JP2002318889A - Scheduling system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve such a problem in GA-based drafting that a schedule generated is not always improved as to evaluation values about the number of reservations and the order of precedence, so that it is difficult to generate a schedule which is prior to reduce the number of reservations and a schedule which is prior to allocate setting conditions which are higher in the order of precedence. SOLUTION: Candidates for the setting conditions of some opposite subjects to be drafted, for two schedules selected from among existing schedules, are interchanged to produce a new schedule. Within one schedule selected from among the existing schedules, one or more subjects whose drafting should be reserved are selected and the candidates for the setting conditions are changed to produce a new schedule. In the one schedule selected from among the existing schedules, one or more portions of the candidates for the setting conditions which are lower in the order of precedence are selected and the candidates for the setting conditions are changed into those which are higher in the order of precedence, to produce a new schedule. From all the new schedules, the schedule which is high in three evaluation values is obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a schedule for operating a satellite, and has a plurality of different objectives (items to be optimized). It relates to a scheduling system applied to create.

[0002]

2. Description of the Related Art As an example of a scheduling problem having a plurality of purposes, it is conceivable to formulate an operation schedule for maintenance of an artificial satellite. The planning of this type of operation schedule is based on the operation of the control station (hereinafter referred to as the operation station) used by each satellite when performing communication for maintenance and the setting items such as the operation day of the week, operation time and date. Specify in the form of setting conditions required for. further,
The setting items such as the operation station, the operation day of the week, and the operation time specified by the operation request are assigned priorities, and a plurality of setting condition candidates are specified. In the planning of the operation schedule, the operation request is “planned”. For each operation request, an operation schedule is created by selecting one of the setting condition candidates for each of the operation station, the operation day of the week, and the operation time and assigning it as the setting condition. FIG.
FIG. 4 is an explanatory diagram for describing a configuration example of an operation schedule. FIG. 11A shows three items for each setting item.
This shows a case where each setting condition candidate is specified. Each setting condition has a priority. If an operation request is created by simply assigning one setting condition candidate to each of the three setting items, an operation schedule having 27 operation requests can be created as shown in FIG.

As described above, in a scheduling problem in which a plurality of setting condition candidates to be assigned to a planning object are assigned priorities and designated, a setting condition candidate having the highest priority is assigned to each planning object. It is desirable to assign. However, when setting condition candidates having a high priority are assigned to all planning objects, setting condition candidates assigned to a plurality of planning objects may overlap.
For example, in the planning of the satellite operation schedule described above, the same station may be allocated to a plurality of operation plans in the same time period on the same day. Such duplication of assignment is called "contention". Since one operation station cannot communicate with a plurality of satellites at the same time, the operation schedule in which contention occurs becomes an unexecutable schedule. Therefore, when a conflict occurs, one of the conflicting operation requests is left based on a predetermined criterion (for example, the priority order assigned to each satellite), and the remaining operation requests are determined. Treats the assignment of setting conditions such as operation date and time and operation station as undecided. The operation request for which the allocation is undecided will be referred to as a “pending operation request”.

[0004] An example where the same scheduling problem occurs is a staffing schedule. The staff dispatch schedule is targeted at the dispatch destinations, and dispatch items and dispatch date / time are set items for each dispatch destination.
In this case, duplication occurs because each setting item is assigned from among the setting condition candidates that are prepared and designated in advance. Considering the skills of personnel as setting items,
The scheduling becomes more complicated.

[0005] A plurality of setting condition candidates to be assigned to a planning object such as a satellite operation schedule and a staffing schedule are assigned priorities and designated plurally, and when a conflict occurs between planning objects, one planning object is left. , For scheduling that puts the rest on hold,
The purpose of the planning is to "assign setting condition candidates having the highest priority as possible to each planning object, and at the same time reduce the number of planning objects to be put on hold as much as possible".

A planning problem such as an operation schedule or a staffing schedule is a combination optimization problem, and an optimum solution can be found by examining all combinations. However, when the scale (the total number of combinations) increases, it becomes difficult to check all combinations within a limited time. In such a case, a genetic algorithm (hereinafter referred to as G)
A) has been proposed to apply a meta-heuristic represented by A) to obtain a solution that satisfies the user within a limited time.

[0007] Consider a case in which the GA is applied to solve the operation schedule planning problem. One operation schedule is represented by a one-dimensional array, and this is defined as one individual in GA. For example, it is assumed that three setting items of operation day, operation time, and operation station and setting condition candidates for them are assigned to each operation request, and there are five operation requests (operation request 1 to operation request 5) in total. Consider the case where an operation schedule is planned. At this time, the GA individual is $ 1,
2,3,2,3,1,3,4,1,1,2,2,3
It is expressed as {1,1}. Each number represents the priority of the assigned setting condition candidate, and the first to third (that is, 1, 2, 3) are the priority of the setting condition candidate assigned to the operation request 1. Is an operation day, operation time, and operation station setting item order. Therefore, the operation request 1 is assigned the operation day of the first priority, the operation time of the second priority, and the operation station of the third priority. Similarly, the fourth to sixth (ie, 2, 3, 1) indicate the setting condition candidates assigned to the operation request 2 and the seventh to ninth (ie, 3, 4, 4)
1) sets the setting condition candidate assigned to the operation request 3 to 1
The 0th to 12th (that is, 1, 2, 2) indicate the setting condition candidates assigned to the operation request 4, and the 13th to 1st.
The fifth (that is, 3, 1, 1) represents a setting condition candidate assigned to the operation request 5.

The GA always keeps a plurality of individuals representing different schedules, and generates a new individual based on two individuals (called parent individuals) randomly selected from among the retained individuals. . There are two types of processing for generating a new individual. One is called "crossover" and the other is called "mutation". In the crossover, two new individuals (new individuals C and new individuals D) are generated by partially alternately replacing two parent individuals (parent individuals A and B).
In the case of planning an operation schedule having the above five operation requests, the parent individual A is set to $ 3,2,2,1,4.
3, 3, 2, 1, 2, 4, 2, 4, 3, 1｝, parent individual B
To {1,1,1,1,2,2,3,2,4,4,1,
When 1, 2, 1, 2}, the new individual C and the new individual D are, for example, {3, 2, 2, 1, 2, 2,
3,2,4,2,4,2,4,1,2} and {1,1,
1,1,4,3,3,2,1,4,1,1,2,3
It looks like 1｝. In this case, the new individual C includes the first to fourth (that is, 3, 2, 2, 1) and the tenth to thirteenth (that is, 2, 4, 2, 4) the parent individual A
From the fifth to ninth (that is, 2, 2, 3,
2,4) and 14th to 15th (that is, 1,
2) is the value of the parent individual B. The new individual D is the opposite. In this case, between the fourth and fifth values, the ninth and one
The intersection between the zeroth value and the thirteenth and fourteenth values is called an intersection, and the position of the intersection is determined randomly each time the intersection is performed. Mutation replaces some of the new individuals created by crossover with different values. The position to replace the value is chosen randomly. For example, 1 in the new individual C
A new individual is generated by changing the 0th value from the current “2” to another value.

[0009]

In the conventional scheduling system, as described above, the purpose of planning the operation schedule and the staffing schedule is to assign setting condition candidates with the highest priority as possible to each planning object.
Although it is to minimize the number of planning objects to be reserved, these two objectives are often in a trade-off relationship, and it is difficult to obtain a schedule that is 100% satisfactory at the same time. Therefore, both the evaluation value for the priority and the evaluation value for the number of suspensions not only create a schedule that is good on average, but depending on the purpose of planning, even if a low-priority setting condition candidate is Alternatively, a schedule to which setting condition candidates with high priorities are assigned even if the number of reservations is increased is required. However, when planning by applying GA, the place where the value assigned in the processing of crossover or mutation is changed is selected stochastically, so the newly generated schedule is not necessarily the number of reservations, the priority, However, there is a problem that it is difficult to generate a schedule that prioritizes the reduction of the number of reservations and a schedule that prioritizes the assignment of setting condition candidates with high priority.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and includes a schedule having a good overall evaluation value, a schedule having a good evaluation value regarding hold, and a schedule having a good evaluation value regarding priority. It is an object of the present invention to obtain a scheduling system in which different types of schedules are simultaneously drafted and a user can select a required one from three types of schedules.

[0011]

A scheduling system according to the present invention comprises a plurality of planning objects, each of which has a plurality of setting items, and a plurality of setting items which are designated in advance and prioritized for each setting item. In a scheduling system in which a plurality of schedules formed by allocating one of the setting condition candidates is specified, and one of the setting condition candidates is allocated to each plan target while repeatedly selecting the existing schedule, and a new schedule is generated. , A set condition candidate randomly selected from two stored schedules and assigned to any opposing planning object of the two schedules selected by the processing of crossover and mutation in the genetic algorithm. Generate or store a new schedule by replacing or changing From among the existing schedule, which is 1
One schedule was selected at random, and one or more planning objects that were suspended due to conflicting setting condition candidates in one selected schedule were randomly selected and assigned to the selected pending planning objects. A new schedule is generated by changing the setting condition candidate, one schedule is randomly selected from the stored schedules, and the setting condition candidate having a lower priority in the selected one schedule is assigned. Randomly select one or more, change the setting condition candidate at the selected location to a setting condition candidate with a higher priority, generate a new schedule, and suspend the planning object held for all generated new schedules Calculate the evaluation value for the number of items, the evaluation value for the priority order, and the total evaluation value And outputs the planning target minimization three priority schedule a higher priority set condition the allocation schedule giving priority to the candidate overall evaluation value is good schedule number of the schedule.

[0012] The scheduling system according to the present invention determines a schedule having the best overall evaluation value, a schedule having the highest evaluation value regarding hold, and a schedule having the highest evaluation value regarding priority for the next new schedule generation processing. It is something that is always left.

[0013] The scheduling system according to the present invention stores the determined number of schedules in the order of good overall evaluation value, good evaluation value for hold, and good evaluation value for priority in the next new schedule generation process. It is something to leave for.

The scheduling system according to the present invention comprises a plurality of planning objects, each of which has a plurality of setting items, and one of a plurality of setting condition candidates which are designated in advance and prioritized for each setting item. In a scheduling system in which a plurality of schedules formed by allocating a schedule are designated and one of the setting condition candidates is allocated to each plan target while repeatedly selecting the existing schedule, a new schedule is generated. Initial schedule generating means for generating a plurality of initial schedules, a new schedule storing means for storing the initial schedule and a new schedule generated thereafter, and setting between schedule planning objects stored in the new schedule storing means. Checks whether a conflict of condition candidates has occurred. , 1 the highest planning target of priority within the planning target there is a conflict if a conflict occurs
All schedules other than those scheduled are held as unscheduled holds, and there are a plurality of schedules including an evaluation calculation means for calculating an evaluation value regarding the hold, an evaluation value regarding the priority of each setting condition candidate, and an overall evaluation value, and a new schedule holding means And a schedule holding means for selecting and storing a predetermined number of schedules from among the schedules, and transferring the initial schedules stored in the new schedule holding means to the schedule holding means at the start stage of the planning processing, and terminating the planning processing. Then, a predetermined number of schedules are selected from the new schedule holding means and all the schedules stored in the schedule holding means in descending order of the total evaluation value, and the selected predetermined number of schedules and the schedule holding means are selected. To replace the same number of schedules stored in Joule and updating means selects from among the already schedule stored in the schedule memory unit two schedules randomly selected by crossover and mutation processes in the genetic algorithm 2
New schedule generating means A for generating a new schedule by replacing or changing the setting condition candidates assigned to arbitrary planning objects opposite to one schedule and storing the new schedule in the new schedule holding means, and stored in the schedule holding means. 1 out of the existing schedule
Randomly select one schedule, randomly select one or more pending plan objects in the selected one schedule, and change the setting condition candidates assigned to the selected pending plan objects to create a new schedule. A new schedule generating means B for generating a schedule and storing it in the new schedule holding means; and randomly selecting one schedule from the existing schedules stored in the schedule holding means, and having a lower priority in the selected one schedule. One or more locations to which setting condition candidates are assigned are randomly selected, the setting condition candidates at the selected locations are changed to setting condition candidates with higher priority, a new schedule is generated, and the new schedule is stored in the new schedule holding means. New schedule generating means C and storage schedule of schedule holding means End determination means for determining whether or not to continue the planning process based on whether or not the total evaluation value has been replaced with the schedule with the best overall evaluation value, and the total evaluation value among the planned schedules stored in the schedule holding means, A schedule display unit for selectively displaying a schedule having a best evaluation value regarding a good schedule and a hold, and a schedule having a highest evaluation value regarding a priority order.

[0015] In the scheduling system according to the present invention, the schedule updating means always includes a schedule having the best overall evaluation value, a schedule having the highest evaluation value regarding suspension, and a schedule having the highest evaluation value regarding priority. It is something to choose.

[0016] In the scheduling system according to the present invention, the schedule updating means selects a predetermined number of schedules in the order of good total evaluation value, in the order of good evaluation value for hold, and in the order of good evaluation value for priority. It is like that.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below. Embodiment 1 FIG. FIG. 1 is a block diagram showing a functional configuration of a scheduling system according to Embodiment 1 of the present invention. In the figure, reference numeral 51 denotes an initial schedule generating means for generating a plurality of initial schedules having different setting conditions at the start of planning. 52 is a new schedule holding means for storing a newly generated schedule. 53 is an evaluation value calculation means for calculating and obtaining an evaluation value of the new schedule stored in the new schedule holding means 52. 57 is a schedule holding means for storing a schedule for selecting and storing a predetermined number of schedules from a plurality of currently obtained schedules. Numeral 54 denotes a schedule generating means A, which randomly selects two schedules from currently obtained schedules stored in the schedule holding means 57 using meta-heuristics such as a genetic algorithm, and selects the selected two schedules. This is a means for changing a setting condition assigned to an arbitrary planning target based on a schedule and generating a new schedule. 55 is a schedule generating means B, which randomly selects one schedule from the currently obtained schedules stored in the schedule holding means 57, and further selects some of the planning objects suspended in the schedule. Is selected, and the setting conditions assigned to those planning targets are changed to generate a new schedule. Reference numeral 56 denotes a schedule generating means C, which randomly selects one schedule from currently obtained schedules, further selects planning objects to which setting conditions with low priorities are assigned in the schedule, and plans them. This is a means for generating a new schedule by allocating setting conditions having higher priorities to objects. Reference numeral 58 denotes a schedule updating unit that selects a predetermined number of schedules based on a predetermined criterion from the schedules stored in the new schedule holding unit 52 and the schedule holding unit 57, respectively. And the schedule holding means 57
Is a means for replacing the schedule stored in the selected schedule with the selected schedule. Reference numeral 59 denotes an end determining means for determining whether to end the repetitive planning process. Reference numeral 60 denotes a schedule display means for displaying a schedule meeting the designated condition from the schedules stored in the schedule holding means 57.

Next, the operation will be described with reference to an example of planning an operation schedule of the artificial satellite. The feature of this scheduling system is to create an optimal schedule by repeatedly generating and updating the schedule.
At this time, a new schedule is generated by changing a part of a plurality of different schedules obtained so far. FIG. 2 is a flowchart showing an operation flow of the scheduling system according to the first embodiment of the present invention. First, in step ST101, the initial schedule generation unit 51 generates P different operation schedules at the start of schedule planning, and stores the generated schedules in the new schedule holding unit 52. It is assumed that the value of P is determined in advance.

As a method of generating an operation schedule, there are a method of randomly assigning a setting condition candidate to each operation request, and a method of assigning a setting condition candidate having the highest priority. 4 shows an example of processing in the initial schedule generation means 51. The content of one operation request (hereinafter referred to as operation request X) {Setting item: Setting condition candidate} is {Operation day: Monday, Tuesday, Wednesday, Thursday, Friday / Operation time: 6:00, 7:00, 8:00, 9 Time / operating station: given as station A, station B, station C #. The numbers enclosed by ○ represent the priority. In the case of operating two more stations, for the operation request X, one of the five setting condition candidates for the operation day, one of the four setting condition candidates for the operation time, and three of the operation station It suffices to select two from the setting condition candidates. When the initial schedule is randomly generated, one is randomly selected from the setting condition candidates for the operation day, one is randomly selected from the setting condition candidates for the operation time, and the random is selected from the setting condition candidates for the operation station. Two are selected and assigned to operation request X.
On the other hand, when generating an operation schedule in which the setting condition candidates having the highest priority are assigned, the operation day is Monday,
When the operation time is 6:00, the operation station selects the station A and the station B as candidates and assigns them to the operation request X.

Next, the process proceeds to step ST102. The evaluation value calculation means 53 calculates the evaluation value of the operation schedule stored in the new schedule holding means 52. At this time, the evaluation value calculation means 53 checks whether or not a conflict has occurred between the operation requests. If a conflict has occurred, the evaluation value calculation unit 53 leaves one of the conflicting operation requests and leaves the remaining operation requests. All requests are unscheduled. The operation request for which the schedule has not been determined is called a pending operation request. In the process of leaving only one of the competing operation requests, the priority given to each operation request (for example, the priority between artificial satellites issuing each operation request) is used to give the highest priority. Leave high operational requirements.

FIG. 3 shows how operation requests compete. Station A is assigned to operation request 1, operation request 2, operation request 5, operation request 6, operation request 7, and operation request 9, and station B is assigned to operation request 3, operation request 4, and operation request 8. I have. Of these, three of operation request 1, operation request 6, and operation request 9, two of operation request 5 and operation request 7, and two of operation request 3 and operation request 8 are in conflict with each other. FIG. 4 shows a state in which the contention is resolved and the operation request 6, the operation request 7, the operation request 8, and the operation request 9 are suspended.

The evaluation values of the operation schedule calculated by the evaluation value calculation means 53 are three values: an evaluation value relating to suspension, an evaluation value relating to priority, and an overall evaluation value. The evaluation value regarding suspension is, for example, the number of suspension operation requests. The evaluation value regarding the priority is, for example, an average value of the priorities of the setting condition candidates assigned to each operation request. The comprehensive evaluation value is a summary of the evaluation value regarding the suspension and the evaluation value regarding the priority. There are several methods for obtaining the overall evaluation value. For example, an evaluation value relating to suspension is represented by S, its weight coefficient is represented by C, an evaluation value relating to priority is represented by T, and its weight coefficient is represented by D. Is determined by C * S + D * T. When the evaluation value calculation for all the schedules stored in the new schedule holding means 52 is completed, the evaluation value calculation means 53 stores the schedule including the hold operation request generated for the evaluation value calculation,
The original schedule held by the new schedule holding means 52 is replaced.

When the process of step ST102 is completed, the process proceeds to step ST103. In step ST103, the schedule updating unit 58 sets the new schedule holding unit 52
Is moved to the schedule holding means 57, and at the same time, the schedule held by the new schedule holding means 52 is emptied.

Next, the process proceeds to step ST104. Steps ST104 to ST107 are processing for generating a new schedule in the schedule generating means A54. First, in step ST104, the schedule generation unit A54 randomly selects two schedules from the operation schedules stored in the schedule holding unit 57, and proceeds to step ST105. In step ST105, a new schedule is generated by crossover and mutation processing in the genetic algorithm.

An example of generating a new schedule by crossover and mutation will be described with reference to FIGS. Here, in order to simplify the description, an operation schedule having five operation requests is used as an example. It is assumed that one setting condition candidate is assigned to each of the five operation requests for the operation day, operation time, and operation station. P1 and P2 in FIG. 5 are the two operation schedules selected in step ST104. In both cases, the operation schedules are shown in the order of operation requests 1 to 5. For example, the first three
The two numbers indicate the priority of the setting condition candidates regarding the operation day, operation time, and operation station to be assigned to the operation request 1. In the case of the operation schedule P1, the setting condition candidate of the priority 1 is assigned to the operation day of the operation request 1, the setting condition candidate of the priority 2 is assigned to the operation time, and the setting condition candidate of the priority 3 is assigned to the operation station. It is shown that.

In the crossover process, a new schedule is generated by selecting some crossover points at random and exchanging values sandwiched between the selected crossover points between the two operation schedules P1 and P2. In FIG. 5, CP1 to CP4 represent intersection points, and C1 and C2 are two new schedules generated by the intersection. The new schedule C1 inherits the value of the operation schedule P1 from the beginning to the intersection point CP1, from the intersection point CP2 to the intersection point CP3, and from the intersection point CP4 to the end, and from the intersection point CP1 to the intersection point. The operation schedule P between CP2 and the intersection CP3 to the intersection CP4.
Inherit the value of 2. The new schedule C2 is the reverse. In the example of FIG. 5, four intersections are selected, but the actual number of intersections is not limited to four.

Next, the mutation process will be described with reference to FIG. The mutation is performed on a new schedule generated by the crossover. In FIG. 6, C2 is a schedule for performing mutation, and is a new schedule generated by crossover. C3 is a new schedule generated by mutating the new schedule C2. MP1 is a point where mutation is performed (mutation point). First, the mutation point M for the new schedule C2
P1 is randomly determined. Next, the value indicating the priority of the setting condition candidate at the mutation point MP1 is randomly changed. In the example of FIG. 6, a new schedule C3 is generated by changing to a value “2” different from the current value “4”. In FIG. 6, the number of mutation points is one, but the actual number of mutation points is not limited to one.

When the crossover and mutation processing is completed, the process proceeds from step ST105 to step ST106, and the value of L representing the number of times of crossover and mutation has been increased by one. Next, the process proceeds to step ST107 to determine whether or not the value of L has reached a preset value. If the value has not reached the set value, the process returns to step ST104. If the set value has been reached, the schedule generation means A54 stores all the generated new schedules in the new schedule holding means 52, and proceeds to step ST108.

Steps ST108 to ST11
The processing up to 1 is processing for generating a new schedule in the schedule generating means B55. First, step ST10
In 8, the schedule generation unit B55 randomly selects one operation schedule from the operation schedules stored in the schedule holding unit 57, and proceeds to step ST109. In step ST109, the schedule generation means B55 randomly selects one or more operation requests held in the operation schedule selected in step ST108. Then, the setting condition candidate assigned to the selected pending operation request is changed, and a new schedule is generated.

An example of generating a new schedule by the schedule generating means B55 in step ST109 will be described with reference to FIG. In FIG. 7, P3 is a step S
This is the operation schedule selected by the schedule generation means B55 in T108. XP <b> 1 to XP <b> 3 are portions (changes) where setting condition candidates to be assigned in the suspended operation request are changed. C4 is a new schedule generated from the operation schedule P3 by the schedule generation means B55.

The operation schedule P3 represents a schedule having five operation requests of operation requests 1 to 5. It is assumed that one setting condition candidate regarding the operation day, operation time, and operation station is assigned to each operation request. For example, in the operation request 1, the setting condition candidate of the priority 2 is assigned to the operation day, the setting condition candidate of the priority 4 is assigned to the operation time, and the setting condition candidate of the priority 1 is assigned to the operation station. First, in the operation schedule P3, some of the suspended operation requests are selected. In FIG. 7, operation request 2, operation request 4, and operation request 5
Are pending, and operation request 2 and operation request 4 are selected from among them. Further, the selected operation request 2 and operation request 4
Represent the setting condition candidates assigned to
) Are randomly selected as changes.
By changing the value, a new schedule C4 is generated. In FIG. 7, one change point XP1 is selected from the operation request 2 and its value is changed from 2 to 3. From the operation request 4, two change points XP2 and XP3 are selected, and the respective values are changed from 4 to 1; 1
The schedule changed from to 2 is a new schedule C4.

When one new schedule is generated, the process proceeds to step ST110, where the number of generations of the new schedule M
Is incremented by one. Next, the process proceeds to step ST111, where M
Is determined to have reached a preset value. If the preset value has not been reached, the process returns to step ST108. If the set value has been reached, the schedule generation means B55 stores all the generated new schedules in the new schedule holding means 52, and proceeds to step ST112.

Steps ST112 to ST11
Steps up to 5 are processes for generating a new schedule in the schedule generating means C56. First, step ST11
In 2, the schedule generation unit C56 randomly selects one operation schedule from the operation schedules stored in the schedule holding unit 57, and proceeds to step ST113. In step ST113, the schedule generating means C56 randomly selects one or more locations to which setting condition candidates with low priorities are assigned in the operation schedule selected in step ST112, and further sets the setting condition candidates in the selected location. A new schedule is generated by changing to a higher priority.

An example of generating a new schedule by the schedule generating means C56 in step ST113 will be described with reference to FIG. In FIG. 8, P4 corresponds to step S
This is the operation schedule selected by the schedule generation means C56 in T112. YP1 and YP2
Is a change that changes the assigned setting condition candidate. C5 is a new schedule generated by changing the setting condition candidate of the change point YP1 and the change point YP2 in the schedule P4 to another setting condition candidate.

The operation schedule P4 represents a schedule having five operation requests of operation requests 1 to 5. It is assumed that one setting condition candidate regarding the operation day, operation time, and operation station is assigned to each operation request. For example, in the operation request 1, the setting condition candidate of the priority 2 is assigned to the operation day, the setting condition candidate of the priority 4 is assigned to the operation time, and the setting condition candidate of the priority 1 is assigned to the operation station.

The schedule generating means C56 randomly selects some of the locations where the setting condition candidates having a certain priority or lower are allocated. For example, when selecting a location to which a setting condition candidate with a priority of 4 or less is assigned,
In the operation schedule P4, setting condition candidates of priority 4, priority 5, and priority 4 are assigned to the second of the operation request 1, the second of the operation request 2, and the third of the operation request 3, respectively. Choose some randomly from. In FIG. 8, two of the changed points YP1 and YP2 are shown.
You have selected a location. Next, the values of the selected changed points YP1 and YP2 are changed to those having a higher priority than the current one, and the changed operation schedule is set as a new schedule C5.

When one new schedule is generated, the process proceeds to step ST114, where the number N of times of generation of the new schedule is set.
Is incremented by one. Next, the process proceeds to step ST115 to determine whether the value of N has reached a preset value. If the value of N has not reached the set value, step ST112
Return to If the set value has been reached, the schedule generating means C56 stores all the generated new schedules in the new schedule holding means 52, and proceeds to step ST116.

In step ST116, the evaluation value calculation means 53 calculates the evaluation values of all the new schedules stored in the new schedule holding means 52. This process is the same as step ST102, and a description thereof will be omitted. When the evaluation values of all the new schedules have been calculated, the process proceeds to step ST117.

Steps ST117 to ST119
Then, the schedule updating means 58 updates the operation schedule stored in the schedule holding means 57. First, in step ST117, P operation schedules are selected from all the operation schedules stored in the schedule storage unit 57 and the new schedule storage unit 52 in the order of good overall evaluation value. It is assumed that the value of P is set in advance. Next, step S
In T118, the selected P operation schedules are exchanged with the P operation schedules stored in the schedule holding unit 57, and at the same time, the schedule held by the new schedule holding unit 52 is deleted.

Next, the process proceeds to step ST119. In step ST119, the end determining means 59 determines the operation schedule having the best overall evaluation value among the schedules held in the schedule holding means 57 in step ST11.
It is checked whether or not the operation schedule has been replaced by the processing in step 8, and if not, the count value T indicating the number of times that the operation schedule having the best overall evaluation value has not been continuously updated.
Is incremented by 1, and the process proceeds to step ST120. If the operation schedule with the best overall evaluation value has been replaced by the process in step ST118, the count value T is cleared to zero and the process proceeds to step ST120.

In step ST120, the end determination means 5
In step 9, it is determined whether the count value T has reached a preset value t. If the count value T has not reached the preset value t, the process returns to step ST <b> 104. On the other hand, if the set value t has been reached, the process proceeds to step ST121.

In step ST121, the operation schedule stored in the schedule holding unit 57 by the schedule display unit 60 is the operation schedule with the best overall evaluation value, the one with the best evaluation value on hold, and the evaluation value on the priority order. Selectively display the operation schedule specified by the user, such as the best one.

As described above, according to the first embodiment, in the system for planning the operation schedule of the artificial satellite, in addition to the process of generating a new schedule by the genetic algorithm, A process for generating a new schedule by changing the assigned setting condition candidates, and generating a new schedule by assigning a higher priority setting condition candidate to an operation request assigned a lower priority setting condition candidate In addition to operating schedules with good overall evaluation values, such as operating schedules that prioritize minimizing the number of holds and operating schedules that prioritize the assignment of high-priority setting condition candidates, Different operation schedules at the same time, and users can The effect is obtained that the schedule can be selected.

Embodiment 2 In the first embodiment described above, a case has been described in which the schedule updating means 58 selects the operation schedule in the order of the overall evaluation value and updates the schedule stored in the schedule holding means 57. In the second embodiment, in the schedule selection process performed by the schedule updating unit 58, the operation schedule with the best overall evaluation value, the operation schedule with the best evaluation value on hold, and the operation schedule with the best evaluation value on the priority order Is always selected. The functional configuration of the scheduling system according to the second embodiment is the same as that of FIG. 1 described in the first embodiment, and a description thereof will be omitted.

Next, the operation will be described. FIG. 9 is a flowchart showing an operation flow of the scheduling system according to the second embodiment. Step ST201
The process from step ST216 to step ST216 is the same as the process from step ST101 to step ST116 described in FIG. 2 of the first embodiment, and thus the description is omitted here. When the calculation of the evaluation value of the new schedule in step ST216 is completed, the process proceeds to step ST217. The processing from step ST217 to step ST219 is performed by the schedule updating means 58.

In step ST217, the schedule updating means 58 selects the schedule having the best overall evaluation value from all the operation schedules stored in the new schedule holding means 52 and the schedule holding means 57. This is called the best schedule α. Next, of all the operation schedules stored in the new schedule holding means 52 and the schedule holding means 57, the operation schedule having the best evaluation value regarding the hold is selected from among the operation schedules excluding the best schedule α.
This is called the best schedule β. Next, of all the operation schedules stored in the new schedule holding means 52 and the schedule holding means 57, the operation schedule having the highest evaluation value regarding the priority order is selected from among those excluding the best schedule α and the best schedule β. This is called the best schedule γ. When the selection of the best schedules α, β, γ is completed, the process proceeds to step ST218.

In step ST218, the schedule updating means 58 removes the best schedules α, β, and γ selected in step ST217 from all the schedules stored in the new schedule holding means 52 and the schedule holding means 57. , P-3 operation schedules are selected in ascending order of the total evaluation value, and step ST2
Proceed to 19.

In step ST219, the schedule updating means 58 performs the operations in steps ST217 and ST2.
At the same time as replacing the P operation schedules selected at 18 with the P operation schedules stored in the schedule holding means 57, the new schedule holding means 5
The schedule held by 2 is deleted.

Next, the process proceeds to step ST220. In step ST220 and step ST221, the end determination means 59 determines whether to end the planning process. First, in step ST220, end determination means 59
Are the operation schedule with the best overall evaluation value, the operation schedule with the best evaluation value on hold, and the operation schedule with the best evaluation value on the priority among the operation schedules held in the schedule holding means 57. It is checked whether or not all of the three operation schedules have been exchanged by the process of ST219. If all of these three operation schedules have not been exchanged, the count value T representing the number of times the schedule has not been updated continuously is increased by 1, and the process proceeds to step ST221. If any one of the three operation schedules is replaced by the process in step ST219, the count value T is cleared to zero and the process proceeds to step ST221.

In step ST221, the end determination means 5
No. 9 checks whether or not the count value T has reached a preset value t. If the count value T has not reached the set value t, the process returns to step ST204, and if it has reached the set value t, the process proceeds to step ST222. .

In step ST222, from the operation schedules stored in the schedule holding means 57 by the schedule display means 60, the one with the best overall evaluation value, the one with the best evaluation value on hold, and the evaluation value on the priority order are displayed. Displays the operation schedule specified by the user, such as the best one.

As described above, according to the second embodiment, in the system for planning the operation schedule of the artificial satellite, in addition to the process of generating a new schedule by a genetic algorithm, A process of changing the assigned setting condition candidates to generate a new schedule, and generating a new schedule in which a higher priority setting condition candidate is assigned to an operation request to which a lower priority setting condition candidate is assigned. The processing is performed, and the operation schedule with the best overall evaluation value, the operation schedule with the best evaluation value on hold, and the operation schedule with the highest evaluation value on the priority order are always left, so that the overall evaluation value is good. Not only the operation schedule, but also the operation schedule giving priority to minimizing the number of holds, Operation schedules with different objectives, such as an operation schedule that prioritizes the assignment of setting condition candidates with higher priorities, can be drafted at the same time, and the best one can be generated among them. The effect is that the schedule can be selected accurately.

Embodiment 3 In the second embodiment,
The schedule updating means 58 calculates an operation schedule with the best overall evaluation value (the best schedule α) and an operation schedule with the best evaluation value related to the reservation excluding the best schedule α (the best schedule β).
And the case where the operation schedule (the best schedule γ) having the best evaluation value regarding the priority order except for the best schedule α and the best schedule β is selected. In the third embodiment, in the schedule selection process performed by the schedule updating means 58, the order in which the total evaluation value is good, the order in which the evaluation value related to suspension is good,
A plurality of operation schedules are selected in order of the evaluation value regarding the priority. The functional configuration of the scheduling system according to the third embodiment is the same as that described with reference to FIG. 1 of the first embodiment, and a description thereof will not be repeated.

Next, the operation will be described. FIG. 10 is a flowchart showing a flow of the operation of the scheduling system according to the third embodiment. The processing from step ST301 to step ST316 is the same as that shown in FIG.
Steps ST101 to ST described in
Since the process is the same as the process up to 116, the description is omitted here.

When the evaluation value of the new schedule is calculated in step ST316, the process proceeds to step ST317. The processing of steps ST317 and ST318 is performed by the schedule updating means 58.
In step ST317, the schedule updating means 58
First, from among all the operation schedules stored in the new schedule holding means 52 and the schedule holding means 57, P / 3 operation schedules are selected in descending order of the total evaluation value. These selected operation schedules will be referred to as a group of best schedules for comprehensive evaluation values. Next, of all the operation schedules stored in the new schedule holding means 52 and the schedule holding means 57, P / 3 schedules are sorted in ascending order of the evaluation value related to suspension from among the total evaluation value best schedule group. select. These selected schedules will be referred to as a group of reservation evaluation value best schedules. Next, out of all the operation schedules stored in the new schedule holding means 52 and the schedule holding means 57, the evaluation value related to the priority is selected from among the total evaluation value best schedule group and the pending evaluation value best schedule group. P / 3 operation schedules are selected in descending order. These selected operation schedules will be referred to as a priority evaluation value best schedule group. Upon completion of the above-mentioned schedule selection processing, the flow proceeds to step ST318.

In step ST 318, the schedule updating means 58 sets P operation schedules and schedule holdings, which are the total evaluation value best schedule group, pending evaluation value best schedule group, and priority evaluation value best schedule group selected in step ST 317. Means 57
Are replaced, the schedule held by the new schedule holding unit 52 is deleted, and the process proceeds to step ST319.

Step ST319 and step ST3
At 20, the end determination means 59 determines whether to end the planning process. The processing in step ST319 is the same as the processing in step ST220 in the second embodiment, and the processing in step ST320 is the same as the processing in step ST221 in the second embodiment. Therefore, the description of the processing in each step is omitted here. . Also, the processing in step ST321 is the same as step ST222 in the second embodiment, and a description thereof will be omitted.

As described above, according to the third embodiment, a setting condition candidate having the highest priority is assigned to each operation request, and the operation schedule of the artificial satellite for minimizing the number of operation requests to be held is minimized. In the planning system, in addition to the process of generating a new schedule by the genetic algorithm, the process of generating a new schedule by changing the setting condition candidates assigned to the pending operation requests, and the setting of a low priority A process is performed to generate a new schedule in which setting condition candidates with higher priorities are assigned to operation requests to which condition candidates have been assigned. By leaving a plurality of operation schedules in the order of the evaluation value related to the rank, the operation schedule with the better overall evaluation value Users can simultaneously create different operation schedules for different purposes, such as an operation schedule that prioritizes the minimization of the number of holds and an operation schedule that prioritizes the assignment of high-priority setting condition candidates. It is possible to select an operation schedule according to the purpose of planning, and to leave data with a good evaluation value. This has the effect of improving the work efficiency in response to the next scheduling.

[0059]

As described above, according to the present invention, two schedules are randomly selected from the stored schedules, and the two schedules selected by the crossover and mutation processes in the genetic algorithm. A new schedule is generated by replacing or changing the setting condition candidates assigned to arbitrary planning objects opposite to each other, and one schedule is randomly selected from the stored schedules, and the selected one is selected. Generates a new schedule by randomly selecting one or more planning objects that are suspended because of conflicting setting condition candidates in one schedule, and changing the setting condition candidates assigned to the selected pending planning objects Then, one schedule is randomly selected from the stored schedules, and the selected schedule is selected. One or more places where low priority setting condition candidates are allocated in one schedule are randomly selected, and the setting condition candidates of the selected places are changed to higher priority setting condition candidates to generate a new schedule. Calculate the evaluation value for the number of planning objects suspended for all new schedules generated, the evaluation value for priority, and the comprehensive evaluation value summarizing these two evaluation values. The system is configured to output three types of schedules: a schedule that prioritizes minimization and a schedule that prioritizes the assignment of setting condition candidates with high priorities, and a schedule that has a comprehensively good evaluation value. Users do not need to perform complex scheduling every time There is an advantage of being able to choose the combined schedule.

According to the present invention, the schedule with the best overall evaluation value, the schedule with the best evaluation value on hold, and the schedule with the best evaluation value on priority are always left for the next new schedule generation process. With this configuration, it is possible to simultaneously create operation schedules with different purposes and generate the best one among them. The user does not need to perform complicated scheduling one by one, and the operation schedule according to the planning purpose There is an effect that can be selected accurately.

According to the present invention, the order in which the total evaluation value is good is
The schedules with the determined number in the order of good evaluation value for suspension and good evaluation value for priority are configured to be left for the next new schedule generation process. Enables planning, allowing the user to select a schedule that meets the planning purpose without having to perform complicated scheduling one by one.Also, by leaving data with good evaluation values, it is possible to respond to the next scheduling and work efficiency There is an effect that can be achieved.

According to the present invention, an initial schedule generating means for generating a plurality of predetermined initial schedules at the start of planning, a new schedule holding means for storing the initial schedule and a new schedule generated thereafter, It checks whether or not a conflict of the setting condition candidates has occurred between the planning objects of the schedule stored in the schedule holding means, and when a conflict has occurred, has the highest priority among the competing planning objects. An evaluation calculating means for calculating an evaluation value relating to the reservation, an evaluation value relating to the priority of each setting condition candidate, and an overall evaluation value, and a new schedule holding means; Schedule for selecting and storing a predetermined number of schedules from a plurality of schedules At the start of the planning process, all the initial schedules stored in the new schedule holding device are transferred to the schedule holding device, and at the end of the planning process, all of the initial schedules stored in the new schedule holding device and the schedule holding device are transferred. Schedule updating means for selecting a predetermined number of schedules from the schedules in descending order of the total evaluation value, and replacing the selected predetermined number of schedules with the same number of schedules stored in the schedule holding means; 2 out of the existing schedules stored in the holding means
Two schedules are selected at random, and the set condition candidates assigned to any opposing planning objects of the two schedules selected by the crossover and mutation processing in the genetic algorithm are replaced or changed with each other to create a new schedule. One schedule is randomly selected from a new schedule generation unit A generated and stored in the new schedule storage unit and an existing schedule stored in the schedule storage unit, and is held in the selected one schedule. A new schedule generating means B for randomly selecting one or more planning objects, changing a setting condition candidate allocated to the selected suspended planning objects, generating a new schedule, and storing the new schedule in the new schedule holding means; Schedule already stored in the means Randomly select one schedule from among the selected schedules, randomly select one or more locations to which low priority setting condition candidates are allocated in the selected one schedule, and prioritize the setting condition candidates at the selected location A new schedule generating means C for generating a new schedule by changing to a setting condition candidate having a higher rank and storing it in the new schedule holding means, and whether or not the storage schedule of the schedule holding means has been replaced with the schedule having the best overall evaluation value. An end determination unit for determining whether or not to continue the planning process; and a schedule having the best overall evaluation value, a schedule having the highest evaluation value regarding suspension, and a priority order among the planned schedules stored in the schedule holding unit. Select the schedule with the best evaluation value Since it is configured to have schedule display means for displaying, it is possible to plan different schedules at the same time and generate the best schedule among them, and the user does not need to perform complicated scheduling one by one. This has the effect that a schedule according to the planning purpose can be selected accurately.

According to the present invention, the schedule updating means always selects the schedule having the best overall evaluation value, the schedule having the highest evaluation value relating to the hold, and the schedule having the highest evaluation value relating to the priority. With this configuration, it is possible to simultaneously plan the best operation schedules for different purposes, and the user can select an operation schedule that meets the planning purpose without having to perform complicated scheduling one by one. effective.

According to the present invention, the schedule updating means selects a predetermined number of schedules in the order of the total evaluation value, in the order of the evaluation value for the hold, and in the order of the evaluation value for the priority order. With this configuration, it is possible to simultaneously plan the optimal schedules with different objectives, and the user can select a schedule that meets the planning purpose without having to perform complicated scheduling each time. The effect of this is that the efficiency of the work can be improved in response to the next scheduling.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a functional configuration of a scheduling system according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing an operation flow of the scheduling system according to the first embodiment of the present invention.

FIG. 3 is an explanatory diagram showing a state of contention of operation requests according to the operation of the first embodiment.

FIG. 4 is an explanatory diagram showing a state in which an operation request according to the operation of the first embodiment is suspended.

FIG. 5 is an explanatory diagram showing an example of generating a new schedule by crossover according to the operation of the first embodiment;

FIG. 6 is an explanatory diagram showing an example of generating a new schedule by a mutation according to the operation of the first embodiment.

FIG. 7 is an explanatory diagram for describing an example of generating a new schedule according to the first embodiment;

FIG. 8 is an explanatory diagram for describing an example of generating a new schedule according to the first embodiment;

FIG. 9 is a flowchart showing an operation flow of the scheduling system according to the second embodiment of the present invention.

FIG. 10 is a flowchart showing a flow of an operation of the scheduling system according to the third embodiment of the present invention.

FIG. 11 is an explanatory diagram showing a configuration example of an operation schedule.

[Explanation of symbols]

51 initial schedule generating means, 52 new schedule holding means, 53 evaluation value calculating means, 54 schedule generating means A, 55 schedule generating means B, 56
Schedule generation means C, 57 schedule holding means, 58 schedule update means, 59 end determination means, 60 schedule display means.

Claims (6)

[Claims] 1. A plurality of planning objects, each of which has a plurality of setting items, and is formed by allocating one of a plurality of setting condition candidates designated in advance and assigned priorities to each setting item. In a scheduling system in which a plurality of schedules are designated and one of the setting condition candidates is assigned to each of the planning objects while repeatedly selecting the existing schedules, a new schedule is generated. Two schedules are selected at random, and the setting condition candidates assigned to arbitrary planning objects opposite to each other of the two schedules selected by the processing of crossover and mutation in the genetic algorithm are replaced or changed with each other to create a new schedule. A schedule is generated, and one schedule is selected from the stored schedules. A module is randomly selected, and one or more planning objects that are suspended due to conflicting setting condition candidates in the one selected schedule are randomly selected and allocated to the selected suspended planning objects. A new schedule is generated by changing the set condition candidates that have been set, and one schedule is randomly selected from the stored existing schedules, and a set condition candidate with a low priority is assigned to the selected one schedule. Randomly select one or more locations, change a setting condition candidate of the selected location to a higher priority setting condition candidate, generate a new schedule, and generate a new schedule for all of the generated new schedules. An evaluation value for the number of pending planning objects, an evaluation value for priority, and a comprehensive evaluation summarizing these evaluation values A value is calculated, and three types of schedules are output: a schedule that prioritizes minimization of the number of planning objects to be suspended and a schedule that prioritizes allocation of setting condition candidates with high priorities and a schedule that has an overall good evaluation value. A scheduling system, characterized in that: 2. A schedule having the best overall evaluation value,
2. The scheduling system according to claim 1, wherein the schedule having the best evaluation value regarding the suspension and the schedule having the highest evaluation value regarding the priority are always left for the next new schedule generation processing. 3. The method according to claim 1, wherein a predetermined number of schedules are left for the next new schedule generation process in the order of good overall evaluation value, good evaluation value on hold, and good evaluation value on priority. The scheduling system according to claim 1, wherein 4. A plurality of planning objects, each of which has a plurality of setting items, and is formed by allocating one of a plurality of setting condition candidates designated in advance and assigned priorities to each setting item. In a scheduling system in which a plurality of schedules are designated and one of the setting condition candidates is assigned to each of the planning objects while repeatedly selecting an existing schedule, a new schedule is generated. Initial schedule generating means for generating an initial schedule of: a new schedule holding means for storing the initial schedule and a new schedule generated thereafter; and setting condition candidates between schedule planning targets stored in the new schedule holding means. Determine if there is a conflict, and if so, In such a case, all the planning objects except the one with the highest priority among the competing planning objects are held pending unscheduled, and the evaluation value of the holding and the evaluation of the priority of each setting condition candidate are set. Evaluation calculation means for calculating a value and an overall evaluation value; schedule holding means for storing the predetermined number of schedules selected from the plurality of schedules of the new schedule holding means; and All of the initial schedules stored in the new schedule holding means are transferred to the schedule holding means, and at the end of the planning process, the new schedule holding means and all the schedules stored in the schedule holding means are selected in advance. Select the determined number of schedules in the order of best overall evaluation value, and select the Schedule updating means for replacing a predetermined number of schedules with the same number of schedules stored in the schedule holding means; and randomly selecting two schedules from the existing schedules stored in the schedule holding means. ,
Generating a new schedule by mutually replacing or changing setting condition candidates assigned to arbitrary planning objects opposite to each other in the two schedules selected by the processing of crossover and mutation in the genetic algorithm; A new schedule generating means A to be stored in the schedule storing means, and randomly select one schedule from the already-stored schedules stored in the schedule holding means, and select a planning object held in the selected one schedule. A new schedule generating means B for randomly selecting one or more and changing a setting condition candidate allocated to the selected pending planning object to generate a new schedule and storing it in the new schedule holding means; Said schedule stored in the means One schedule is selected at random from among the schedules, and at least one location to which a setting condition candidate with a low priority is assigned in the selected one schedule is randomly selected, and the setting condition candidate of the selected location is selected. Is changed to a setting condition candidate having a higher priority, a new schedule is generated by generating a new schedule and stored in the new schedule holding means, and a storage schedule of the schedule holding means is replaced by a schedule having the best overall evaluation value. End determination means for determining whether or not to continue the planning process based on whether or not the schedule has the best overall evaluation value among the planned schedules stored in the schedule holding means, The best evaluation value for good schedule and priority Scheduling system characterized by comprising a schedule display means for selectively displaying a schedule. 5. The schedule updating means always selects a schedule with the best overall evaluation value, a schedule with the best evaluation value on hold, and a schedule with the best evaluation value on priority. The scheduling system according to claim 4, wherein 6. The schedule updating means selects a predetermined number of schedules in the order of the total evaluation value, the order of the hold evaluation value, and the order of the priority evaluation value. Item 6. The scheduling system according to Item 4.