JPH0337201B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a scheduling device.

Scheduling means predetermining a rational procedure for a production process (including processing steps, processing steps, etc.) and executing it according to the schedule sequence. FIG. 1 is a block diagram showing an example of the scheduling target, in which 1 is a parts supply port, 2 is a product outlet, 3a, 3b, 3c, 3
d are processing devices 1, 2, 3, and 4, respectively,
Processing equipment 1 (3a) and 2 (3b) are pre-processing equipment with mutually equivalent functions, and processing equipment 3 (3b) is
c) and 4 (3d) are post-processing devices having mutually equivalent functions. Further, 4 is a processing skip path when no post-processing is required. The parts supplied to the parts supply port 1 are pre-processed by device 1 (3a) or 2 (3b), and then post-processed by device 3 (3c) or device 4 (3d), or post-processed by path 4. is omitted and reaches the product outlet 2. In this case, from 1 to 3
A schedule is required for distribution to c, 3d, and 4.

Figure 2 is a flowchart showing the conventional schedule algorithm.
6, 7, and 8 are program steps, respectively. First, it is determined in step 5 whether or not the material has been pre-processed, and if NO, the device 1 is determined in step 6.
(3a) or 2 (3b) is selected. However, in the example shown in Figure 2, both devices 1 (3a) and 2 (3b)
It is assumed that there will be no BUSY (work in progress and prohibited status). If YES in step 5, it is determined in step 7 whether or not post-processing is necessary. If YES, the part is supplied to the jump path 4, and if NO, the part is sent to device 3 (3c) or 4 (3d) based on the determination in step 8. is selected. However, in the example shown in FIG.
Assume that both c) and 4 (3d) are never BUSY.

In conventional devices, scheduling is executed as described above, so it is necessary to create a scheduling algorithm that suits the scheduling target in advance, and if the target changes, the algorithm must be changed, so the scheduling algorithm must be changed in advance. It was not possible to respond immediately to changes in the subject. Also,
Conventionally, scheduling algorithms were created in a non-real-time manner by manually converting scheduling rules into scheduling algorithms statically, that is, by knowing information about the scheduling target in advance before scheduling. In order to do this, it is necessary to know in advance information such as the statistical frequency of occurrence of the scheduling target, and there is also the disadvantage that manual conversion takes time and requires a skilled programmer.

This invention was made in order to eliminate the drawbacks of the conventional ones as described above, and it is a system that stores information on the functions of the scheduling target, information on the status of the scheduling target, and information on the scheduling rules. It is an object of the present invention to provide a scheduling device that allows a data processing device to perform scheduling speculatively based on information.

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 3 is a block diagram showing an embodiment of the present invention, in which numeral 9 denotes an inference mechanism that sequentially applies knowledge information to proceed with calculations, and is composed of a program memory and a processing device that reads and executes the program. ,1
0 is functional information of the scheduling target, 11 is status information of the scheduling target, and 12 is information regarding scheduling rules. 10, 11, and 12 are collectively referred to as knowledge information, and the knowledge information is stored in the data memory. 13 is input and is input to update the status information 11. 14 is an output, and the schedule result is output.

FIG. 4 is a flowchart showing the program steps executed in the inference mechanism of FIG.
6, 17, and 18 are program steps, respectively. Step 15 is a judgment unit that determines whether there is a rule to be applied. Step 16 is a rule application unit that calculates the rule by the inference mechanism 9 if there is a rule to be applied in step 15 (YES). Step 17 is a calculation unit that determines whether there is a rule to be applied. The output judgment unit, step 18, which judges whether there is an output at the time of completion, is the step 1.
7, if there is an output result (scheduled result) (YES), it is an output processing unit that outputs it. Incidentally, if there is no rule to be applied in step 15 or if there is no output result in step 17 (NO), the presence or absence of a rule to be applied is determined from the beginning again in step 15.

That is, in the inference mechanism program shown in FIG. 4, the schedule result is calculated by dynamically pattern matching the knowledge information 10, 11, and 12 in real time.

FIG. 5 shows an example of the functional information 10 in FIG. 3,
FIG. 6 shows an example of the status information 11 in FIG. 3, and FIGS. 7 and 8 show examples of the information 12 regarding scheduling rules. The information 12 regarding the scheduling rules is the description of the PREMISE section (condition judgment section) as shown in FIGS. 7 and 8.
It consists of a description of the ACTION part (processing part).

When the scheduling target is shown in FIG. 1, the scheduling target function information 10 should describe the functions of devices 1 to 4 and the functions necessary for processing parts 1, 2, etc. as shown in FIG. Good (hereinafter,
For convenience of explanation, it is assumed that there are only parts 1 and 2. ) Also, schedule target status information 11
are the states of parts 1 and 2 and device 1 as shown in Figure 6.
It is sufficient to describe the states of ~4. The status information is input 13 by providing detection means (not shown) for it.
or by the output processing unit 18 in FIG. In addition, scheduling rule 12 is shown in FIG. 7 (RULE: 1), and the rule for supplying parts to device 1 or 2 is shown in FIG. 7 (RULE: 2).
Figure 8 (RULE: 3) is the rule for determining parts to be supplied to device 3 or 4, and Figure 8 (RULE: 4) is the rule for determining parts to be supplied to device 3 or 4.
The rules for supplying parts to the jumping route 4 are shown respectively. In these rules, X is a component and Y is a device.

The inference is performed using the algorithm shown in FIG. In step 15, it is determined whether there are any rules to be applied. (RULE: 1) means that when checking the storage of scheduled status information 11 (Fig. 6), part 2 conforms to X, and from the storage of scheduled function information 10 (Fig. 5), devices 1 and 2 are Also, from the storage of the state information 11 (FIG. 6), the device 1
Since it is NO BUSY, it is determined that (RULE: 1) can be applied, and this (RULE: 1) is applied in step 16, and the processing result of "supplying part 2 to device 1" is generated, and step 1
At step 7, it is determined that there is an output, and the output at step 18 is activated, and the processing result ``supply part 2 to the device'' is output. That is, generally speaking, this inference process derives ACTION B if PREMISE A is given in (if A then B), and is an expanded and mechanized version of the operational function based on modus ponens in logic.

Next, (RULE:2) is determined to be applicable because Part 1 is pre-processed from Fig. 6 and Part 1 requires post-processing from Fig. 5. A processing result "This is a machined part" is generated, output, and passed to (RULE:3). It is also possible to start (RULE:2) from (RULE:3). The "post-processed part" in the information "X is a post-processed part" in the PREMISE part 1 of (RULE: 3) is neither mere functional information nor state information, so it is It doesn't exist either. (RULE:2) ACTION part is “X
is a post-processed part, so (RULE: 3) is (RULE:
2) is activated, and if the PREMISE part of (RULE:2) is established, the information of the ACTION part can be obtained.
In other words, in RULE: 3, X is part 1 and Y is part 5.
Fig. 6 (In Fig. 6, the description of the status information of devices 3 and 4 is omitted from the drawing, but “Device 3 is NO
BUSY), it is device 3, and a processing result of ``supplying component 1 to device 3'' is generated. In the example shown in FIGS. 5 and 6, (RULE: 4) does not satisfy the PREMISE portion and the determination at step 15 is NO. Scheduling in the present invention is performed as described above.

In addition, in the above embodiment, the scheduling rule 12 was explained as referring to the scheduling target function information 10 and the scheduling target status information 11, but the function information 10 and the status information 11 are compatible.
Scheduling rule 1 with PREMISE part
2 can also be called.

As described above, according to the present invention, in order to configure a scheduling system, functional information of a scheduling target, status information of a scheduling target, and scheduling rules are stored as knowledge information, and scheduling is determined using an inference mechanism. Therefore, it is possible to respond to changes in the scheduling target by simply changing the knowledge information, eliminating the need to change the program as in the past.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing an example of a scheduling target, Fig. 2 is a flow chart showing a conventional scheduling algorithm, Fig. 3 is a block diagram showing an embodiment of the present invention, and Fig. 4 is the inference mechanism of Fig. 3. FIG. 5 is a flowchart showing the program executed in FIG. 3, FIG.
FIG. 7 and FIG. 8 are diagrams showing examples of scheduling rule information. 9... Reasoning mechanism, 10... Scheduling target function information, 11... Scheduling target state information, 12... Scheduling rule, 13... Input, 14... Output, 16... Rule application, 18... Output processing. Note that the same reference numerals in the figures indicate the same or equivalent parts.

Claims (1)

[Claims] 1: means for storing information on the function to be scheduled; means for storing the state of the scheduled object; means for storing information regarding the scheduling rule in the relationship between the description of the condition determination unit and the description of the processing unit; By dynamically pattern matching the description of the condition judgment part in the above scheduling rule and the storage of information and status information of the function to be scheduled in real time, the description of the processing part of the above scheduling rule is and an inference mechanism that calculates the relationship between the two, and the scheduling device is configured to control the scheduling target based on the output of the inference mechanism.