction a knowledge base (KB ), in which in digital form the current state of the production system (SYS) and semantic information (ASM, SKM, POM, PRM) are heaped, the semantic information (ASM, SKM, POM, PRM) data on the production system (SYS), on basis functions (SK), which as encapsulated programs by the components (Dl, D2, D3) are executed in the production system include about processable in the production process of partial products and the production process executable Produktionsschrit ^¬ te and their dependencies, the apparatus comprising a reading unit, an inference engine (RE), a planning ^¬ program (PE ) and an execution program (EE) with which the following steps can be carried out:

a) reading a digital specification (SPEC) for the production of a predetermined product by the

Reading unit;

b) scheduling a production process for the predetermined product on the basis of the read-in specification (SPEC) using the planning program (PE) and the inference program (RE), wherein the planning program (PE) the production process in the form of a sequence of basis functions (SK) generated using information derived by inference from the knowledge base via the inference program (RE);

c) controlling the execution of the sequence of basic functions (SK) by the execution program (EE).

12. The device according to claim 11, characterized in that the device is arranged for carrying out a method according to one of claims 2 to 10.

A computer program having a program code for carrying out a method according to any one of claims 1 to 10, when the program code is executed on a computer.

14. A computer program product having a program code stored on a machine carrier for performing a method according to any one of claims 1 to 10 when the program code is executed on a computer.