JPS6024637A - System flow generating system - Google Patents

Abstract

PURPOSE:To automate half generation of a system flow chart, and to design efficiently software by storing knowledge of an expert of a software design in a computer, and using it. CONSTITUTION:A knowledge base 2 is used for accumulating knowledge for generating a system flow chart, and stored in a storage device formed by an IC memory, etc. The knowledge base 2 is constituted of a simple system flow generating knowledge 3 and an efficient system flow generating knowledge 4. Both the generating knowledges 3, 4 are formed by inputting know-how heared out of an expert of a software design to a computer in accordance with a knowledge expression form. Also, both of generating knowledge 3, 4 break down, and are structured as the knowledge base 2 of the computer by using the lowest layer level as a knowledge source (a group of knowledge expressed in the shape of a production rule). An inference is executed by utilizing this knowledge base, a part whose knowledge is short is replenished by inquiring of a user, and a semiautomatic system flow is generated.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to a tool that supports the design of software such as computer processing and control systems, and is mainly used to design a computer processing system using a specific number of software functional components. The present invention relates to a system 70-creation method of processing using sometimes suitable knowledge.

[Background of the invention]

Conventional system flow creation methods for computer processing, etc., are created by hand, or by using a graphic editor function on a computer, etc., and humans decide on the connection structure of all software components, and then give detailed instructions on how to write it. Since this method requires drawing, it requires an experienced software designer to consider the connection structure of software components, and it also requires a large amount of time to create.

[Purpose of the invention]

An object of the present invention is to solve the above problems and provide a method for storing the knowledge of software design experts in a computer and creating a system flow plan by inference using this knowledge.

In other words, traditionally, a software designer would think of a system flow that satisfies the required system specifications and create a design drawing by hand, or a human would determine all the connection structures in detail and create a system flow using a graphic editor using a computer. The aim is to semi-automate the process of creating diagrams and streamline software design.

[Summary of the invention]

When designing software, you may organize software created in the past and know that it will consist of a specific number of software components. It has been empirically found that in the field of patch-type business calculations centered on file and report creation, control PCs, etc., systems can be constructed from several functional components. Under these boundary conditions, software designers partially define the connection relationships between software components based on logical rules (for example, a certain process or input is required in order for a certain item to appear), Partly based on empirical rules, system 70-
is being created.

This software designer's knowledge is expressed in the form of production rules, and inferences (
By pattern matching (symbol string pattern matching), software functional components are searched for and combined to create a system flow that meets the required system specifications.

[Embodiments of the invention]

Hereinafter, the present invention will be explained in detail with reference to Examples.

First, the configuration of the present invention will be explained with reference to FIG.

The knowledge base 2 stores knowledge for creating a system flow diagram, and is stored in a storage device such as an IC memory. The knowledge base 2 is composed of simple system flow creation knowledge 3 and efficient system flow creation knowledge 4. Simple system flow creation knowledge 3 and efficient system 7
01 Creation Knowledge 4 is the knowledge developed from software design knowledge (-) that is put into a computer according to the knowledge expression format. Simple System Flow Creation Knowledge 3
and efficient system flow creation knowledge 4 are broken down as shown in Figure 2, and the lowest level is divided into knowledge 7-s (7).
``a group of knowledge expressed in the form of reduction rules)
This will be constructed as a computer knowledge base 2. Each knowledge source is created with the following structure (8 LISP prefixes).
(represented by the formula).

((Knowledge source startup prerequisites) (Knowledge source common information) ((Production Runil 1) (Rule 1 (IF (Prerequisite Clause) (Prerequisite Clause)) (THEN (PROPO8E (Action) (SELi
! , CT-INF prerequisites) (PR6PO8E...
...))) (Production rule n))) The knowledge source consists of knowledge source activation preconditions, knowledge source common information, and production rule #. The blackboard 5 stores the state during the inference. It is a storage device. The frame 6 is a memory table for storing the combined architecture of the system flow that has been inferred using the knowledge base 2, and uses a frame-type knowledge representation of knowledge engineering. The terminal device 7 is used to apply the required system specifications of the system to be created manually and from inference. terminal 4&ms
is used to plot system flow diagrams on a graphical display. The inference control information area 9 is used to determine which knowledge source is to be incremented next.

Next, the self-answers of Knowledge Base 2 will be explained.

In the batch type business calculation system handled in this implementation series, the 61ii! Functional parts of d (
File conversion second check, collation, update.

Extraction, report creation), and it has been empirically known that the coupling relationships between each functional component are expressed as shown in Figure 3 ((b). This knowledge can be expressed in the form of production rules. , simple system flow creation knowledge 3. Efficiency system 70 - creation knowledge 4 assumes that a simple system flow has been created.For example, if there are multiple input forms to file conversion and the processing cycles are different, For example, software design expertise such as the idea that dividing file conversion can improve computer processing efficiency is expressed and accumulated in the form of production rules.

Efficiency System Flow Creation Knowledge 4 is a knowledge source about 'what' to divide or integrate what under what conditions, and 'h' about how to divide or integrate when it is decided to divide or integrate.
Knowledge sources related to Ow 11 are stored separately.

Next, the present invention will be explained in detail. The basic control of inference, which is the core of the operation, is shown in FIG. First, the required system specifications that have been input to the user-defined initial processing function shown in the figure are written on the blackboard 5.37゜At the end of this user-defined initial processing function processing, the knowledge source to be activated first is specified.37
In this embodiment, the knowledge source is set so that inference proceeds from the output item side to the input item side, so the report creation job step setting knowledge source is activated. Next, the processing of the interpreter shown in FIG. 4 begins.

The 12 parts of the production rule in the report creation job step setting knowledge source specified in step 37 are matched with the events written on the blackboard, and the then part of the rule for which the matching is established is executed. 3
9. Determine the inference end condition (when the creation of the system flow is written on the specified mate on the blackboard),
The interpreter process is repeated until it is established. The selection 40 of the knowledge source to be interpreted is determined by using the control information in the production rules applied up to that point stacked in the inference control area 9.

Next, the operation of inferring the connection structure of the simple system flow will be explained. To create a simple system flow, for each functional component module (31 to 36 in Figure 3), the presence or absence of placement in the system 70-, input/output, and various properties of the functional component are inferred. Infer the bond structure of. Inference for each functional component module is performed sequentially based on the output information, inferring the functional component module that creates output information, then inferring the functional component module that creates input information for that functional component module, and so on. Continue reasoning until you arrive at the information. To create a simple system flow, we first infer the report generation functional component module. As a result of this inference, the following inference intermediate state is created.

Report creation ...... Node (object) placement - Yes ...... Attribute - value pair output □ Contract detailed table reference master Supplier mask code master Next, the above By inference, the connection structure of functional component modules (the context of job steps and the arrangement of intermediate files between job steps) that determines the system is inferred. After the above-mentioned inference for each functional component module, the inference of the connection structure is performed using the following production rule.

(Rule 7-2 (IP (Extraction Placement Yes) (Extraction Input No) (Update Placement No) (Verification Placement Yes)) (THEN(PROPO8E (Extraction Input Intermediate File 3) (Extraction Pre-Step Matching) (SELECT- INF knowledge 8)))) (Explanation) If there is an extraction job step arrangement, no input is set for the extraction job step, there is no update job step arrangement, and there is a matching job step arrangement, then the extraction The input of the job step is intermediate file 3, and the step before the extraction job step is the collation.

As information for selecting the next knowledge source to launch,
Register knowledge 8.

An example of a simple system throw created by inference is shown in FIG. Each symbol in the figure is the same as shown in FIG.

Since the processing is simple, the flow after input is monotonous in a single line.

Next, the operation of inferring the efficiency system 70- will be explained based on the blackboard that stores the inference results of the simple system flow. A knowledge base is created by dividing the knowledge possessed by software designers to improve efficiency into what should be divided or integrated under what conditions (what-type reasoning) and how to divide or integrate it (how-type reasoning). Store it in. First, by what-type reasoning, job steps that should be made more efficient are determined and written on the blackboard. An example of a production rule for Wllat-type inference is shown below.

(Rule 9-1 (IF' (7-isle conversion arrangement Yes) (File conversion split check No) ($5IYOFLISA Engineering KU
RU file conversion))(THEN(PROPO8E(
File conversion split check Yes) (Split work split file conversion) (SELECT-INF split knowledge)))) (Explanation) If the file conversion job step is placed incorrectly, there is no split check for the file conversion job step, If there are multiple inputs for the conversion job step and the processing cycles are different (user-defined function $S IYORI 5AIKtJRU judgment), enable the file conversion split check and set the split job step of the split operation to file conversion. .

As information for selecting the next knowledge source to launch,
Use divided knowledge.

Using this production rule, create the following state on the blackboard.

Dividing work: Node to be divided Job Step - File conversion: At the same time as writing on the attribute-value blackboard, information to be activated next is stacked in the inference control area. Inference ftj! The knowledge source to be interbred next from the stack information in area 9 using a first-in, first-out method is a divided knowledge source that is a 9-knowledge source for h□w type reasoning.

Create an efficient system flow by selecting from integrated knowledge sources and vertically divided knowledge sources. FIG. 5 shows how efficiency is improved by hqw-type reasoning (only a standard example is shown). FIG. 7 shows an example of an efficient system flow created as a result of whatm inference and JIOW type inference.

Each symbol shown in the figure is the same as that shown in FIG. 8, as in the case of FIG. In this flow, the knowledge of the division process shown in FIG. There is.

The system flow diagram obtained as a result of the inference is output on the graphic display 8 and presented to the user of the present invention, and if necessary, correction information is input through the input/output device 7 to modify and complete the system flow. .

According to this embodiment, a system flow diagram of a batch type business calculation system can be semi-automatically created. Further, by using the method of dividing and managing knowledge according to this embodiment, it is possible to easily collect and organize knowledge, and to improve the inference speed.

The above is a detailed description of the embodiment of the batch type calculation system.
However, in addition to this, there is a computer system used by U.

It goes without saying that the present invention can be similarly implemented in the field of creating flow diagrams by combining program components such as some OA systems.

〔Effect of the invention〕

According to the present invention, in conventional computer processing, etc., a software design expert considers the connection relationship of functional component modules that appear in the system flow, and also specifies the placement of each block in the system flow in detail for drawing. What previously had to be done is now semi-automated. That is,
The present invention stores and utilizes the knowledge of software design experts in a knowledge base to make inferences, but the parts of the knowledge base that are lacking in knowledge can be provided to the users of the system of the present invention. It is possible to supplement and semi-automatically draw the system 70-, which has the effect of saving the labor and brain power of software design experts.

Furthermore, since the software designer's knowledge for creating the system 70 is organized and stored in the computer, there is also the effect that the knowledge can be passed on smoothly.

[Brief explanation of drawings]

Figure 1 is a block diagram showing the overall configuration of the present invention, Figure 2 is a configuration diagram of knowledge in the knowledge base of the present invention, and Figure 3 is a functional component module in the simple system flow of the present invention. Fig. 4 shows the flow q, g5, iL, which shows the overall control procedure of the inference of the present invention. Division of functional component modules to create an efficient system flow,
FIG. 6 is an example of the output diagram of a simple system 70 according to the present invention; FIG.
FIG. 8 is an example of an output diagram of a flowchart of an efficient system according to the present invention. FIG. 8 shows the contents of the symbols in FIGS. 6 and 7. l... Reasoning processing device, 2... Knowledge base, 3...
Simple system 7, creation knowledge, 4...Efficiency system 702 creation knowledge, 5...Blackboard, 6...Frame (for storing system flow connection structure), 7...Required system specifications Terminal device for input/time response, 8...
System rotation Figure 3 Figure 4

Claims (1)

[Claims] In a method of creating a system 70- by combining software functional components, the steps include: accumulating system flow creation knowledge in a knowledge base; deducing a combination structure of software components that satisfies required system specifications; and based on the inference. A system flow creation method comprising the step of creating a system flow diagram.