JPS6324444A - Event scheduling system - Google Patents

Abstract

PURPOSE:To improve an operation efficiency by introducing a hierarchism into the constitution of an event to combine the existing events by an operator and form the new event. CONSTITUTION:The events are combined hierarchically to constitute the new event, the schedule of the slave event of a constituting element of the new event is set by a relative time viewing from a master event constituted of the slave events, the schedule of the slave event is converted into an absolute time by the start of the execution of the master event to decide whether the slave event is executed or not. Namely, the event has a list in which all the slave event names of the constituting element, a standby slave event list at a remarkable time and an executing slave event list. The schedule of the slave event is described in the relative time viewing from the event (relative time making the start time of the event a reference), when the event is started to be executed, the schedule of the respective slave events is converted into the absolute time and when the event is executed, whether the respective stand by events are executed or not is checked.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a scheduling method for arbitrary events on a computer, such as events in computer simulation and operations in computer animation.

(Prior Art) The event referred to here is any event on a computer, and in particular, it can be an event that occurs at a certain time in a computer simulation, or the movement of an object.

It is assumed that an event starts at a certain time (start time) and ends at a certain time (end time), and the time relationship between multiple events is set by the sequential relationship between their start time and end time. If an event starts at a certain time t1, the start time of the event may be set to tl. Also, event A is event B
If the start time of event A is started later than the start time of event B, the start time of event A may be set to a value later than the start time of event B. Also,
By setting a qualitative magnitude relationship between the start time or end time of each event, the start time of an event can be determined from the time relationship with other events without having to clearly set the start time of the event. It was also possible to do so.

For example, use the predicate hold to set the event event to start executing at time time, and set h
old(event, time) Furthermore, a predicate was provided to express the time relationship between events, and a schedule was set using this predicate. For example, the fact that event A occurs before event B is expressed using the predicate before as follows.

before (A, B) For details, see Mizoguchi et al., [Prototype of an action planning system including time information J (Information Processing Society of Japan 32nd (first half of 1985) National Conference Proceedings pp, 1559-1560).

(Problems to be solved by the invention) However, with the above method, it is cumbersome to describe and execute complex events, and the reusability of previously defined events is poor, making it inefficient to define new events. Moreover, there is a problem in that it is necessary to check whether all events are to be executed during scheduling.

The purpose of the present invention is to solve the above problems by introducing the idea of configuring other events hierarchically by combining events, and by combining events hierarchically, it is possible to generate and execute a new event. By doing so, it is an object of the present invention to provide an event scheduling method that can efficiently define and execute events.

(Means for solving the problem) The event scheduling method of the present invention configures a new event by hierarchically combining the above-mentioned events, and schedules child events that are constituent elements of the new event. It is characterized in that the time is set as a relative time viewed from a parent event made up of events, and the schedule of the child event is converted to an absolute time based on the start of execution of the parent event to determine whether or not to execute the child event. do.

(Operation) In the present invention, an event has a list of all child event names that are its constituent elements, a list of waiting child events, and a list of child events being executed at the main time.

The schedule of child events of the event is described in terms of relative time from the event (for example, relative time with respect to the start time of the event). Once the event begins execution, the schedule of each of its child events is converted to absolute time. When the event is executed, it checks whether each of its waiting child events is executed. Among the waiting child events, the child event determined to be executed is moved to the currently executing child event list and executed. The child events registered in the executing child event list are executed as they are. After execution, it is checked to see if it has finished, and if it has finished, it is removed from the running child event list. Furthermore, the child events registered in the waiting child event list are also checked for completion, and events whose end time has been reached without being started are removed from the waiting event list. By performing the above processing recursively,
A hierarchy of events is realized.

(Example) Hereinafter, one embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 2 is a conceptual diagram showing an example of the time relationship of occurrence of events. The start time and end time of each event are shown on the time axis. In this figure, event A starts at time L] and time t
Ends in 2. Event B starts at time t2 and ends at time t4. Event C starts at time t3 and ends at time t5. The event begins at time 4 and ends at time t6.

Here, each time tl, . . . , t6 may have an undetermined value. Therefore, for example, if time t2 is undetermined, then FIG. 2 expresses the qualitative property that event B starts at the same time as event A ends. Similarly, if t2 = t3, event B and event C occur at the same time t
2 (if t2 and t3 are undetermined, only the qualitative time relationship of [simultaneous] is expressed).

FIG. 3 is a conceptual diagram showing an example of the time relationship of hierarchized events. Event A is composed of its child events AI, A2 and A3. Event A occurs at time t1 and ends at time t2. The child event A1 occurs at a relative time tll with respect to the start time t1 of the event A, and ends at a relative time t13. The same applies to other child events. The time relationship of each event is the second
Interpreted in the same way as the figure.

FIG. 4 is an explanatory diagram for representing the contents of the event. The contents of the event are the name, the execution body representing the execution content of the event, the name of the parent event that includes various events as child events, the start time variable name, the end time variable name, the end flag, the name list of child events, and the child being executed. It consists of a list of event names and a list of names of waiting child events. The name of an event may be any unique name that can be distinguished from other events, but it is convenient to use a path name that includes the names of the parent event and all of the parent events above it. The execution body stores the contents to be executed by various events. In FIG. 4, as an example, execution contents for printing the character "hello" are stored. Although the name of the parent event is not necessarily required, it is convenient for implementing this method. The start time variable name is a pointer to a memory area for storing the start time, and the value of the start time variable is either given directly by the user or determined by the time synchronization mechanism described later. The end time variable name is a pointer to a storage area for storing the end time. The value of the end time variable is determined in the same way as the start time variable. These two numbers are collectively called a time parameter, and the event schedule is set by this time parameter. However, the value of the time parameter is not necessarily determined in advance, and may be dynamically determined based on the time relationship with other events. The details will be described later. Termination flags are used to indicate that events have fulfilled their purpose during execution. The end flag is 1. It is activated by the events themselves when the Kaneko event of the events ends and the processing of the execution body ends. When the end flag is activated, the events are ended, and the time at that time is stored in the end time variable. The child event name list is a list of all the names of child events that events include as constituent elements. The list of names of child events that are being executed is a list of all the names of child events that have already started at the time of interest and have not yet ended. The waiting child event name list is a list of all the names of child events that have not yet been executed at the time of interest.

Events are classified into standby events and execution events based on their status.

A standby event is one that has not yet occurred at the time T of interest and is in a state of waiting to start. An execution event is one that is already being executed at time T. Waiting events are registered in the waiting event list, and execution events are registered in the execution event list. In the initial state of scheduling (time O or any reference time), all events are first registered in the waiting event list.

In addition, by actually using equipment such as a computer, figures 2 and 3 can be
When events occur according to the time relationship shown in the figure, the time axis becomes discrete, and times are counted every certain unit time.

FIG. 5 is a block diagram showing an overall outline of an event scheduling system as an embodiment of the present invention. An event storage means 1 for storing and referencing the contents of an event, an event execution means 2 for determining whether an event of interest is executed or terminated at a certain time, and an event execution means 2 for determining whether an event of interest is executed or terminated at a certain time; A time synchronization means 3 for synchronization, an event execution means 4 for executing the event execution body, and a tally clock generation means 5 for providing time are provided. A connection line 6 for inputting the contents of an event to the event storage means 1, a connection line 7 for transmitting the contents of the event stored in the event storage means 1 to the event execution determination means 2, and a connection line 7 for inputting the contents of the event stored in the event storage means 1 to the event execution determination means 2. a connection line 8 for transmitting state transition information of execution or standby of a fact to the event storage means 1; an event execution determination means 2;
A connection line 9 is used to transmit time synchronization information consisting of the values of time parameters (start time variables and end time variables) and the time at that time to the time synchronization means 3, and the time parameter values determined by the time synchronization means 3 are transmitted to the event. A connection line 10 for transmitting to the storage means 1, a connection line 11 for transmitting information on whether or not to execute the event determined by the event execution determination means 2 and the time at that time to the event execution means 4, A connection line 12 for outputting the execution result and a connection line 13 for transmitting the time generated by the clock generation means 5 to the event execution determination means 2 are provided.

This is shown in Figure 4. The contents of the event are stored in the event storage means 1 shown in FIG. At the time of storage, a table for event reference is generated by combining the event base and the address of the storage area where the event contents are stored.

FIG. 1 is a flowchart showing an event scheduling system as an embodiment of the present invention, and shows processing in the event execution determining means 2, time synchronizing means 3, and event executing means 4 of FIG. A method of scheduling a certain event X at a certain time T will be described below as an example.

In step 101, it is determined from the value of the time parameter of event X and time T whether event X is executed at time T. If event X starts exactly at time T, synchronization with other events is performed based on the start time variable of event X and time T. Details will be described later. If it is determined in step 101 to execute, step 102 advances the process to steps 103, 104, and 105 for execution. If it is determined not to execute, step 106
Processing will proceed. If event X does not have any child events in step 103, the main body of event X is executed in step 104. If event X has a child event, step 105
Scheduling is performed for each of its child events. Scheduling of each child event is also performed in the same way as the process shown in FIG. In other words, the process is recursive. At this time, the schedule of each child event is converted into an absolute time by adding the value of the start time variable of event X to the start time variable and end time variable of each child event. What should be noted here is that the value of the time parameter (relative value seen from event X) of each child event is not necessarily fixed. Therefore,
The value of the time parameter of each child event is dynamically determined by propagation of the value in a time-related network, which will be described later. Finally, in step 106, it is determined whether event X ends at time T. If it ends, event
Synchronization with other events is performed based on the end time variable and time T. Details will be described later. Note that the order of scheduling for each child event in step 103 and the main execution order of step 104 may be reversed.

The above explanation is based on a certain event X at a certain time T.
However, if a certain time [T1.
T2], the process shown in FIG. 1 may be repeated for each event at each time T while advancing time T by one unit time from time T1 to time T2. However, there is no need to perform a start check for events registered in the execution event list.

FIG. 6 is a flowchart for explaining step 101 of the start check in FIG. 1 in detail. Consider time T. First, in step 21, it is checked whether the value of the start time variable ts of event X is already determined or undetermined. To indicate that it is undetermined, it is sufficient to provide an undetermined flag of ts. If it is undetermined, it is assumed that event X will not be executed at time T, and event X will remain registered in the waiting event list. If ts is determined, step 2
2, the value of ts and time T are compared in magnitude. ts
If the value of is greater than time T, event X remains registered in the waiting event list, assuming that it has not yet started at time T. If the value of ts is less than or equal to time T, it is interpreted as being executed. At that time, if it is determined in step 23 that the value of ts is equal to time T, it is assumed that the event X starts exactly at time T, and in step 24 ts
Time synchronization with other events is established based on the time T and the time T. A method for synchronizing will be described later. Events interpreted in step 22 to be executed are moved in step 25 from the waiting event list to the execution event list.

FIG. 7 is a flowchart for explaining step 106 of the completion check in FIG. 1 in detail. Consider time T. First, in step 31, it is checked whether the end time variable te of the event X is undetermined or determined (this is done in the same way as the start check). If te is determined, the value of te and time T are compared in step 32, and if te is less than or equal to T, time synchronization with other events is established based on te and time T in step 35. . If te is undetermined or te>T, it is checked in step 33 whether the end flag of event X is on. The end flag indicates that the Kaneko event of various events has ended, that is,
(The waiting core event list and running core event list are empty)
, and is activated by the various images themselves when the processing of the execution body is completed. If the end flag is activated, it is assumed that event X ends at time T, and in step 34 the value of end time variable te of event X is set to T, and in step 35 synchronization with other events is established. If the termination flag is not activated, it will not terminate. If it is determined at step 32 or step 33 that the event has ended, event X is removed from the running event list and the waiting event list at step 36.

Next, a method of setting time relationships between events will be explained.

An example of a method for setting time relationships between events is shown in FIG. 8(a). When setting the time relationship, specify the setting time axis indicating whether it is an absolute time or a relative time from a certain event. When defining the content of a parent event and setting the time relationship of its child events, it is convenient to set the set time axis to be a relative time viewed from the parent event. In Figure 8 (a>,
The time relationship of the events shown in FIG. 3 is taken as an example. The fact that event A1 in event A starts at relative time 3 is expressed using the predicate occur. The predicate seq means that event B starts immediately after event A, and the predicate 5tart means that event A1 and event A2 start at the same time.
1 and event A3 end at the same time is the predicate flnish
The predicate seq expresses that event A3 starts immediately after event A2.

Based on the time relationships set as above, a network of time parameters consisting of start time variables and end time variables is constructed, and synchronization between events is achieved by propagating the values of the time parameters on the network. . In particular, time synchronization of the start time variable of the parent event and the start time variable of the child event realizes conversion from relative time to absolute time in the schedule of the child event.

Next, a method of synchronizing events using a time-related network will be explained. A time-related network has a form in which time parameters are connected by directed edges, and a certain calculation formula is attached to the directed edges. The time parameter on the starting point side of the directed branch will be called an active variable, and the time parameter on the end point side of the directed branch will be called a dependent variable. FIG. 8(b) is an explanatory diagram showing an example of a method for configuring a network of time relationships of events. 8th
FIG. 8(b) shows a network configuration method corresponding to the setting method of FIG. 8(a). It is configured as a triple table of active variable name, dependent variable name, and calculation formula. When a value is stored in the active variable, the calculation result of the calculation formula is stored in the corresponding dependent variable. Further, if the dependent variable is also registered as an active variable, similar processing is performed using the dependent variable as the active variable. That is, the result of a value stored in a certain variable is propagated one after another on the network. Therefore, when a certain event starts or ends, the values of the time parameters of various events and the time parameters of other events connected on the network are determined one after another, and the times are synchronized.

FIG. 9 is a flowchart showing an example of a procedure for propagating values on a network of time-related events. Now, suppose that values are propagated on the network starting from a time parameter tp (which may be a start time variable or an end time variable).

First, in step 41, a propagation path on the network starting from tp is determined. To find the propagation path, based on the table in Figure 7, with tp as the active variable, all corresponding dependent variables and a list of pairs of calculation formulas ((tl, El
), (t2. E2), ... (tn, En)), and use this as a search route list. However, tl.

t2. ..., tn is a time parameter, El, E2.
..., En are calculation formulas corresponding to each time parameter. Next, tl.

t2. . . . For each tn as an active variable, all corresponding pairs of dependent variables and calculation formulas are obtained in the same way as above, and merged into the search route list that has already been obtained. However, time parameters already on the search route list are excluded. By repeating the above operations one after another, the propagation path is found (this is basically a horizontal search. The search algorithm is described in, for example, the literature self-published by Tsujii, Artificial Intelligence (Sonic Shoten, 1982), Vol. (described in Chapter 3). Based on the propagation list obtained in step 41, the values of the time parameters are propagated in step 42. Value propagation can be accomplished by sequentially extracting pairs of time parameters and calculation formulas from the propagation list, and storing the calculation results of the calculation formulas in the time parameters.

(Effects) As described above, the event scheduling method of the present invention introduces hierarchy into the configuration of events. therefore,
Workers can generate new events by combining existing events, increasing work efficiency. Moreover, since the schedule of the child event can be set at a relative time as seen from the parent event, the setting becomes easy. Furthermore, since scheduling is also performed hierarchically, each child event of an event that is not executed at a certain time does not need to be scheduled at that time, improving execution efficiency.

[Brief explanation of the drawing]

FIG. 1 is a flowchart showing an implementation method of the present invention, FIG. 2 is a conceptual diagram showing an example of the time relationship of events, FIG. 3 is a conceptual diagram showing an example of the time relationship of hierarchical events, and FIG. Figure 5 is an explanatory diagram showing the contents of an event, Figure 5 is a block diagram showing the entire event scheduling method, Figure 6 is a flowchart showing an example of a method for checking the start of a certain event at time T, Figure 7 The figure is a flowchart showing an example of a method for checking the completion of a certain event at time T; FIG.
b) is an explanatory diagram showing an example of a method for configuring a time-related network of events, and FIG. 9 is a flowchart showing an example of a procedure for propagating the value of a time parameter on the time-related network of events. In the figure, 1 is an event storage means, 2 is an event execution determination means, 3 is a time synchronization means, 4 is an event execution means, and 5 is a clock generation means. ／−′″＼ Representative Patent Attorney Susumu Uchihara″−゛・. Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 (a) Figure 9 13t

Claims (1)

[Claims] In the event scheduling method, the above events are hierarchically combined to form a new event, and the schedule of the child event that is a component of the new event is determined based on the relative time seen from the parent event composed of the above child events. an event scheduling method, characterized in that the schedule of the child event is converted into an absolute time based on the start of execution of the parent event, and it is determined whether or not to execute the child event.