JPH0573320A - Schedule preparation supporting device - Google Patents

Abstract

(57) [Summary] [Purpose] To provide a schedule creation support device that facilitates construction and maintenance of a knowledge base for schedule creation. [Structure] A schedule creation support device divides a problem of schedule creation into a plurality of parts, expresses each sub-problem obtained by the division as a node, and a node corresponding to a sub-problem and a higher order of the sub-problems. Knowledge base (FIG. 1; FIG. 1; FIG. 1; FIG. 1; which stores a tree structure obtained by connecting nodes corresponding to the sub-problems of FIG. 1) and a procedure group for solving sub-problems corresponding to each node of the tree structure. 12) and searching the tree structure of the knowledge base,
A schedule creating unit (FIG. 1; 14) for creating a schedule by sequentially executing the procedure corresponding to the node obtained by the search is provided.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a schedule creation support device for creating a schedule by utilizing the knowledge of a specialist.

[0002]

2. Description of the Related Art In a conventional schedule creation support device (for example, Japanese Patent Laid-Open No. 3-25621), the knowledge necessary for creating a schedule is the if rule type production rule generally used for diagnostic problems. Described in. Moreover, these rules are stored in a single knowledge base. In general, if-then-style rules basically indicate that under such conditions described after if, such a state change described after then is generated. In order to create a schedule, it is not enough to describe it. In addition, it is necessary to describe in what order the schedule will be created later. That is, in addition to the original role of changing the state, it is necessary to include a parameter that changes the flow of control of what to do next, and the readability of the rule will fall, and the knowledge will be described. There was a problem that was complicated.

As a conventional schedule creation support device that partially solves the problem, a plan-making means, an evaluation means, an analysis means, and an improvement plan making means are provided with a unique knowledge base for each schedule making phase. There is one (Japanese Patent Laid-Open No. 2-185351). Each knowledge base stores the rules used in that phase.

[0004]

Since it is important for the expert to take a problem in creating a schedule, it is necessary to proceed with the process according to this procedure. However, in any of the conventional techniques, the schedule creation procedure and the schedule allocation procedure are not separated, and the rule must be written so that the rule is fired according to the schedule creation procedure, which is a burden on the knowledge base builder. .. The present invention aims to solve such problems. That is, an object of the present invention is to provide a schedule creation support device that facilitates the construction and maintenance of a knowledge base for schedule creation.

[0005]

A schedule creation support apparatus of the present invention divides a problem of schedule creation into a plurality of parts, expresses each sub-problem obtained by the division as a node, and divides it into a certain sub-problem. Information on the schedule creation procedure consisting of a tree structure obtained by connecting the corresponding nodes and the nodes corresponding to the upper sub-problems of the sub-problems, and the procedure group for solving the sub-problems corresponding to each node of the tree structure A knowledge base (FIG. 1; 12) for storing a knowledge base and a tree structure of the knowledge base, and a procedure for creating a schedule by sequentially executing procedures corresponding to the nodes obtained by the search (FIG. 1; FIG. 1; 14) and.

[0006]

According to the present invention, the knowledge about the schedule creation procedure representing the problem solving process of the expert is stored in the knowledge base as a tree structure. The depth-first search of this tree structure enables control of inference similar to the problem solving process of the expert.

Generally, in a schedule creation system,
Knowledge of schedule creation procedure and schedule allocation method is required. In the present invention, since these knowledge items are stored separately in the knowledge base, the independence of both knowledge items is increased, and the knowledge base item is easily created and maintained.

Further, according to the present invention, knowledge about a schedule creating procedure is expressed as a tree structure. Individual schedule assignment work will be performed at the leaf nodes of the tree. Since this work is a repetition of the process in the node that is a sibling of the leaf node, in the present invention,
It is possible to collectively describe the processing in the leaf node in the node that is the parent of the leaf node. In this way, it is not necessary to define individual leaf nodes before entering the schedule creation process, and the required number of leaf nodes are created during the schedule creation process.

[0009]

1 is a block diagram showing the schematic construction of an embodiment of the present invention. This schedule creation support device is a database 11 that stores various data groups necessary for creating a schedule and data of a schedule creation result.
And a knowledge base 12 that is a storage unit that stores knowledge necessary for inference and that includes a schedule assignment procedure storage unit 121 that stores a schedule assignment procedure and a schedule procedure storage unit 122 that stores a tree structure that represents a schedule creation procedure. , A tree structure management unit 13 that manages operations on a tree structure that represents a schedule creation procedure, and a database 1
A schedule creation unit 14 that creates a schedule using the data of No. 1, the data of the knowledge base 12, and the data of the tree structure management unit 13.

Next, the process of creating a schedule with the above configuration will be described by taking the creation of a work schedule of a nurse in a hospital as an example. (1) Examples of targeted problems Described is the problem of creating a work schedule for nurses. This problem is to assign the work schedule shown in FIG. 2 to three types of work, namely day shift, night shift, and holiday.

(2) Schedule Creation Procedure First, let us consider the procedure for solving this problem. In this example, there are three types of work patterns, so the problem of "nurse's work schedule" is "assignment of day shift".
And "night assignment" and "holiday assignment" can be subdivided into three sub-problems. Assuming that the target period for schedule creation is one month, each of the sub-problems described above can be further subdivided into 30 sub-problems of "allocation of day shift for x days". The solution process of this problem is represented by a tree structure as shown in FIG. Next, we define a schedule creation procedure tree structure corresponding to the problem solving process.
In the present invention, since the node corresponding to the leaf node of the problem solving process tree structure does not have to be defined, the schedule creation procedure tree structure is as shown in FIG.
It is defined as the flow of processing in the order of holiday assignment and day shift assignment. The schedule creation procedure tree structure is stored in the schedule procedure storage unit 122 in the knowledge base.

FIG. 5 shows the flow of processing for creating the schedule creation procedure tree structure of FIG. In the first step (step S51), a node corresponding to the subproblem is created. The following instructions are used to create a node. CreateNode (node_name)

In the second step (step S52), the attribute value of the node created in step S51 is set. In this embodiment, the node has two types of attributes, “repeat” and “process”. The command for instructing the setting of the attribute value is shown below. SetNodeAttribute (node_nam
e, attribute_name, attribute
ute_value)

In the third step (step S53), the vertical relation of the nodes created in step S51 is defined. The instructions used for this are: CreateLink (upper_node_nam
e, lower_node_name)

When the above-mentioned instruction is used, the contents of processing for creating the schedule creation procedure tree structure of FIG. 4 are as shown in FIG. This tree structure creation processing is executed by the tree structure management unit 13 prior to the schedule creation.

(3) Schedule assignment procedure Each node appearing in the tree structure of FIG. 4 defines a procedure for solving the corresponding subproblem. These procedures are stored in the schedule allocation procedure storage unit 1 in the knowledge base 12.
21 is stored. Schedule allocation procedure storage unit 12
The structure of 1 is shown in FIG. As shown in FIG. 7, the schedule assignment procedure storage unit 121 includes a node name in a tree structure, a table including a pointer to a procedure to be executed in the node, a node name, a repetition condition, and a content of a procedure process. And a procedure content storage section including.

(4) Process Flow in Schedule Creation Unit In the present invention, the process flow in the schedule creation unit 14 is based on the depth-first search of the tree structure shown in FIG. Here, the flow of processing of the schedule creating unit 14 in this problem will be described along this flow. FIG. 8 is a diagram showing the flow of the processing. The search is started from the top node, that is, the node called "Nurse's work schedule" (step S80).
Then, it is checked whether or not the node has a child node (step S81). If there is a child node, the child node of this node is searched next (step S82).
If there is no child node, it is checked whether there is a leaf node (step S83). A flag for the leaf node is prepared for the node, and the presence or absence of the leaf node can be known by checking this flag. In the example,
Since the node "Nurse's work schedule" has child nodes, in step 82, the "children's assignment" node, which is the first child node, is searched.

When it is determined in step S83 that there is a leaf node, it is checked whether the repeating condition is satisfied (step S84). As for this check, when the schedule procedure storage unit 122 is viewed, the “night shift assignment” node assigns the night shift from 1 to 30 days as shown in FIG. 7. However, since the node (leaf node) for allocating the night shift for each day has not been created, first, the "allocation for night shift for one day" node is created (step S85). Next, the “allocation of night shift” node is searched (step S86), and the “allocation of night shift” stored in the schedule procedure storage unit 122 is searched.
Use the procedure to assign a night shift (step S8)
7). When this process ends, the parent node is searched (step S88). In the example, the "night shift assignment" node, which is the parent node of the "1 day shift assignment" node, is searched again.

A process similar to this one day process is performed from two days to three days.
By performing until the 0th day, the processing in the "night shift assignment" node ends. That is, when the allocation process is completed up to 30 days, it is understood that the repeating condition is not satisfied in step S84 and the allocation process in the "night shift allocation" node is completed, and then the parent node is searched (step S88). In the example, the node searched is moved to the "Nurse's Work Schedule" node.

Next, in step S81, it is checked whether or not there is any unsearched child node. In the above example, since there is a child node, the next child node "holiday allocation" is searched in step S82. In subsequent steps S81 to 87, holiday allocation processing is performed by the same processing as the night shift allocation processing described above. Furthermore, the same processing as the "night shift assignment" node is performed in the "day shift assignment" node. As the process proceeds, it is determined that the searched node has no unsearched child node in the determination of step S81, does not have a leaf node in the determination of step S83, and is the highest node in the determination of step S89. When this happens, the search for all nodes is completed, and the nurse's work schedule is completed.

As an example of the "execution of schedule assignment work" (step S87) shown in FIG. 8, assignment of the night shift will be described. At the workplace handled in this example, the number of nurses is 20 and the number of night shifts on each day is 2. FIG. 9 shows the flow of night shift assignment processing based on the knowledge of a specialist who assigns a work schedule in this workplace.

Step S91 is a process of collecting people who have a long interval from the previous night shift. As a result of the execution of this step, Mr. A and Mr. B, Mr. C, Mr. D, and Mr. E are selected as candidates for night shift assignment.

Step S92 is a process for sorting the set of candidates obtained in step S91 in ascending order of the number of night shifts. As a result of executing this step, candidates for night shift assignment are {A, C, D, B, E} in descending order of priority.

In step S93, the order of priority is as follows:
This is a process of selecting two people to which the night shift is assigned. As a result of the execution of this step, the night shift of the day will be assigned to Mr. A and Mr. C.

The night shift assignment work from 2 to 30 days
The processing content is as described above.

According to the present invention, the knowledge about the schedule creation procedure representing the problem solving process of the expert is stored in the knowledge base as a tree structure. The depth-first search of this tree structure enables control of inference similar to the problem solving process of the expert.

Further, according to the present invention, the schedule allocation procedure is prepared corresponding to the nodes in the tree structure, and therefore the knowledge about the schedule creation procedure and the knowledge about the schedule allocation method are separately stored in the knowledge base. The independence of the knowledge of both parties becomes high, and the maintenance of the knowledge base becomes easy.

When the knowledge about the schedule creation procedure is expressed as a tree structure as in the present invention, individual schedule assignment work is carried out at the leaf nodes of the tree. Since this work is a repetition of the process in the node that is the sibling of the leaf node, in the present invention, the process in the leaf node can be collectively described in the node that is the parent of the leaf node.
In this way, it is not necessary to define individual leaf nodes before entering the schedule creation process, and the required number of leaf nodes are created during the schedule creation process. Therefore, the present invention facilitates the creation of a knowledge base.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a work schedule.

FIG. 3 is a diagram showing an example of a tree structure representing a problem solving process.

FIG. 4 is a diagram showing a schedule creation procedure tree structure of the present invention.

FIG. 5 is a diagram showing a flow of processing for creating a schedule creation procedure tree structure.

FIG. 6 is a diagram showing processing contents for creating a schedule creation procedure tree structure.

FIG. 7 is a diagram showing the structure of a schedule assignment procedure storage unit.

FIG. 8 is a diagram showing a flow of diaper processing in the schedule creation unit.

FIG. 9 is a diagram showing a flow of night shift assignment processing based on expert knowledge.

[Explanation of symbols]

11 ... Database, 12 ... Knowledge base, 121 ... Schedule allocation procedure storage section, 122 ... Schedule procedure storage section, 13 ... Tree structure management section, 14 ... Schedule creation section, 25 ... I / O controller.

Claims (1)

[Claims] 1. A schedule creation problem is divided into a plurality of parts, each sub-problem obtained by the division is expressed as a node, and a node corresponding to a sub-problem and a sub-problem higher than the sub-problem are expressed. A knowledge base that stores information about a schedule creation procedure including a tree structure obtained by connecting corresponding nodes and a procedure group that solves a partial problem corresponding to each node of the tree structure; and the tree of the knowledge base. By searching the structure and sequentially executing the procedure corresponding to the node obtained by the search,
A schedule creation support device comprising: a schedule creation means for creating a schedule.