DE102021110469A1 - Method and system for selecting a process control instruction using a computer unit using fuzzy logic - Google Patents

Abstract

The invention relates to a method and system for selecting a process control instruction using a computer unit using fuzzy logic, a plurality of different actual process parameters being stored in the computer unit and a process control instruction being assigned to each actual process parameter. The invention also provides for entering a plurality of process parameters representing different basic process conditions into the computer unit and using fuzzy logic to determine a target process parameter based on one of the process parameters as a reference variable and as a function of at least the other process parameters, using linguistic Fuzzy sets stored in variables, which subdivide each of the process framework conditions as linguistic input variables into different classes, whereby a respective class affiliation is determined for each process parameter using fuzzy logic and based on the reference variable and depending on the respective class affiliation of each process parameter for determining the target process characteristic is weighted according to a set of rules stored in the computer unit, in particular a set of rules that can be adapted to the reference variable. From the plurality of process control instructions, the process control instruction is then selected whose assigned actual process characteristic value comes closest to, in particular corresponds to, the determined target process characteristic value.

Description

The invention relates to a method for selecting a process control instruction using a computer unit using fuzzy logic, particularly suitable for fuzzy-based selection control of the assignment of a process model on the basis of competing default parameters, and a suitable system for selecting a process control instruction.

Every project that has a project goal, such as the development of new software, a new car, a new building, a new factory, a new production line or a new automation process, to name just a few examples, starts with an analysis of the framework conditions and the Selection of a process model to be used as a basis for the implementation of the project. The process model serves as a guide or management instruction in the project in order to be able to carry out the various complex tasks in high quality and in the appropriate order. There are a number of process models or procedure models, e.g. B. the "waterfall model", the "spiral model", the "V model", the "OEP model", the "TDD model", the "MDD model", the "incremental process model", the "BDD model", the "prototyping model", the "XP model", the "RUP model", the "serum model" or the "Kanban model", to name just the best known.

With so many options, it is usually difficult to choose the right model for a particular project. This is also made more difficult by the fact that the distinction between the individual process models is sometimes only slight and the entire course of a project can often not be fully or only rudimentarily estimated at the beginning.

In addition, when selecting the process model, the respective quality characteristics of the result to be developed must usually also be taken into account. For example, in the case of software to be developed, according to the ISO 25000 valid on the filing date, these are in particular performance, compatibility, usability, reliability, security, maintainability, portability ( Portability), Functional Suitability. Depending on the place of use and other framework conditions, these quality features or others must be taken into account differently and their characteristics also influence the process model.

If an unfavorable process model is selected, the system development can be delayed or the project can fail completely. Extensive technological knowledge is therefore required for the correct selection of the process model, which enables the correct differentiation of the process models, with all possible process models, in particular the process models mentioned above, being known in the rarest of cases, and a sensible selection can consequently be made. Choosing an unsuitable process model can result in economic damage, and choosing a completely wrong process model can also cause the project to fail overall.

The object of the invention is to provide or define a system and a method for the automatic, also adaptive selection of a process control instruction, which in particular also enables the assignment of a process model for a selected project on the basis of definable project requirements and a wide variety of project goals.

The solution to the problem is given by a method and a fuzzy control system having the features of the appended claims 1 and 5, respectively. Expedient configurations are the subject matter of the respective dependent claims.

The invention is also achieved by using a fuzzy control system according to claim 9 and a computer-readable medium which according to claim 10 includes instructions for carrying out the method.

Accordingly, the invention proposes a method for selecting a process control instruction using a computer unit using fuzzy logic, in which a plurality of stored or storable, different actual process parameters are each assigned a process control instruction and based on one of a plurality of input or inputtable process parameters as setpoint and as a function of at least the other process parameters, a target process parameter is determined, with each process parameter for determining the setpoint process parameter being weighted according to a set of rules stored in the computer unit, in particular a set of rules that can be adapted to the setpoint, and the process control instruction from the plurality of stored Process control instructions is selected whose assigned actual process characteristic value comes closest to the determined target process characteristic value, in particular corresponds to it.

By means of the fuzzy logic applied here according to the invention, the process framework conditions can consequently be subdivided into different classes on the basis of stored fuzzy sets described as a linguistic input variable, a respective class affiliation is determined for each process parameter and the weighting is carried out based on the reference variable and depending on the respective class affiliation.

Accordingly, the invention also proposes a fuzzy control system for selecting a process control instruction with a computer unit that uses fuzzy logic, in which fuzzy sets are stored for describing linguistic variables, which subdivide different process framework conditions into different classes as linguistic input variables, and a set of rules is stored , in particular a set of rules that can be adapted to a reference variable.

The control system is set up here to store a plurality of different actual process parameters and to assign one process control instruction to each actual process parameter, and also to use a process parameter from a plurality of process parameters that can be entered and represent the various basic process conditions as a reference variable . In addition, the control system is set up to determine a target process characteristic using fuzzy logic based on the process parameter used as a reference variable and as a function of at least the other process parameters, specifically by determining a respective class affiliation for each process parameter using the stored fuzzy sets and based on this on the reference variable and depending on the respective class affiliation for weighting each process parameter to determine the target process characteristic according to the set of rules stored in the computer unit, wherein the control system is also set up to select the process control instruction from the plurality of stored process control instructions, the associated actual Process parameter the determined target process parameter comes closest, in particular corresponds.

Significant advantages are therefore that the use of fuzzy logic makes it possible for the first time to select a suitable process control instruction, even on the basis of incomplete and inaccurate information, in particular also on the basis of competing and/or dependent process parameters, and thus in particular to select one that is then considered a suitable , process model described at the beginning can serve as a guide or management instruction in a project. This information and in particular also mutually competing and/or dependent process parameters are evaluated using fuzzy logic, i.e. in particular the fuzzy sets and the stored set of rules, and the optimal process control instruction is then determined from a plurality of stored process control instructions. As a result, the appropriate process control instruction can be determined at the beginning of a process, even if some process parameters are not yet known.

As a result of the possible use of a wide variety of process parameters, the necessary process parameters for determining a suitable process control instruction can also be identified depending on the process.

With the invention, a tool is consequently made available with which a suitable process control instruction can be automatically determined, in particular selected, on the basis of process parameters that are respectively present for different process framework conditions.

The fuzzy logic, i.e. in particular the fuzzy sets and the stored set of rules, can be suitably adapted to specific processes and the specific framework conditions associated with them by changing the variables in each case. By using the technology of fuzzy logic, an individual selection of a suitable process control instruction, in particular a suitable process model based on the respective existing process infrastructure, is thus made possible in an automated manner for the first time, even with a dynamically changeable process infrastructure, in particular one that can be specified by the framework conditions.

Further advantages and features of the invention will become apparent from the following description of preferred exemplary embodiments with reference to the accompanying drawings.

• 1 eine verallgemeinert dargestellte Prinzipskizze einer eines Fuzzy-Reglers als Basis für eine Fuzzylogik anwendende Rechnereinheit des erfindungsgemäßen Fuzzy-Regelsystems,
• 2 bevorzugte Eingangsgrößen für eine erste Fuzzy-basierte Komponente der Rechnereinheit gemäß erfindungsgemäßen Fuzzy-Regelsystem,
• 3a, 3b bevorzugte Eingangsgrößen für eine zweite und dritte Fuzzy-basierte Komponente der Rechnereinheit gemäß erfindungsgemäßen Fuzzy-Regelsystem,
• 4 eine vereinfacht dargestellte Skizze eines Fuzzy-Regelsystems mit fünf Fuzzy-basierten Komponenten,
• 5 bis 18 am Beispiel eines ein Projektziel betreffendes Projektes als ein im Rahmen der Erfindung möglicher Prozess zweckmäßige Fuzzy-Sets umfassend eine Variable Projektumfang, eine Variable Projektbedingen, eine Variable Bekanntheit der Anforderungen, eine Variable Stabilität, eine Variable Dokumentation, eine Variable Vorhandene Dokumentation, eine Variable Organisation, eine Variable Räumliche Verteilung, eine Variable Team Erfahrung, eine Größe des Teams, eine Expertenwissen, eine Variable Kohäsion der Komponenten, eine Variable Wirtschaftliche Kritikalität und eine Variable Sicherheitsanforderung als Beispiele von linguistisch beschriebenen, einem jeweiligen Prozess als Prozessrahmenbedingungen zuweisbaren Variablen zur Veranschaulichung verschiedener Ausführungsformen eines erfindungsgemäßen Fuzzy-Regelsystems,
• 19, 20 zweckmäßige Fuzzy-Sets in Bezug auf eine Variable Ausgangswert Inkrementell und eine Variable Ausgangswert Iterativ als weitere Beispiele von linguistisch beschriebenen, einem jeweiligen Prozess zuweisbaren Variablen zur Veranschaulichung verschiedener Ausführungsformen eines erfindungsgemäßen Fuzzy-Regelsystems,
• 21 ein Abbildungsschema betreffend ermittelte Soll-Prozesskennwerte und hinterlegte Ist-Prozesskennwerte für hierauf basierte Auswahl einer Prozessführungsanweisung, und
• 22, 23 Prinzipskizzen bevorzugter Ausführungsbeispiele eines erfindungsgemäßen Verfahrens zur Auswahl einer Prozessführungsanweisung mittels einer besonders zweckmäßigen Ausgestaltung eines Fuzzy-Regelsystems.

In the drawings show:

• 1 a generalized schematic diagram of a computer unit of the fuzzy control system according to the invention that uses a fuzzy controller as a basis for fuzzy logic,
• 2 preferred input variables for a first fuzzy-based component of the computer unit according to the fuzzy control system according to the invention,
• 3a , 3b preferred input variables for a second and third fuzzy-based component of the computer unit according to the fuzzy control system according to the invention,
• 4 a simplified sketch of a fuzzy control system with five fuzzy-based components,
• 5 until 18 using the example of a project relating to a project goal as a possible process within the scope of the invention, useful fuzzy sets comprising a variable project scope, a variable project conditions, a variable knowledge of the requirements, a variable stability, a variable documentation, a variable existing documentation, a variable organization , a variable spatial distribution, a variable team experience, a size of the team, an expert knowledge, a variable cohesion of the components, a variable economic criticality and a variable safety requirement as examples of linguistically described variables that can be assigned to a respective process as process framework conditions to illustrate different embodiments a fuzzy control system according to the invention,
• 19 , 20 expedient fuzzy sets in relation to an incremental output value variable and an iterative output value variable as further examples of linguistically described variables that can be assigned to a respective process to illustrate different embodiments of a fuzzy control system according to the invention,
• 21 a mapping scheme relating to determined target process parameters and stored actual process parameters for the selection of a process control instruction based on this, and
• 22 , 23 Schematic sketches of preferred exemplary embodiments of a method according to the invention for selecting a process control instruction using a particularly useful embodiment of a fuzzy control system.

1 shows a generalized outline of a fuzzy controller, the basic principle of which is also used for a computer unit using fuzzy logic within the scope of the invention. As is known, fuzzy controllers belong to the class of map controllers, which correspond to the theory of fuzzy logic. Usually, three sub-steps are carried out in each control step, i.e. the step of fuzzification, inference and finally the step of defuzzification, with the individual inputs and outputs being described as linguistic variables, each of which includes fuzzy sets.

The fuzzy sets, in turn, are formed from a set of membership functions, also called fuzzy sets. These membership functions or fuzzy sets are usually preconfigured and can, for example, take the form of a triangle, trapezium or curve (cf. 5 until 20 ), whereby in this regard and for the documentation of the specialist knowledge of the person skilled in the art to be applied here, the full extent to the published application DE 10 2013 008 151 A1 and the one mentioned there DE 195 02 230 A1 is referenced. The fuzzy sets classify the linguistic input variables into "fuzzy" classes and assign a class membership value of between 0% and 100% to the linguistic input variables corresponding to measurements or "sharp" input variables obtained in some other way.

Correspondingly, the “sharp” output values to be output for a linguistic output variable can be based on membership functions, and can have a membership value in relation to the fuzzy classification from 0% to 100%. A fuzzy controller thus generally determines a manipulated variable to be output from the control difference, ie from the difference between a desired value and an actual value, by processing the steps of fuzzification, inference and defuzzification. In fuzzification, fuzzy sets assigned to the input variables described by means of linguistic variables are evaluated and values for class memberships are determined. In the inference step, the rules for calculating the output are specified (set of rules) based on the values for class memberships, and this also determines a weighting of the variables included in the calculation. In the defuzzification, the output values described by means of linguistic variables are assigned to certain membership functions based on the rules for the calculation and weighting and are determined according to a membership value. A fuzzy controller is therefore a non-linear map or state controller and basically has no internal dynamics.

4 shows a greatly simplified schematic diagram of a fuzzy control system which can be used to carry out a method within the scope of the invention. The fuzzy control system consequently has a computer unit that uses fuzzy logic and is appropriately based on the principles of a fuzzy controller described above. Fuzzy sets are used to describe linguistic variables, such as in the 5 until 20 stored, which process framework conditions are divided into different classes as linguistic input variables. A set of rules is also stored, which can be appropriately adapted to a reference variable, as described in more detail below.

Such a fuzzy control system can in particular by combining suitable hardware and be given software. Correspondingly, the computer device set up according to the invention can also be given in particular by a combination of suitable hardware and software. For example, individual components can also run as software on existing hardware, for example an existing processor. However, hardware can also be provided specifically for individual components. A computer program which includes program code means for carrying out the steps of the method according to the invention can also be provided for the procedure according to the invention, the computer unit and/or for the fuzzy control system according to the invention. A computer-readable medium, e.g. a CD-ROM or DVD or a USB or flash memory, can also be used for this purpose, which includes instructions which, when executed on a computer with at least one computer unit applying fuzzy logic, cause the computer to carry out the method according to the invention. However, such a computer-readable medium does not have to be exclusively a physical medium, but can also be present, for example, in the form of a data stream and/or a signal that represents a data stream.

That with 4 Outlined fuzzy control system has a number of fuzzy-based components for the selection of a process control instruction, in the example shown a fuzzification of a first to n-th input variable, in particular in the 2 , 3a and 3b carry out the input variables shown and implement the subsequent inference and defuzzification for the selection of the process control instruction as the output variable, in the example shown for the determination of a process model as the output variable.

The fuzzy control system is set up in particular so that a plurality of process parameters representing different basic process conditions can be entered as input variables. Within the scope of the invention, process framework conditions can in particular include the respective process environment, the process team and/or the process criticality 4 referred to as project, team or criticality. The respective process parameters can then be defined as input variables for these or other groups and then evaluated and combined. The result of the groups is in turn combined using the fuzzy logic and the stored set of rules and an initial value is determined. With the help of this initial value, the suitable process control instruction is selected, ie in particular one that can then serve as a suitable, initially described process model in a project as a guide or control instruction. Specific process framework conditions can consequently be adjusted by changing the input variables or selecting other input variables.

the 2 shows in this context preferred process parameters within the scope of the invention as input variables for a first fuzzy-based component relating to the process environment, in particular for determining a characteristic value Pu-K related to the process environment, ie in particular for the characteristic of a process, eg a specific project. the 3a shows preferred process parameters as input variables for a second fuzzy-based component relating to the process team, in particular for determining a characteristic value Pt-K related to the process team, ie in particular for the characteristic of a team used or required for the implementation of the process. 3b shows preferred process parameters as input variables for a third fuzzy-based component relating to the process criticality, in particular for determining a parameter Pk-K related to the process criticality, ie in particular for the characteristic of the criticality in relation to the process to be implemented.

• Projektumfang: Der Projektumfang wird z.B. bestimmt durch die Projektkosten, die Entwicklungsdauer (Zeit), die Komplexität, Qualität und Anzahl der zu realisierenden Funktionen. Ein entsprechendes, zur Beschreibung der diesbezüglichen linguistischen Variablen hinterlegtes Fuzzy-Set ist z.B. mit 5 skizziert.

So can according to 2 preferred process parameters relating to the process environment as input variables include:

• Project scope: The project scope is determined, for example, by the project costs, the development period (time), the complexity, quality and number of functions to be implemented. A corresponding fuzzy set stored for the description of the relevant linguistic variables is, for example, 5 sketched.

Project conditions (especially internal project conditions): The project conditions are determined, for example, by the characteristics of the customer and the stakeholders (eg the legislature, works council, etc.), as well as the contractual conditions. With a linguistic variable, as through the with 6 For example, the customer's involvement in the project can also be taken into account in the fuzzy set outlined. The more frequently the customer is involved in the project, the more unstable the project conditions can be. Consequently, stability or instability caused by project-internal framework conditions can also be taken into account.

Awareness of the requirements: This can be used to express, for example, the extent to which the requirements for the finished product are known. With the related linguistic variable, as by the with 7 reproduced in the outlined fuzzy set, the quality of the formulated requirements is also included. In the case of incomplete and contradictory requests a lower value should be selected. If complete and specific requirements are available, then a high value should be selected for this variable. The relevant process parameter, which can in particular be input, can consequently also expediently represent a desired process result.

Stability (especially external reasons): In a dynamic environment (economy, political and social changes) with many changes, the environment is usually more unstable and a corresponding value should be selected for this variable. This variable also describes the expected stability of the requirements. With a related linguistic variable, as by the with 8th If the fuzzy set is reproduced, stability or instability caused by external framework conditions can also be taken into account.

Documentation: With one of these relevant process parameters, for example, by means of a like with 9 outlined fuzzy sets are described via a corresponding linguistic variable of the importance and scope of the documentation in the project.

Existing documentation: As a rule, follow-up projects have documentation that varies in scope and quality. It must therefore usually also be possible to take these into account during the course of the project and thus also determine the possible process models. The more extensive and specific the documentation from old projects is, the greater its influence on the selection of the process model should be. A corresponding fuzzy set stored for the description of the relevant linguistic variables is, for example, 10 sketched.

• Organisation: Dieser Parameter bzw. die entsprechende linguistische Variable (vgl. 11) ist der Gruppe Team zugeordnet und beschreibt die Eigenschaften des Teams beschreibt. Je homogener das Fachwissen im Team (so sind z.B. ein Entwickler Team und ein Test Team unterschiedliche Teams und/oder ein Team kann in verschiedenen Organisationseinheiten der Firma angesiedelt sein) desto spezifischer ist in der Regel die Organisation des Teams und ein entsprechender Wert anzugeben. In einem interdisziplinären Team arbeiten z.B. Tester und Entwickler zusammen in einem Team.

Preferred process parameters affecting the process team are at 3a shown and can include the following process parameters as input variables:

• Organization: This parameter or the corresponding linguistic variable (cf. 11 ) is assigned to the Team group and describes the properties of the team. The more homogeneous the expertise in the team (e.g. a development team and a test team are different teams and/or a team can be located in different organizational units of the company), the more specific the organization of the team and a corresponding value should be specified. In an interdisciplinary team, for example, testers and developers work together in one team.

Spatial distribution of the team: A team can work in one organizational unit and still be spatially separated. In the case of physical distance, communication within the team is impaired and can therefore be appropriately taken into account when selecting the process control instruction. With the help of such a variable (cf. also 12 ) a corresponding value can therefore be specified for this.

Team experience: If a team already knows many different methods for managing a process, e.g. for developing software, and has already worked with them, the team experience is usually higher (cf. also 13 ) to estimate.

Size of the team: A related to this variable (cf. also 14 ) applicable process parameter is related to the size of the project, for example: the larger the project, the larger the team. The smaller the team, the more agile the team can work.

• Expertenwissen: Je spezifischer das erforderliche Expertenwissen (z.B. Spezialkenntnisse bezüglich des Einsatzgebietes für ein zu entwickelndes Produkt) desto höher ist die Kritikalität für eine diesbezügliche linguistische Variable (vgl. auch 15) einzuschätzen.

Preferred process parameters relating to the process criticality are at 3b shown and can include the following process parameters as input variables:

• Expert knowledge: The more specific the required expert knowledge (e.g. special knowledge regarding the area of application for a product to be developed), the higher the criticality for a relevant linguistic variable (cf. also 15 ) to estimate.

Cohesion of the components: If many partner systems and components have to be taken into account, the higher the cohesion of the components in the project (cf. also 16 ). A high degree of cohesion and dependency of the components requires in particular classic process models as management instructions.

Economic criticality: The function of a product can also have different effects on the profitability of a company. An online shop often has a higher criticality than a website with information from a canteen, for example (cf. also 17 ).

Safety requirements: Using a process parameter representing the safety requirements, the safety requirements for a process, in particular a project, and/or a product to be developed can be assessed using the corresponding variable (cf. also 18 ). The area of application for the product can also be taken into account. It can be considered, for example, that a calendar in a security critical physical environment must also meet the requirements of this environment.

The control of the system according to the invention is expediently set up in such a way that one of the process parameters representing the various basic process conditions is used as a reference variable, in particular a process parameter which represents a desired process result and can preferably be entered individually is used. Because if the desired process result is e.g. The “awareness of the requirements” described above can serve as such a process parameter. Depending on the process and/or basic process conditions, other process parameters to be used as reference variables can preferably also be selected within the scope of the invention, which is why the system expediently has a selection device set up for this purpose.

Based on the process parameter used as a reference variable and depending at least on the other process parameters, the control system according to the invention is then also set up to determine a target process parameter using fuzzy logic, specifically by determining, using the stored fuzzy sets, a respective class membership for each process parameter and based on the reference variable and depending on the respective class affiliation for weighting each process parameter to determine the target process characteristic according to the set of rules stored in the computer unit. The process parameter used as a reference variable can also be used as one of the other process parameters and/or the stored set of rules can be adapted to the respective reference variable, in a preferred embodiment also adapting itself automatically to the respective reference variable.

The target process characteristic makes it clear how the process control instruction to be selected should be designed, e.g. how incremental and iterative a process model must be designed.

As in 23 For this purpose, the control system is expediently set up for storing a plurality of different actual process parameters 1...N and for assigning one process control instruction 1...N to each actual process parameter 1...N, so that, particularly within the framework of Defuzzification (cf. 4 ), the process control instruction, in particular the process model, can be selected from the plurality of stored process control instructions whose associated actual process characteristic value comes closest to the determined target process characteristic value, in particular corresponds to it.

In a particularly preferred embodiment, the computer unit with the stored set of rules for determining the target process parameter is also set up to first carry out two different weightings when weighting the process parameters and then initially a first process part parameter 4 denoted by process value incrementally, and at least one second process part characteristic value 4 referred to as process value iteratively. For each determined process part characteristic value, a linguistic process control variable subdivided into different classes is again appropriately kept in the computer unit, for example with the 19 and 20 outlined fuzzy sets, so that using the fuzzy logic to determine a respective class membership for each process part parameter and based on the reference variable and depending on the respective class membership each process part parameter to determine the target process parameter is to be weighted again according to the set of rules stored in the computer unit before the process control instruction is selected from the plurality of process control instructions whose associated, stored actual process characteristic value comes closest to, in particular corresponds to, the determined target process characteristic value. 21 shows an example of a mapping scheme relating to the weighting of two process part parameters, one iterative and one incremental, determined target process parameters (in the table according to 21 ) and stored actual process parameters (below the table according to 21 ) with respectively assigned process control instructions for the selection of a suitable process control instruction based on this. Respective different process parameters are attached 21 illustrated by different hatchings.

If the weighting of the two process part characteristic values leads to a target process characteristic value that lies between two actual process characteristic values, for example, and in particular therefore actually lies between different process control instructions, a combination of the different process control instructions, e.g. the combination of two process models, can also be carried out in a preferred embodiment appropriate selection is recommended. For example, agile methods (planning poker or backlog) can be used in a classic process environment, especially project environment. Process control instructions, such as V-model and spiral model as exemplary process models, can also be combined in a preferred manner and their advantages can thus be used.

By means of a particularly useful embodiment of a fuzzy control system is based on the 22 and 23 Schematic sketches of preferred exemplary embodiments shown once again illustrate a preferred method according to the invention for selecting a process control instruction in the process, in particular using the example of the selection of a process model for a selected project, with a project goal that is at least largely defined, for example, a desired process result, in particular via the "awareness of the requirements", representing process parameters as reference variable.

The fuzzy control system is for in the 22 and 23 The principle sketches shown are therefore set up in particular with a first fuzzy logic software application running in the computer unit that uses the fuzzy logic, based on the reference variable and as a function of process parameters that represent the process environment-based, process team-based and process criticality-influencing process framework conditions, a respective characteristic value related to the process environment Pu-K, a parameter Pt-K related to the process team and a parameter Pk-K related to the process criticality. Furthermore, the fuzzy control system is set up, in particular with a second fuzzy logic software application running in the computer unit applying the fuzzy logic, based on the reference variable and depending on the characteristic value Pu-K related to the process environment, the characteristic value Pt-K related to the process team and of the parameter Pk-K related to the process criticality, first to determine a first process part parameter Ptk-In representing the relevance of an incrementally aligned process control instruction and a second process part parameter Ptk-It representing the relevance of an iteratively aligned process control instruction, and eg with a in which the fuzzy logic third fuzzy logic software application running in the third fuzzy logic software application used by the computer unit, based on the reference variable and as a function of the first process part parameter and the second process part parameter, to determine the target process parameter.

In particular, fuzzy-based components can be used to determine the respective characteristic values, the modes of action of which are the same as those previously described in relation to the 2 , 3a , 3b and 4 correspond to the components described.

After storing a plurality of different actual process parameters 1...N, a process control instruction 1...N can be assigned to each of these in the computer unit ( 23 ). Process control instructions 1 . . . N assigned to the actual process parameters 1 . can, such as those according to 21 process models listed below the table.

Depending on the specific process, a plurality of process parameters representing different basic process conditions can be entered into the computer unit and one of these process parameters can be selected as the reference variable. As described above, the process framework conditions are usually specified in particular by a specific process environment, a specific process team and a specific process criticality. The process parameters representing these process framework conditions can consequently be subdivided or grouped in particular into process environment-based, process team-based and process criticality-influencing parameters.

The computer unit then uses fuzzy logic to determine a target process parameter based on the process parameter used as a reference variable and as a function of at least the other process parameters, and uses this to select the process control instruction, with this being done in particular by using fuzzy logic technology, i.e. by the processing of steps of fuzzification, inference and defuzzification characterizing a respective fuzzy logic takes place. The fuzzy logic-based regulation is consequently configured for the first time for the selection of process instructions, in particular also process models, such that a suitable process instruction, in particular a suitable process model, is selected, in particular determined, in a short time on the basis of imprecise and incomplete process parameters.

According to the invention, on the basis of fuzzy sets stored for describing linguistic variables, which subdivide each of the process framework conditions as linguistic input variables into different classes, a respective class affiliation is determined for each process parameter using fuzzy logic and based on the reference variable and depending on the respective class affiliation at least everyone the further process parameters for determining the setpoint process characteristic value are weighted according to a set of rules stored in the computer unit, in particular a set of rules that can be adapted to the reference variable. Finally, from the plurality of process control instructions, the process control instruction whose associated actual process characteristic value comes closest to, in particular corresponds to, the determined target process characteristic value is selected.

Based on the subdivision or grouping shown above into process environment-based, process team-based and process criticality-influencing parameters, a respective characteristic value Pu-K related to the process environment, to the process team-related characteristic value Pt-K and a the process criticality-related parameter Pk-K can be determined.

In addition or as an alternative, a first process part parameter Ptk-In representing the relevance of an incrementally aligned process control instruction and a second process part parameter Ptk-It representing the relevance of an iteratively aligned process control instruction can also be determined as part of the weighting. As a result, at least two different weightings are carried out when weighting the process parameters, appropriately in accordance with the set of rules stored in the computer unit, with a linguistic process control variable also being kept available in the computer unit for each determined process part characteristic value and broken down into different classes and using fuzzy logic, a respective one Class membership determined for each process part parameter and based on the reference variable and depending on the respective class membership of each process part parameter to determine the target process parameter according to the set of rules stored in the computer unit is again weighted.

The characteristic value Pu-K related to the process environment, the characteristic value Pt-K related to the process team and the characteristic value Pk-K related to the process criticality are also particularly preferred for determining such process part characteristic values Ptk-In, Ptk-It of a correspondingly configured value Fuzzy-based component supplied as input variables.

The properties of the respective process, in particular of the respective project, consequently determine the input parameters for the fuzzy-based control. In particular, the characteristic of the process, in particular of the project, is consequently determined by a series of framework conditions, based on which in turn the input parameters for the fuzzy-based control are made available.

For example, if the process relates to the construction of an industrial hall, this represents a complex project. The manufacturer often has clear ideas/requirements as to which functions the industrial hall must fulfil, even before construction begins. Depending on the manufacturing process, the industrial hall must meet different lighting conditions, temperatures, power supply, resistance to chemicals, transportation of goods, communication, data collection, etc. In the industrial hall there should also be different areas with different requirements. These specifications can be clearly formulated by the client and these specifications / requirements will only have to be adjusted slightly during the construction of the hall.

Nevertheless, the general conditions for the construction of the hall can change, for example if the soil condition was incorrectly assessed and further measures to stabilize the subsoil are now required. There may also be changes in the suppliers of building materials. Based on this scenario, the input parameters for the fuzzy-based control for the selection of a process control instruction can now be determined and weighted by appropriate input.

• • Projektumfang: hoch 0,3; sehr hoch 0,8;
• • Projektbedingungen (intern): stabil 1,0; sehr stabil 0,3;
• • Bekanntheit der Anforderungen: mittel 0,3; hoch 1,0;
• • Stabilität (extern): instabil 0,75; stabil 0,7;
• • Dokumentation: hoch 1,0; sehr hoch 0,3;
• • Vorhandene Dokumentation: gering 0,7; normal 0,7;

With regard to the process environment, for example:

• • Project scope: to the power of 0.3; very high 0.8;
• • Project conditions (internal): stable 1.0; very stable 0.3;
• • Familiarity with the requirements: medium 0.3; to the power of 1.0;
• • Stability (external): unstable 0.75; stable 0.7;
• • Documentation: to the power of 1.0; very high 0.3;
• • Existing documentation: low 0.7; normal 0.7;

• • Organisation: sehr interdisziplinär 0,4; interdisziplinär 0,8;
• • Räumliche Verteilung: mittel 1,0; entfernt 0,2;
• • Team Erfahrung: mittel 0,35; hoch 1,0;
• • Größe des Teams: groß 1,0; sehr groß 0,2;

Regarding the process team, for example:

• • Organisation: very interdisciplinary 0.4; interdisciplinary 0.8;
• • Spatial distribution: average 1.0; removed 0.2;
• • Team experience: medium 0.35; to the power of 1.0;
• • Size of the team: large 1.0; very large 0.2;

• • Expertenwissen: interdisziplinär 1,0;
• • Kohäsion der Komponenten: hoch 1,0; sehr hoch 0,4;
• • Wirtschaftliche Kritikalität: mittel 0,2; hoch 1,0;
• • Sicherheitsanforderungen: mittel 1,0;

With regard to the process criticality, for example:

• • Expert knowledge: interdisciplinary 1.0;
• • Cohesion of components: to the power of 1.0; very high 0.4;
• • Economic criticality: medium 0.2; to the power of 1.0;
• • Security requirements: medium 1.0;

Based on this, for example, an initial value between normal 1.0 and low 0.55 is determined as an incremental partial characteristic value and an initial value between normal 1.0 and low 0.5 is determined as an iterative partial characteristic value.

With that, as with 21 indicated by the dashed circle shown in the table, for the special industrial hall project, for example, the spiral model and the V-model were determined. The entire hall should therefore be built as a monolith, whereby the individual areas with different requirements can be built independently of each other. This corresponds to the spiral model. Methods of the V-model (continuous review) should be integrated in order to minimize the uncertainties of the external framework and to determine appropriate measures to react to the changed framework.

If the process concerns the development of an online shopping app, for example, which is to be developed for general use by every smartphone user, further interesting requirements often arise during the use of the first versions. There are also high security requirements, since both personal data and bank data are used.

• • Projektumfang: gering 1,0; mittel 0,2;
• • Projektbedingungen (intern): instabil 1,0;
• • Bekanntheit der Anforderungen: niedrig 1,0;
• • Stabilität (extern): instabil 1,0; sehr stabil 0,1;
• • Dokumentation: niedrig 1,0; mittel 0,3;
• • Vorhandene Dokumentation: mittel 0,4; hoch 1,0;

If, using this scenario, the input parameters for the fuzzy-based control for the selection of a process control instruction are determined and weighted by appropriate input, the following results regarding the process environment, for example:

• • Project scope: low 1.0; mean 0.2;
• • Project conditions (internal): unstable 1.0;
• • Familiarity with the requirements: low 1.0;
• • Stability (external): unstable 1.0; very stable 0.1;
• • Documentation: low 1.0; mean 0.3;
• • Existing documentation: medium 0.4; to the power of 1.0;

• • Organisation: mittel 1,0; Spezifisch 0,5;
• • Räumliche Verteilung: lokal 1,0;
• • Team Erfahrung: sehr hoch 0,1; hoch 1,0;
• • Größe des Teams: klein 1,0; mittel 0,3;

Regarding the process team, for example:

• • Organization: average 1.0; Specific 0.5;
• • Spatial distribution: local 1.0;
• • Team experience: very high 0.1; to the power of 1.0;
• • Team size: small 1.0; mean 0.3;

• • Expertenwissen: interdisziplinär 0,15; mittel 1,0;
• • Kohäsion der Komponenten: niedrig 1,0;
• • Wirtschaftliche Kritikalität: sehr hoch 0,1; hoch 1,0;
• • Sicherheitsanforderungen: hoch 1,0; sehr hoch 0,25;

With regard to the process criticality, for example:

• • Expert knowledge: interdisciplinary 0.15; medium 1.0;
• • Cohesion of components: low 1.0;
• • Economic criticality: very high 0.1; to the power of 1.0;
• • Security requirements: high 1.0; very high 0.25;

Based on this, an initial value of, for example, raised to the power of 1.0 is determined as an incremental partial characteristic value and an initial value of normal 1.0 as an iterative partial characteristic value;

This means that the online shopping app is in accordance with the special project 21 the project models MDD - Model Driven Development and TDD - Test Driven Development by means of fuzzy-based control as suitable process control instructions, in particular project models. In this case, they correspond to the predominantly incremental character and enable the further development of the app as soon as new requirements are formulated through user feedback. The low cohesion of the components is also taken into account here, which can be used to advantage, particularly in model-driven development.

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 102013008151 A1 [0020]
• DE 19502230 A1 [0020]

Claims (10)

Method for selecting a process control instruction by means of a computer unit using fuzzy logic, with the following steps: - storing a plurality of different actual process parameters in the computer unit, - Allocation in the computer unit to an actual process parameter in each case one process control instruction, - Entering a plurality of process parameters representing different process framework conditions into the computer unit, - Determination by the computer unit using fuzzy logic based on one of the process parameters as a reference variable and depending on at least the other process parameters of a target process parameter, using fuzzy sets stored for describing linguistic variables, which each of the process framework conditions as linguistic input variables in different Subdivide classes, whereby a respective class affiliation is determined for each process parameter using fuzzy logic and based on the reference variable and depending on the respective class affiliation of each process parameter for determining the target process characteristic value according to a set of rules stored in the computer unit, in particular a set of rules that can be adapted to the reference variable , is weighted, - Selection by the computer unit of the process control instruction from the plurality of process control instructions whose associated actual process characteristic value comes closest to the determined target process characteristic value, in particular corresponds to it. procedure after claim 1 , wherein at least two different weightings are carried out for the weighting of the process parameters in order to determine the target process parameter in accordance with the set of rules stored in the computer unit, and a first process part parameter and at least one second process part parameter are subsequently determined, with for each process part parameter determined in the computer unit a linguistic process control variable is kept available and divided into different classes and using fuzzy logic a respective class affiliation is determined for each process part characteristic value and based on the reference variable and depending on the respective class affiliation each process part characteristic value for determining the target process characteristic value according to the set of rules stored in the computer unit is weighted again. Procedure according to one of Claims 1 or 2 , wherein the process parameter applied as a reference variable is selected or permanently specified in the computer unit before the step of determining, in particular input process parameters representing a desired process result being used for the process parameter applied as a reference variable. Procedure according to one of Claims 1 until 3 , the process parameter used as a reference variable also being used as one of the other process parameters. Fuzzy control system for selecting a process control instruction, with a computer unit that uses fuzzy logic with stored fuzzy sets for describing linguistic variables, which subdivide different process framework conditions into different classes as linguistic input variables, and a stored set of rules, in particular a set of rules that can be adapted to a reference variable, - wherein the control system is set up to store a plurality of different actual process parameters and to assign one process control instruction to each actual process parameter, and is also set up to use a process parameter from a plurality of process parameters that can be entered and that represent the various basic process conditions as a reference variable , and - wherein the control system is also set up to determine a target process characteristic value using fuzzy logic based on the process parameter used as a reference variable and as a function at least of the other process parameters, specifically by determining a respective class affiliation for each process parameter using the stored fuzzy sets and based on the reference variable and depending on the respective class affiliation for weighting each process parameter for determining the target process characteristic value according to the set of rules stored in the computer unit, - wherein the control system is also set up to select the process control instruction from the plurality of process control instructions whose associated actual process characteristic value comes closest to the determined target process characteristic value, in particular corresponds to it. Fuzzy control system according to previous claim 5 which is also set up to automatically adapt the set of rules stored in the computer unit to the reference variable and/or which has a selection device set up to select the process parameter to be used as a reference variable. Fuzzy control system according to previous ones claims 5 or 6 , which is also set up to determine the target process characteristic value according to the set of rules stored in the computer unit, to carry out at least two different weightings when weighting the process parameters and, as a result, to first determine a first process part characteristic value and at least one second process part characteristic value, with each determined process part characteristic value in the computer unit is provided with a linguistic process control variable subdivided into different classes, and using fuzzy logic to determine a respective class affiliation for each process part characteristic value and, based on the reference variable and as a function of the respective class affiliation, each process part characteristic value for determining the target process characteristic value according to the Computing unit stored set of rules to weight again. Fuzzy control system according to one of claims 5 until 7 , which is set up, - in particular with a first fuzzy logic software application running in the fuzzy logic computer unit, based on the reference variable and as a function of process parameters which represent process environment-based, process team-based and process criticality-influencing process framework conditions, a respective related to the process environment characteristic value (Pu-K), characteristic value (Pt-K) related to the process team and a characteristic value (Pk-K) related to the process criticality, and which is further set up, - in particular with a second fuzzy logic running in the fuzzy logic applying computer unit -Software application, based on the reference variable and depending on the characteristic value related to the process environment, the characteristic value related to the process team and the characteristic value related to the process criticality, initially a first one that represents the relevance of an incrementally aligned process control instruction to determine the resulting process part parameter and a second process part parameter representing the relevance of an iteratively aligned process control instruction, and which is set up, in particular with a third fuzzy logic software application running in the computer unit that uses fuzzy logic, based on the reference variable and as a function of the first process part parameter and the second Process part characteristic to determine the target process characteristic. Use of a fuzzy control system with a computer unit that uses fuzzy logic with fuzzy sets stored for describing linguistic variables, which subdivide process framework conditions into different classes as linguistic input variables, and a stored set of rules, in particular a set of rules that can be adapted to a reference variable, for carrying out a method according to one of the Claims 1 until 4 , to select a process control statement. Computer-readable medium comprising instructions which, when executed on a computer with at least one computer unit applying fuzzy logic, cause the latter to carry out the steps of the method according to one of Claims 1 until 4 to perform.

Images