JPH04171538A - Repetitive data processing scheduling system - Google Patents

Abstract

PURPOSE:To reduce the part of the processing that is abandoned halfway due to the shortage of the allocated processing time by controlling the start and the end of a task based on the estimated data processing time as well as a task executing order plan and monitoring the time passed for execution of the task. CONSTITUTION:A data processing time estimating part 11 reads the input data to be processed in a repetitive period and estimates the time for the data processing based on the quantity, the type, the value, etc., of the input data. A task control data storage part 12 stores the optimum task executing order set for each estimated data processing time as the table form data. A task execution control/monitor part 13 selects an optimum task executing order based on the data stored in the parts 11 and 12 and controls the start and the end of each data processing. Then the part 13 monitors the time passed for execution of a due task. In such a way, one of plural task executing order plans set previously is selected based on the processing priority of each task. Based on this selected plan, the optimum control is assured to the execution of the task within a fixed repetitive period.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a repetitive data processing scheduling method, and in particular, processing time prediction for repetitive data processing at fixed intervals, such as radar data processing and device control based on the processing of radar data. The present invention relates to an iterative data processing scheduling method that optimally schedules the execution order of each data process based on information and priority information of a plurality of data processes to be performed within a certain period.

[Conventional technology]

Conventionally, when controlling equipment by processing track data such as radar data, the input data to be processed is input at regular intervals, and the equipment to be controlled also requires control data at regular intervals. In such data processing, when input data from non-periodic human systems is added and must be processed within the above fixed period, the time span within the fixed period may be A method has been adopted in which data is divided into fixed time periods and the necessary data processing is performed sequentially.

In such a method, each data processing is required to be completed within the time span allotted to it, and if the processing time varies greatly depending on the value and amount of input data, it is necessary to complete each data processing within the time span allotted to it. It was necessary to give them ample time.

[Problem to be solved by the invention]

However, in the conventional example described above, when complex data processing is required, it is difficult to design a time allocation that fits within a given repetition cycle (for example, radar data input interval). Ta. In other words, if the sum of the maximum estimated processing times required for each data processing exceeds the repetition period, lengthen the repetition period (for example, lengthen the radar data input period to increase the response time of the entire system). ), or accept that low-priority data processing may be abandoned due to interruption.

In either case, there is a problem that when the amount of data to be processed is small, the time allocated to each data process cannot be used effectively.

[Means to solve the problem]

The present invention aims to solve the above problems and automatically estimates the execution time required for each data process (hereinafter, a program that executes this data process is referred to as a task) based on input data to be processed. Based on this value, one of the predetermined multiple task execution order floating points is selected based on the processing priority of each task, and based on this, the purpose is to optimally control task execution within a certain repetition period. There is.

Therefore, the task execution control method of the present invention is capable of processing time-sharing periodic data consisting of a plurality of data processing tasks whose processing is required to be completed by a certain point within a fixed repetition period and which are data-coupled with each other. In the processing system, a data processing time prediction unit estimates the execution time of each task from the processing target, input data, and previous processing time results for each repetition period, and for each range of the size of the predicted data processing time. A task control data storage section that stores a plurality of predetermined task execution order floating points, and an optimal task execution order is selected based on the data stored in the data processing time prediction section and the task control data storage section. It is equipped with a task execution control/monitoring unit that controls the start and end of processing (tasks) and monitors the elapsed execution time of tasks necessary for this, and predicts data processing time from the contents of input data to be processed. It is configured by performing optimal task execution scheduling based on the following.

〔Example〕

The present embodiment will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, and FIG. 4 is a flowchart of processing of the repetitive data processing scheduling mechanism of this embodiment according to FIG.

The data processing time prediction unit 11 reads input data to be processed within the repetition period, and calculates an estimated value of the time required for processing the data using a calculation formula given in advance based on the amount, type, value, etc. do. Furthermore, if it is more advantageous to correct the estimated value calculated using a specific formula using the previous actual value, the previous actual value is read from the task control data storage unit 12 and the correction is performed.

The task control data storage unit 12 stores the optimal task execution order determined for each estimated data processing time obtained by the function of the data processing time prediction unit 11 as tabular data (-
(An example is shown in FIG. 3), this data is determined in advance as a system design value, and is read into and held in the task control data storage section 12 at the time of initializing this system.

Further, the task control data storage unit 12 stores numerical values used for correcting the above-mentioned data processing time calculation value.

This numerical value is calculated based on the difference between the previous estimated value and actual value of similar (same category) data processing transmitted via the task execution control/monitoring unit 13, and is stored in the form of no numerical value or deviation value. Ru. This value is used for similar data processing (e.g.
It is updated every time a manual action (processing for the same type of manual action) is performed.

The task execution control/monitoring unit 13 has a timer function that monitors the time at regular intervals (for example, 10 milliseconds), and based on the data processing completion status data notification from each task, the data processing time prediction unit 11 and the inter-task It controls the start and end of each data processing task (18A to 18D in FIG. 1) connected to the data storage section 18 for data exchange.

The execution order of each data processing task is determined by the task execution order data read from the task control data storage section based on the data processing time estimate notified from the data processing time prediction section 11 as processing end state data. Based on the determined task execution order, the next task is started while confirming the successful completion of the previous task, but if the actual processing time exceeds the estimated time and the scheduled task is not completed by the specified elapsed time. In this case, the processing of the task is forcibly terminated, and the output data transmitter (indicated by 16 and 17 in FIG. 1) is notified that the output data of the task is invalid. It also notifies the task control data storage unit 12 of the actual data processing time value used to correct the estimated data processing time value.

Next, the operation of this embodiment will be described with reference to FIG. 4, focusing on the repetitive data processing scheduling mechanism.

The time requirements required for the system of this embodiment, the execution time of each processing task, and the data processing priority are shown in FIG. FIG. 3 shows an example of task execution order data. At this time, task sequence numbers 1 to 6
shows the task execution order corresponding to .

First, the start of periodic processing is performed by the task execution control/monitoring unit 13.
This is recognized by the timer function, and the data processing time prediction unit 11 is activated (step 41). The data processing time prediction unit reads the target data to be processed in the current processing cycle from the input data storage units 14 and 15, and calculates the estimated required data processing time based on a preset calculation formula (
Step 42).

For specific types of input data, the estimated value is corrected using the previous processing time actual value, and is passed to the task execution control/monitoring unit 13 along with other processing time estimates that do not require correction (
Step 43) The task execution control/monitoring unit 13 refers to the task execution order data (an example is shown in FIG. 3) in the task control data storage unit 12 and selects a task execution order corresponding to the estimated processing time value. do. If the estimated data processing time is the value shown in Figure 5, the task sequence number in Figure 3 is "
3" execution order r A-) B-* C-+ p"
is selected (step 44), see the scheduling example of FIG.

Next, the task execution control/monitoring unit 13 sequentially starts the tasks according to the above execution order, and determines their end status (normal, normal, etc.).
abnormality) and end time. If the actual processing time exceeds the estimated time and the processing of a predetermined task is not completed by the specified request time (for example, in the example in Figure 5, task C is not completed by the time when aircraft control data generation completion is requested. (in this case), the processing of the task is forcibly terminated and the output data thereof is invalidated (step 45).

When the execution of a series of tasks is completed, the task execution control/monitoring unit 13 stores the actual data processing time value required for the estimated time correction in the task control data storage unit 12. The task execution control/monitoring unit 13 also determines whether to continue the periodic process, and normally starts the next periodic process (step 4).
7).

〔Effect of the invention〕

As described above, according to the present invention, in a process executed by a plurality of task sequences activated at regular intervals, it is possible to reduce the portion of the process that is abandoned midway due to insufficient allocated processing time. Furthermore, in predicting data processing time, the accuracy of prediction can be improved in a learning manner by referring to and correcting the history of the previous processing wait time.

Furthermore, by providing the task startup section with information on multiple task sequences selected according to the size of the predicted data processing time in a table format, it is possible to achieve flexibility and expandability even in fixed-time startup type scheduling systems. This has the effect of realizing data processing.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, FIG. 2 is an explanatory diagram showing an example of the completion time and processing time of each data processing (task) required by the system of this embodiment,
FIG. 3 is an explanatory diagram showing an example of a task execution order plan stored in the task control data storage unit in this embodiment;
The figure is a flowchart of the processing of the iterative data processing scheduling machine of this embodiment according to FIG. 1, and FIG. FIG. 2 is an explanatory diagram showing an example of the execution order of tasks. DESCRIPTION OF SYMBOLS 11... Data processing time opening prediction unit, 12... Task control data storage unit, 13... Task execution control monitoring unit, 1
4.15... Input data storage unit, 16... Aircraft control data transmitting unit, 17... Display display data transmitting unit.

Claims (1)

[Scope of Claims] A time-sharing periodic data processing system comprising a plurality of data processing tasks that are required to complete processing by a certain point in a certain repetition period and that are data-linked to each other, comprising: a data processing time prediction unit that estimates the execution time of each task for each repetition period based on the processing target, input data, and previous processing time results; A task control data storage unit that stores a proposed task execution order, and an optimal task and execution order are selected based on the data stored in the data processing time prediction unit and the task control data storage unit, and each data processing (task) is It is equipped with a task execution control/monitoring unit that controls the start and end of the task and monitors the elapsed execution time of the tasks necessary for this, predicts the data processing time from the contents of the input data to be processed, and calculates the optimal time based on this. A repetitive data processing scheduling method characterized by performing task execution scheduling.