Distribution drawing apparatus, distribution drawing processing method and program

Abstract

To suitably simplify a graph expressed using nodes and edges.SOLUTION: A distribution drawing apparatus comprises: a data acquisition unit for acquiring actual data in a distribution network; a node aggregation unit for creating an aggregated graph structure by aggregating nodes on the basis of the similarity between two nodes in a graph structure with respect to the graph structure showing a connection between steps constituting the distribution network representing the actual data; and a graph output unit for outputting the aggregated graph structure.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a logistics drawing device, a logistics drawing processing method, and a program for processing a graph represented using nodes and edges. [Background technology]

[0002] A graph is a collection of nodes and edges, and can model the interrelationships (networks) between the nodes. Such graphs make it possible to represent the elements (nodes) and connections (edges) of, for example, a transportation network or a logistics network. Based on the graph, it is possible to perform analyses such as efficiency and productivity.

[0003] For example, in the steelmaking process, the final product is obtained through the processing of intermediate materials in many processes. In order to improve productivity, it is important to maximize the operating rate of the equipment in each process, and in order to improve the operating rate of the equipment, it is necessary to accurately predict and control the inventory status of the storage area where intermediate materials are temporarily stored between the equipment. For example, if there is a shortage of materials in the front storage area where materials to be processed by the equipment are stored, the equipment will not have enough materials to process, which will cause the equipment to stop operating. In addition, if the rear storage area where processed materials completed by the equipment are stored becomes congested, the equipment will not be able to discharge the processed materials, which will lead to a decrease in the operating rate of the equipment. For this reason, by visualizing the logistics in the steelmaking process using graphs, it is possible to identify the equipment, storage areas, and transport equipment that are bottlenecks, and to consider measures necessary to improve productivity.

[0004] If all the configurations and connections of the network to be graphed are represented in the graph as is, the more complex the network, the more complicated the graph becomes, and the lower the visibility of the graph becomes. In addition, when a simulation is performed based on the created graph, the simulation load also increases. For this reason, the graph is simplified by aggregating or deleting nodes so that information in the graph is not lost.

[0005] For example, there is a network analysis technique that excavates characteristic subgraphs from a large-scale graph. For example, in the field of graph theory, a network analysis technique has been proposed that simplifies a graph by maximizing an index (modularity) that represents the quality of clustering in the Min-cut method, which divides nodes so that the number of edges within a cluster is maximized and the number of edges between clusters is minimized. Patent Document 1 also discloses a technique in which a node that has only one edge included in input graph data is considered to be in the same cluster as the node and its adjacent nodes, and is aggregated into one node. Patent Document 2 discloses a technique in which graph nodes are divided into N groups, the graph is classified into clusters from the N groups according to modularity, and multiple nodes that are in the same cluster are aggregated into one node. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2013-156698 A [Patent Document 2] JP 2017-204161 A Summary of the Invention [Problem to be solved by the invention]

[0007] However, the network analysis techniques described in Patent Documents 1, 2, etc. focus only on the relationship between two nodes and aggregate the graph structure according to an index that represents the quality of clustering. When the original graph represents a network with a passing order and flow, such as a logistics network, the graph simplified by the techniques described in Patent Documents 1, 2, etc. (hereinafter also referred to as a "simplified graph") does not take into account the passing order of the nodes, resulting in a loss of information in the original graph. As a result, the state of the network cannot be correctly recognized using the simplified graph, or the accuracy of the analysis decreases.

[0008] Therefore, the present invention has been made in consideration of the above problems, and an object of the present invention is to provide a logistics drawing device, a logistics drawing processing method, and a program that are capable of appropriately simplifying graphs represented using nodes and edges. [Means for solving the problem]

[0009] In order to solve the above problems, according to one aspect of the present invention, there is provided a logistics drawing device comprising: a data acquisition unit that acquires performance data in a logistics network; a node aggregation unit that generates an aggregated graph structure by aggregating nodes based on the similarity between two nodes in a graph structure showing connections between processes that make up the logistics network, the graph structure representing the performance data; and a graph output unit that outputs the aggregated graph structure.

[0010] The node aggregation unit may calculate the similarity based on at least one of the following: the physical distance between two nodes, the temporal distance between two nodes, the distance of the probability distribution of the inflow from a parent node to a child node, or the distance of the probability distribution of the outflow to the child node from a parent node.

[0011] Furthermore, the node aggregating unit may determine the similarity between two nodes based on whether or not their logistics positions in the graph structure are the same.

[0012] The node aggregation unit may aggregate nodes that represent the same function among the nodes that are determined to be located at the same logistical location.

[0013] The node aggregating unit may aggregate nodes of performance data of the same product type.

[0014] The graph output unit may cause the output device to display the graph represented by the aggregate graph structure so that the aggregated contents of the nodes aggregated into the node after aggregation by the node aggregating unit can be identified.

[0015] At this time, the graph output unit may display the aggregated contents of the aggregated nodes at the aggregated node of the graph displayed on the output device.

[0016] In addition, in order to solve the above problem, according to another aspect of the present invention, a logistics drawing device is provided that displays an aggregate graph structure that aggregates at least a portion of a plurality of nodes that constitute a graph representing actual logistics data in a logistics network, in a manner that makes it possible to identify the aggregated contents of the nodes aggregated in the aggregated node of the aggregate graph structure.

[0017] Furthermore, in order to solve the above problem, according to another aspect of the present invention, there is provided a logistics drawing processing method including a data acquisition step of acquiring performance data in a logistics network, a node aggregation step of generating an aggregated graph structure by aggregating nodes in a graph structure showing connections between processes that make up the logistics network, representing the performance data, based on the similarity between two nodes in the graph structure, and a graph output step of outputting the aggregated graph structure.

[0018] In addition, in order to solve the above-mentioned problems, according to another aspect of the present invention, there is provided a program for causing a computer to function as a logistics drawing device, comprising: a data acquisition unit that acquires performance data in a logistics network; a node aggregation unit that generates an aggregate graph structure by aggregating nodes in a graph structure showing connections between processes that make up the logistics network, representing the performance data, based on the similarity between two nodes in the graph structure; and a graph output unit that outputs the aggregate graph structure. Effect of the Invention

[0019] As described above, according to the present invention, a graph represented using nodes and edges can be appropriately simplified. [Brief description of the drawings]

[0020] [Figure 1] 1 is a block diagram showing a logistics drawing device according to an embodiment of the present invention; [Diagram 2] 11 is a flowchart showing an example of a logistics drawing processing method according to the embodiment. [Diagram 3] FIG. 2 is an explanatory diagram illustrating an example of a performance data table for storing performance data. [Figure 4] FIG. 11 is an explanatory diagram showing an adjacency matrix in which a graph structure is converted into a matrix, as an example of graph structure data. [Diagram 5] FIG. 5 is an explanatory diagram showing an example of a graph represented by the adjacency matrix of FIG. [Figure 6] FIG. 2 is an explanatory diagram illustrating a physical distance between two nodes, which is an example of the similarity between two nodes. [Figure 7] FIG. 13 is an explanatory diagram illustrating the distance between an inflow probability distribution and an outflow probability distribution, which is an example of the similarity between two nodes. [Figure 8] FIG. 2 is an explanatory diagram illustrating a time distance between two nodes, which is an example of the similarity between two nodes. [Figure 9] FIG. 13 is an explanatory diagram for explaining the aggregation of nodes of performance data of the same product type. [Figure 10] FIG. 13 is an explanatory diagram showing an example of displaying an aggregation node in a simplified graph. [Figure 11] FIG. 13 is an explanatory diagram showing another example of displaying an aggregation node in a simplified graph. [Figure 12] FIG. 13 is an explanatory diagram showing another example of displaying an aggregation node in a simplified graph. [Figure 13] FIG. 13 is an explanatory diagram showing another example of displaying an aggregation node in a simplified graph. [Figure 14] 1 is a block diagram showing an example of a hardware configuration of a logistics drawing device according to an embodiment of the present invention; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0021] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In this specification and the drawings, components having substantially the same functional configurations are designated by the same reference numerals, and duplicated explanations will be omitted.

[0022] [1.Logistics drawing device] First, a logistics drawing device 100 according to an embodiment of the present invention will be described with reference to Fig. 1. Fig. 1 is a block diagram showing the logistics drawing device 100 according to this embodiment.

[0023] A logistics drawing device 100 according to this embodiment includes a data acquisition unit 110, a graph structure acquisition unit 120, a node aggregating unit 130, and a graph output unit 140, as shown in FIG.

[0024] (Data Acquisition Section) The data acquisition unit 110 acquires performance data R of logistics in the logistics network. In the embodiment shown in FIG. 1, the data acquisition unit 110 acquires performance data R from a data storage unit 200 that stores performance data R of logistics in the logistics network. The data storage unit 200 stores the passing order of processes represented by the nodes and edges of a graph as performance data R (see FIG. 3 described later). The nodes represent processes (e.g., equipment, storage yard, transport equipment, etc.), and the directed edges indicate the connections between the nodes and the movement direction of the logistics object. For example, the data storage unit 200 stores the order of processes passed by a coil (logistics object) manufactured in a steelmaking process as performance data R of logistics in the logistics network (see FIG. 3 described later). The data acquisition unit 110 acquires a predetermined number of performance data R or performance data R for a predetermined period.

[0025] The data acquiring unit 110 outputs the acquired performance data R to the node aggregating unit 130. In addition, the logistics drawing device 100 may be configured such that the data acquiring unit 110 outputs the acquired performance data R to the graph structure acquiring unit 120, as shown in FIG.

[0026] (Graph structure acquisition part) The graph structure acquisition unit 120 acquires the graph structure of the logistics network to be processed. The graph structure is information that represents connections between multiple processes included in the logistics network, such as graph structure data S and graph G described below. In the embodiment shown in Fig. 1, the graph structure acquisition unit 120 analyzes the performance data R input from the data acquisition unit 110 and generates a graph structure corresponding to this performance data R, thereby acquiring the graph structure.

[0027] More specifically, the graph structure acquisition unit 120 generates graph structure data S representing a directed graph that represents the relationships between nodes using nodes and edges, from the performance data R. The graph structure data S generated by the graph structure acquisition unit 120 is data representing the graph G before the nodes are aggregated by the node aggregating unit 130 described later. The format of the graph structure data S is not particularly limited, and for example, an adjacency matrix (see FIG. 4 described later) in which the graph structure is converted into a matrix may be used as the graph structure data S. The graph structure acquisition unit 120 generates graph structure data S representing the order of passing through the processes represented by the performance data R, and outputs it to the node aggregating unit 130.

[0028] Note that the present invention is not limited to this example. For example, the graph structure acquisition unit 120 may acquire graph structure data S stored in advance in a storage device such as the data storage unit 200 from the storage device. Alternatively, the graph structure acquisition unit 120 may acquire a graph structure from a graph G, or may acquire a graph G recorded in a storage device such as the data storage unit 200 and generate the graph structure data S by analyzing the graph G.

[0029] (Node aggregation section) The node aggregating unit 130 aggregates nodes based on the similarity between two nodes for the graph structure representing the performance data R. Specifically, as described below, the node aggregating unit 130 calculates the similarity based on the relationship between the nodes obtained based on the performance data R or the graph structure (graph structure data S), and generates an aggregated graph structure (aggregated graph structure data Sa) that simplifies the input graph structure by aggregating the nodes based on the calculated similarity. In the example shown in FIG. 1, the graph structure representing the performance data R is the graph structure data S acquired by the graph structure acquisition unit 120. The node aggregating unit 130 aggregates nodes of the graph structure data S representing the performance data R based on the similarity between two nodes, and generates aggregated graph structure data Sa.

[0030] When the graph represents a network with a passing order and flow, such as a logistics network, each node in the graph has a role in logistics, such as a facility, a storage area, or a transport device. Therefore, by aggregating nodes with similar roles, the graph can be appropriately simplified without compromising the passing order of the nodes in the graph before simplification. The node aggregating unit 130 according to this embodiment identifies nodes with similar roles using the similarity between two nodes.

[0031] For example, the node aggregation unit 130 may calculate the similarity based on at least one of the physical distance between two nodes, the time distance between two nodes, the distance of the probability distribution of the inflow amount from a parent node to a child node, the distance of the probability distribution of the outflow amount to a child node from a parent node, or the difference or similarity of functions. A detailed description of the calculation process of the similarity will be given later.

[0032] The node aggregating unit 130 outputs the aggregate graph structure data Sa generated by aggregating the nodes to the graph output unit 140.

[0033] (Graph output section) The graph output unit 140 outputs the aggregate graph structure generated by the node aggregating unit 130 through aggregation to an external output device 300. That is, the graph output unit 140 outputs at least one of the simplified graph Ga or the aggregate graph structure data Sa to the output device 300. Here, the simplified graph Ga is a graph that depicts an aggregate graph structure that aggregates at least some of the processes (nodes) of the graph structure of the logistics network to be processed, and is a graph that simplifies the graph structure of the logistics network to be processed. When outputting the simplified graph Ga, the graph output unit 140 depicts the simplified graph Ga based on the aggregate graph structure data Sa and outputs it to the output device 300.

[0034] The output device 300 is, for example, a display or a printer, and is communicatively connected to the logistics drawing device 100. The output device 300 displays the simplified graph Ga input from the logistics drawing device 100. This allows the user to visually recognize the simplified graph Ga, and to easily grasp the state of the network at each point in time from the displayed simplified graph Ga.

[0035] The graph output unit 140 may display the simplified graph Ga on the output device 300 so that the aggregated contents of the nodes aggregated by the node aggregation unit 130 can be identified. The aggregated contents may be, for example, at least one of information (identification information, etc.) indicating which nodes in the graph structure S are aggregated, and attribute information such as the types of equipment, storage sites, transport devices, etc. for the aggregated nodes and any process information associated therewith. For example, the graph output unit 140 may display identification information of the aggregated nodes in a node (hereinafter also referred to as an "aggregated node") that aggregates multiple nodes in the simplified graph Ga. At this time, for example, the graph output unit 140 may display a pop-up display of the aggregated contents of the nodes aggregated in the aggregated node when the cursor is placed on the aggregated node in the simplified graph Ga. Furthermore, the graph output unit 140 may display the line type, color, shape, etc. of the frame of the aggregated node so as to be different from that of the non-aggregated node (for example, see FIGS. 10 to 13 described later).

[0036] An example of the configuration of the logistics drawing device 100 according to the present embodiment has been described above. In the above description, the data acquisition unit 110, the graph structure acquisition unit 120, the node aggregation unit 130, and the graph output unit 140 are provided in the same device, but the present invention is not limited to this example. These functional units may be configured by multiple physically different devices connected by a communication line such as a communication network.

[0037] It is also possible to create a computer program for implementing each function of the logistics drawing device 100 and install it in a personal computer or the like. It is also possible to provide a computer-readable recording medium in which such a computer program is stored. The recording medium is, for example, a magnetic disk, an optical disk, a magneto-optical disk, a flash memory, or the like. The above computer program may also be distributed, for example, via a network, without using a recording medium.

[0038] [2. Logistics drawing processing method] Next, a logistics drawing processing method according to an embodiment of the present invention will be described with reference to Fig. 2. Fig. 2 is a flowchart showing an example of the logistics drawing processing method according to this embodiment. The flowchart shown in Fig. 2 shows an example of the logistics drawing processing method performed using the logistics drawing device 100 shown in Fig. 1. In the following description, a logistics network of coils manufactured in a steelmaking process will be considered as a specific example. A graph representing the logistics network shows the order of steps (facilities, storage areas, transport equipment) through which the coils pass in the manufacturing process.

[0039] (S100: Acquire performance data) First, as shown in Fig. 2, the data acquisition unit 110 acquires performance data R from the data storage unit 200 (S100). The data storage unit 200 stores performance data R in the network. Fig. 3 shows an example of a performance data table 210 that stores performance data R of the logistics network of coils.

[0040] The performance data R shown in Fig. 3 indicates three processes that a coil passes through in order, and the coil number (CoilNo), which is the coil's identification number, is associated with the first process, second process, and third process that the coil passes through. For example, it indicates that a coil with coil number C0000 passes through the processes in the order of process A (node ​​A), process AGV1 (node ​​AGV1), and process C (node ​​C).

[0041] Moreover, the performance data R may further include information that can be used to calculate the similarity. For example, the performance data R shown in FIG. 3 includes, for each process, coordinates, which are position information, a process start time, and a process end time. The coordinates may be, for example, position coordinates with an arbitrarily determined reference position P0 as the origin. For example, in the performance data R in FIG. 3, when the process A is set as the reference position P0 (see the right side of FIG. 6 described later), the coordinates (x1, y1) of the process A become the origin (0, 0), and the coordinate values ​​of the other processes are determined based on the process A. Alternatively, the coordinates may be information that indicates the absolute position where the equipment, etc., is installed, such as latitude and longitude. The process start time is the time when the coil is carried into the process. The process end time is the time when the coil is carried out of the process. The process start time and the process end time can be used to obtain the time when the coil stayed in the process.

[0042] Furthermore, the performance data R may include a product number indicating the type of coil, as shown in Fig. 3. By including a product number in the performance data R, it becomes possible to identify performance data R of the same type. For example, the product number of the coil with coil number C0002 is "X", which indicates that it is the same type as the coil with coil number C0000.

[0043] Data acquiring unit 110 acquires a predetermined number of pieces of performance data R or a predetermined period of time from data storage unit 200, and outputs the data to graph structure acquiring unit 120 to generate graph structure data S, and also outputs the data to node aggregating unit 130 to aggregate nodes. Node aggregating unit 130 receives as input graph structure data S input from graph structure acquiring unit 120 and performance data R input from data acquiring unit 110, and generates aggregated graph structure data Sa.

[0044] Note that, when the graph structure data S representing the performance data R acquired by the data acquiring unit 110 is obtained in advance and the graph structure data S can be acquired, the graph structure acquiring unit 120 may acquire the graph structure data S from another storage device or the like that stores the graph structure data S. In such a case, the data acquiring unit 110 may output the acquired performance data R only to the node aggregating unit 130.

[0045] (S110: Obtaining graph structure) Next, the graph structure acquisition unit 120 acquires a graph structure of the logistics network to be processed. In the embodiment shown in Fig. 2, the graph structure acquisition unit 120 acquires a graph structure representing connections between processes included in the logistics network to be processed, using the performance data R input from the data acquisition unit 110 (S110). Specifically, the graph structure acquisition unit 120 generates graph structure data S representing a directed graph that represents relationships between nodes using nodes and edges, from the performance data R. The graph structure data S generated by the graph structure acquisition unit 120 is data representing the graph G before the nodes are aggregated by the node aggregation unit 130 described later.

[0046] For example, the graph structure data S may be an adjacency matrix Sm obtained by converting a graph structure as shown in FIG. 4 into a matrix. The adjacency matrix Sm shown in FIG. 4 represents the graph structure data S corresponding to the graph G shown in FIG. 5. Each row and each column of the adjacency matrix Sm shown in FIG. 4 represents a node of the graph. That is, the adjacency matrix Sm in FIG. 4 indicates that the graph representing the logistics network is composed of nodes representing five processes A to E and nodes representing five conveyance devices AGV1 to AGV5, as shown in the row labels and column labels of the adjacency matrix Sm. In addition, the graph according to this embodiment is a directed graph. When the graph structure data S is the adjacency matrix Sm shown in FIG. 4, for example, the direction of the edge of the graph can be represented by defining the row node as the start point of the edge and the column node as the end point of the edge.

[0047] Each component of the adjacency matrix Sm indicates the presence or absence of an edge connecting the node indicated by the row to the node indicated by the column. The value of the component represents the volume of goods flow in the logistics network. When the value of a component is 0, there is no edge connecting the node indicated by that row to the node indicated by the column. On the other hand, when the value of a component is other than 0, there is an edge connecting the node indicated by that row to the node indicated by the column.

[0048] Moreover, the values ​​of the components of the adjacency matrix Sm shown in FIG. 4 represent the number of coils that have moved between nodes (processes) connected by edges as the logistics amount. In the graph G shown in FIG. 5, the numbers written corresponding to the edges connecting the nodes represent the values ​​of the components of the adjacency matrix Sm shown in FIG. 4. The logistics amount (the number of coils in this example) may be included in the performance data R or may be acquired from the performance data R. That is, in this embodiment, the logistics amount is obtained based on the number of rows (the number of records) in which the combination of the start process and the end process is the same between adjacent processes (for example, the first process and the second process, etc.) in the performance data R, but the performance data R may include information indicating the logistics amount between adjacent processes.

[0049] The graph structure data S is not limited to the adjacency matrix Sm, and may be any information capable of reproducing the graph structure. The graph structure acquisition unit 120 generates graph structure data S representing the passing order of the processes represented by the performance data R, and outputs the data to the node aggregation unit 130.

[0050] As described above, when the graph structure data S representing the performance data R to be acquired by the data acquisition unit 110 is obtained in advance and the logistics drawing device 100 is capable of acquiring the graph structure data S, the graph structure acquisition unit 120 may acquire the graph structure data S of the performance data R directly from the data storage unit 200, etc., without generating it.

[0051] (S120-S130: Calculating similarity between nodes) Then, the node aggregating unit 130 calculates the similarity between the two nodes for the graph structure representing the performance data R (S120).

[0052] As will be described later in step S130, the node aggregating unit 130 aggregates nodes of a graph structure (graph structure data S) representing performance data R based on the similarity between two nodes, and generates aggregated graph structure data Sa corresponding to the simplified graph Ga. The similarity represents the closeness of the roles of the nodes, and the closer the distance (i.e., the higher the similarity of the nodes), the more similar the roles of the nodes are. The similarity between two nodes can be indicated by various indices. Examples of indices indicating the similarity are described below.

[0053] (1) The physical distance between two nodes For example, the physical distance between two nodes may be used as the similarity between two nodes. For example, multiple pieces of equipment performing the same process or multiple storage areas for storing products manufactured through the same process are often located adjacent to each other. For example, in a graph as shown on the left side of FIG. 6, the physical positional relationship of each node is shown diagrammatically as shown on the right side of FIG. 6. From the diagram on the right side of FIG. 6, it can be seen that the physical distance between node D and node E is close. Therefore, it is estimated that node D and node E have similar roles and can be aggregated. Similarly, node H and node I are physically close to each other, so they can be aggregated.

[0054] Therefore, the node aggregating unit 130 may calculate the distance between two processes from the coordinates representing the position of each process included in the performance data R (see FIG. 3) as the physical distance between two nodes, and may use the calculated physical distance between the two nodes as the similarity between the nodes. For example, the positional relationship (physical distance) in the schematic diagram on the right side of FIG. 6 can be calculated based on the coordinates, which are the position information of the process included in the performance data R. In the example of FIG. 3, each performance data R includes the coordinates, but the performance data R does not necessarily need to include the coordinates. For example, the node aggregating unit 130 may calculate the physical distance between two nodes using a table indicating the coordinates for each process, which is prepared separately from the performance data table 210.

[0055] (2) The distance between the probability distributions of two nodes As another example, the similarity between two nodes may be a distance between probability distributions calculated from at least one of the inflow probability distribution p and outflow probability distribution q of the two nodes that are targets for determining whether to aggregate (hereinafter also referred to as "aggregation judgment targets"). Specifically, the similarity may be a distance between at least one of the inflow probability distribution p, the outflow probability distribution q, and the distance between the inflow and outflow probability distributions that are probability distributions including the inflow probability distribution p and the outflow probability distribution q of the two nodes that are aggregation judgment targets.

[0056] The inflow probability distribution p(x) is a distribution that indicates the proportion (probability) of the total inflow from all parent nodes to node x between an arbitrary node (child node) x and all upstream nodes (parent nodes) directly connected to node x by edges.

[0057] The inflow probability distribution p can be calculated by the following procedure. First, any two or more nodes (child nodes) to be aggregated are selected, and all nodes (parent nodes) connected one level upstream of all selected child nodes are extracted. Then, for each child node, the inflow amount from each parent node is calculated, and the ratio of the inflow amount from each parent node to the total inflow amount to the child node is calculated as a probability value (inflow probability). In other words, when there are Nc child nodes and Np parent nodes, the inflow probability of the jth parent node (j=1, 2, , Np) to the ith child node (i=1, 2, , Nc) is calculated as Qj / ΣQj, where Qj is the inflow amount from the jth parent node to the ith child node. The inflow probability distribution p indicates the distribution of the inflow probability Qj / ΣQj.

[0058] Similarly, the outflow probability distribution q(x) is a distribution that indicates the proportion (probability) of the total outflow from an arbitrary node (parent node) x to all downstream nodes (child nodes) directly connected to the node x by edges that flows to each child node.

[0059] The outflow probability distribution q can be calculated by the following procedure. First, select any two or more nodes (parent nodes) to be aggregated and extract all nodes (child nodes) connected one level downstream of all selected parent nodes. Then, for each parent node, calculate the outflow amount to each child node and calculate the ratio of the outflow amount to each child node to the total outflow amount from the parent node as a probability value (outflow probability). In other words, if there are Nc child nodes and Np parent nodes, the outflow probability from the kth parent node (k=1, 2, , Np) to the lth child node (l=1, 2, , Nc) can be calculated as Ql / ΣQl, where Ql is the outflow amount from the kth parent node to the lth child node. The outflow probability distribution q shows the distribution of the outflow probability Ql / ΣQl.

[0060] The inflow / outflow probability distribution of node x is a distribution that combines the inflow probability distribution p(x) and outflow probability distribution q(x) of node x into a single probability distribution.

[0061] By calculating the inflow probability and outflow probability in this way, the node aggregating unit 130 calculates the inflow probability distribution p, the outflow probability distribution q, and the inflow / outflow probability distribution composed of these two probability distributions for an arbitrary node. Then, the node aggregating unit 130 calculates the distance between the probability distributions, which are the inflow probability distribution, the outflow probability distribution, or the inflow / outflow probability distribution, of each of the two nodes.

[0062] The distance between two probability distributions may be expressed, for example, by JS Divergence (Jensen-Shannon Divergence), which is one of the indices for determining the similarity between two probability distributions. JS Divergence can be calculated by the following formula (1) using KL Divergence.

[0063]

number

[0064] In the above formula (1), D JS is the value of JS Divergence, D KLindicates the value of KL Divergence, and P and Q indicate the values ​​of the two probability distributions.

[0065] KL Divergence (Kullback-Leibler Divergence) is an index that shows how far one probability distribution is from another, but it is not symmetric and its value can change if the reference probability distribution is switched. In contrast, JS Divergence is an index that gives symmetry to KL Divergence and is suitable for judging the distance between two probability distributions.

[0066] As a specific example, a case where the similarity is the distance between the inflow and outflow probability distributions will be described with reference to Figs. 6 and 7. In the graph on the left side of Fig. 6, the parent nodes of node D are nodes A and B, and the child nodes are G and H. The parent nodes of node E are nodes A, B and C, and the child nodes are G, H and I. The parent nodes of node F are nodes A, B and C, and the child nodes are G and I. The numbers shown in the graph shown on the left side of Fig. 6 indicate the inflow amounts from nodes A, B and C to nodes D, E and F as the inflow number of coils, and the outflow amounts from nodes D, E and F to nodes G, H and I as the outflow number of coils. There is no connection between the three nodes D, E and F.

[0067] In this case, if the three nodes D, E, and F on the left side of Figure 6, which are not connected to each other, are the targets of aggregation judgment, when calculating the inflow probability, the parent nodes of nodes D, E, and F are nodes A, B, and C that are directly connected to these nodes on the upstream side. Also, when calculating the outflow probability, the child nodes of nodes D, E, and F are nodes G, H, and I that are directly connected to these nodes on the downstream side.

[0068] First, in the graph shown on the left side of Fig. 6, the number of coils flowing in from nodes A, B, and C to nodes D, E, and F, and the number of coils flowing out from nodes D, E, and F to nodes G, H, and I are calculated, and the inflow and outflow amount table T in Fig. 7 is obtained. Q The inflow number and outflow number can be obtained from the performance data R or the graph structure data S (for example, the adjacency matrix Sm).

[0069] Here, the inflow and outflow amount table T in FIG. Q In the above, the number of coils flowing into node D from node C, the number of coils flowing out from node D to node I, and the number of coils flowing out from node F to node H are 0. When the amount of inflow or outflow is 0, it is not possible to calculate probabilities such as the probability of inflow or outflow. Therefore, in order to make the probability calculable, the amount of inflow or outflow is replaced with a small value (for example, 1) in such cases.

[0070] Next, the inflow / outflow table of FIG. 7 (the inflow / outflow table after replacement if there is a cell with a zero inflow or outflow number) T Q By calculating the inflow probability from the number of inflows of the coils shown in FIG. 1 as described above, the inflow probability distribution (p(x), x=D, E, F) of each of the nodes D, E, and F can be obtained. Similarly, the inflow and outflow amount table T Q By calculating the outflow probability from the number of outflows of coils shown in FIG. 7, the outflow probability distribution (q(x), x=D, E, F) of each of the nodes D, E, and F can be obtained (see the probability distribution table T in FIG. 7). P For example, the probability distribution table T in Fig. 7 P In this example, if we focus on node E, the inflow probability distribution is p(E) = {0.6, 0.2, 0.2}, and the outflow probability distribution is q(E) = {0.4, 0.4, 0.2}. If the inflow / outflow probability distribution is r(x), then r(E) = {0.6, 0.2, 0.2, 0.4, 0.4, 0.2}.

[0071] In addition, the inflow and outflow table T in FIG. Q In the probability distribution table T P The probability of inflow from node C to node D and the probability of outflow from node D to node I are calculated by replacing the inflow or outflow value with a value between 0 and 1. For example, the probability of inflow from node C to node D is calculated as 40 / (40+10+1)=0.784314.

[0072] Then, JS Divergence is calculated based on the above formula (1) using the outflow / inflow probability distribution, which is a combination of the inflow probability distribution and the outflow probability distribution. As shown in the right-hand side of the above formula (1), JS Divergence is calculated using KL Divergence. KL Divergence is calculated using the values ​​of two probability distributions P and Q and the average value of the probability distributions P and Q. The average of a probability distribution is obtained by calculating the average of corresponding values ​​in the distribution. For example, the inflow probability from node A to nodes E and F is 0.6 and 0.2, respectively, and the average of these is (0.6+0.2) / 2=0.4. Also, for example, the inflow probability from node B to nodes E and F is 0.2 and 0.2, respectively, and the average of these is 0.2. Probability distribution table T in FIG. 7 P The KL Divergence value calculated based on the value of KL ) into the above formula (1), we obtain the JS Divergence (JS Divergence table T in Figure 7). JS ).

[0073] Figure 7 JS Divergence table T JS Therefore, it can be said that nodes D and E, which have the smallest values, are close in distance to each other in probability distribution and have a high degree of similarity. In such a result, of nodes D, E, and F, nodes D and E can be the targets of aggregation. Note that in the example of FIG. 7, three nodes are aggregated, so three JS Divergences are calculated, but it is also possible to aggregate two nodes. In this case, only one JS Divergence is calculated, but it is also possible to aggregate if it is below a predetermined threshold and not aggregate if it exceeds the threshold.

[0074] In this way, the node aggregation unit 130 may obtain the distance of the inflow / outflow probability distribution between two nodes and use it as the similarity between the nodes. In the above description, the distance of the inflow / outflow probability distribution between two nodes is obtained, but the present invention is not limited to this example. For example, the distance of only the inflow probability distribution between two nodes may be used as the similarity between the nodes. In this case, the probability distribution table T in FIG. P The distance of the probability distribution is calculated using the above formula (1) for only the inflow probability distribution of the node. Alternatively, the distance of only the outflow probability distribution between two nodes may be used as the similarity between the nodes. In this case, the probability distribution table T P The distance of the probability distribution is calculated using the above formula (1) for only the outflow probability distribution of .

[0075] (3) The time distance between two nodes Another example of the similarity between two nodes may be the time distance between the two nodes. The time distance between two nodes is the time required to move between the two nodes, and may be expressed by a representative value such as the average construction period or the maximum construction period. The shorter the time between the two nodes, the shorter the time distance is, i.e., the higher the similarity between the two nodes.

[0076] For example, based on the graph structure data S, the node aggregation unit 130 obtains the time required for all coils passing between two adjacent nodes to move between those nodes, and calculates a representative value (for example, average construction period, maximum construction period) as the temporal distance. Here, since the graph is a directed graph, if the directions in which the coils move between two adjacent nodes are different, the temporal distance is calculated for each direction. Two nodes with a close temporal distance can be aggregation targets.

[0077] As a specific example, the graph shown in Fig. 8 shows the average construction period as the time distance between two adjacent nodes. In the graph in Fig. 8, the time distance between node B and node C is the shortest. Therefore, node B and node C can be aggregation targets.

[0078] (4) Indicators showing differences in logistics locations As another example of the similarity between two nodes, an index showing the difference in the logistical positions of the two nodes may be used. For example, when there are multiple routes (hereinafter also referred to as "branch routes") consisting of at least one node between two nodes, the nodes on each branch route may be considered to be in the same logistical position.

[0079] For example, the graph shown in FIG. 9 indicates that there is actual data R passing through "node A-node B-node C" and actual data R passing through "node A-node D-node C". In the graph of FIG. 9, there are two branch routes, each consisting of one node (node ​​B, node D), between two nodes, node A and node C. Node B and node D on the two branch routes, which branch off from node A and merge with node C, can be considered to be located at the same location in terms of logistics. Two nodes that can be considered to be located at the same location in terms of logistics often have the same role. Therefore, an index indicating the difference or similarity of the logistics positions of two nodes may be used as the similarity between the two nodes.

[0080] An index representing the difference between the logistical positions of two nodes may be determined so that, for example, the value is high between two nodes that can be considered to be in the same logistical position, and the value is low between two nodes that are not in the same logistical position. For example, the value of the index may be determined in the range of 0 to 1, with "1" indicating that the nodes can be considered to be in the same logistical position, and "0" indicating that the logistical positions are not the same. The closer the value of the index is to 1, the more the two nodes can be considered to be in the same logistical position, and the two nodes are likely to be targets for aggregation.

[0081] For example, when there are multiple branch routes between two nodes, the node aggregation unit 130 may determine that a node on one branch route and a node on another branch route are in the same location in the logistics, set an index value (e.g., "1"), and use this as the similarity between the nodes.

[0082] In this manner, the node aggregating unit 130 calculates the similarity between each pair of adjacent nodes in the graph.

[0083] The node aggregating unit 130 may obtain the similarity between two nodes by one of the indices such as physical distance, temporal distance, inflow probability distribution, outflow probability distribution, outflow / inflow probability distribution, and an index indicating the difference or similarity of the logistics position, or may obtain the similarity between two nodes by combining the values ​​of two or more indices. For example, in the example of FIG. 8 described above, the node aggregating unit 130 uses the temporal distance between two nodes as an index indicating the similarity between the two nodes. In this case, the node aggregating unit 130 determines whether an edge exists between two arbitrarily selected nodes based on the graph structure data S, and calculates the temporal distance between the two nodes where an edge exists.

[0084] In step S120 of FIG. 2, the node aggregating unit 130 selects any two nodes from the nodes included in the graph structure data S, and calculates the similarity between the selected two nodes. Then, the node aggregating unit 130 judges whether or not the similarity has been calculated for all combinations of two nodes included in the graph structure data S (S130). If there is a combination of two nodes for which the similarity has not been calculated (S130: NO), the process returns to step S120, and the node aggregating unit 130 repeats the processes of steps S120 and S130 until the similarity has been calculated for all combinations of two nodes. Then, when the similarity has been calculated for all combinations of two nodes (S130: YES), the node aggregating unit 130 proceeds to the process of step S140.

[0085] (S140-S170: Node Aggregation) When the node aggregating unit 130 calculates the similarity between two nodes, it aggregates the nodes based on the similarity. First, the node aggregating unit 130 determines whether or not there is a combination of nodes that satisfies an aggregation condition for aggregating two nodes into one (S140). The aggregation condition may be, for example, a condition that two nodes whose similarity between the nodes is equal to or greater than a predetermined threshold value set in advance are aggregated into one node. Note that when the index representing the similarity is distance (i.e., physical distance, distance based on probability distribution, temporal distance, etc.), the closer the distance, the higher the similarity.

[0086] Specifically, in the case where the similarity is the physical distance between two nodes, it can be said that the closer the physical distance, the greater the similarity. Therefore, the node aggregating unit 130 may aggregate the nodes when the physical distance between two nodes is a value equal to or less than a predetermined threshold set in advance. In addition, for example, in the case where the similarity is a distance based on at least one of the inflow probability distribution, outflow probability distribution, and outflow / inflow probability distribution between two nodes, it can be said that the closer the distance, the greater the similarity. Therefore, the node aggregating unit 130 may aggregate the nodes when the distance of the probability distribution between the two nodes (for example, JS Divergence) is a value equal to or less than a predetermined threshold set in advance. Furthermore, for example, in the case where the similarity is the temporal distance between two nodes, it can be said that the closer the temporal distance, the greater the similarity. Therefore, the node aggregating unit 130 may aggregate the nodes when the temporal distance between two nodes is a value equal to or less than a predetermined threshold set in advance.

[0087] Also, for example, when the similarity is expressed by an index indicating the difference in the logistics position, the node aggregating unit 130 may aggregate two nodes whose index is equal to or greater than a predetermined threshold into one node. That is, the node aggregating unit 130 aggregates two nodes that can be considered to be in the same logistics position.

[0088] The threshold value for determining whether or not to aggregate a node may be set appropriately based on the actual state of the network, performance data R from past operations, and the like.

[0089] Furthermore, when it is determined that two nodes are in the same location in terms of logistics, it may be determined whether or not to aggregate the two nodes depending on whether an additional aggregation condition is satisfied. In this case, when it is determined that two nodes are in the same location in terms of logistics from an index indicating the difference or similarity of the locations in terms of logistics and the additional aggregation condition is satisfied, the two nodes are aggregated into one node, and otherwise the two nodes are not aggregated.

[0090] The additional aggregation condition may be, for example, when the roles of the nodes, such as yard or equipment, are the same. When the roles of the nodes are the same, they can be considered to be in the same location in terms of logistics. The additional aggregation condition may also be, for example, when multiple branch routes as shown in FIG. 9 are routes for the same product type, such as routes through which the same steel type or the same order flows. If the routes are for the same product type, the two nodes can be considered to be in the same location in terms of logistics. The additional aggregation condition may be set to satisfy both of these conditions.

[0091] If the node aggregating unit 130 determines that there is no node combination that satisfies the aggregation condition (S140: NO), the process proceeds to step S170. On the other hand, if it determines that there is a node combination that satisfies the aggregation condition (S140: YES), the node aggregating unit 130 aggregates the two nodes of the combination that satisfies the aggregation condition (S150). When two nodes on different routes are aggregated into one node, the inflow amount of the aggregated node is the sum of the inflow amounts of the two aggregated nodes, and the outflow amount of the aggregated node is the sum of the inflow amounts of the two aggregated nodes.

[0092] If there are multiple node combinations that satisfy the aggregation condition, the node aggregating unit 130 may aggregate only one of the combinations in one processing of step S150. For example, the node aggregating unit 130 may aggregate the nodes with the highest similarity between two nodes (i.e., the nodes with the shortest distance).

[0093] Alternatively, the node aggregating unit 130 may aggregate all combinations of nodes that can be aggregated. For example, if a combination of nodes A and B and a combination of nodes B and C are to be aggregated, and nodes A and B are aggregated first, the aggregation of nodes B and C that is about to be performed may result in two nodes becoming one node due to the previous aggregation. In such a case, for example, by associating and retaining information before aggregation with the node after aggregation, the aggregation of nodes A and B may be performed based on the information before aggregation, thereby aggregating nodes A, B, and C may be performed in one step S150. Alternatively, the node aggregating unit 130 may only aggregate the former (e.g., nodes A and B) and proceed to the next step S160 without aggregating the latter (e.g., nodes B and C). In addition, the node aggregating unit 130 may aggregate nodes A and B into one node, calculate the distance again, and calculate the distance between the node after the aggregation of nodes A and B and node C.

[0094] Then, the node aggregating unit 130 updates the graph structure data S to the graph structure data after the node aggregation in step S150 (S160). When the graph structure data is updated, the graph structure data is simplified by the number of aggregated nodes. For example, when the graph structure data S is an adjacency matrix Sm, the number of rows and the number of columns of the adjacency matrix Sm are reduced.

[0095] Thereafter, the node aggregating unit 130 judges whether or not to end the aggregation of the nodes based on the aggregation end condition (S170). The aggregation end condition may be, for example, when there is no combination of nodes that satisfies the node aggregation condition (i.e., when step S140: NO), when the number of aggregated nodes (i.e., the cumulative number of times nodes are aggregated in step S110 and thereafter in step S150) exceeds a preset maximum number of nodes that can be aggregated, when the number of loops of steps S120 to S170, such as the number of processes of steps S120, S130, S140, or S170, reaches a prescribed number, etc. The aggregation end condition may be set by combining two or more of these conditions, or when any one of the conditions is satisfied. When the node aggregating unit 130 judges that the aggregation end condition is not satisfied (S170: NO), it performs the processes from step S120 on the updated graph structure data.

[0096] For example, consider a case where the aggregation condition is to aggregate two nodes with the highest similarity between them, and the aggregation end condition is when there is no combination of nodes that satisfies the node aggregation condition. In this case, while there is a combination of nodes that satisfies the node aggregation condition (i.e., while step S170: NO), the node aggregating unit 130 aggregates the two nodes with the highest similarity between them (S150), updates the graph structure data S (S160), and repeats the process from step S120. Then, when there is no combination of nodes that satisfies the node aggregation condition (i.e., when step S170: YES), the node aggregating unit 130 ends the node aggregation process.

[0097] Also, for example, consider a case where the aggregation condition is to aggregate two nodes with the highest similarity between the nodes, and the aggregation end condition is when the number of aggregated nodes exceeds a preset maximum number of nodes that can be aggregated. Here, the maximum number of nodes that can be aggregated is Nm (Nm: an integer of 1 or more). In this case, while there is a combination of nodes that satisfies the node aggregation condition (i.e., while step S170: NO), the node aggregation unit 130 aggregates the two nodes with the highest similarity between the nodes (S150), updates the graph structure data S (S160), and repeats the process from step S120. Then, even if there is a combination of nodes that satisfies the node aggregation condition, when the number of aggregated nodes reaches the maximum number of nodes that can be aggregated Nm (i.e., when step S170: YES), the node aggregation unit 130 ends the node aggregation process.

[0098] Then, when it is determined that the aggregation end condition is satisfied (S170: YES), node aggregating unit 130 outputs to graph output unit 140 the graph structure data of simplified graph Ga generated by aggregating the nodes (aggregated graph structure data Sa).

[0099] (S180: Graph output) Thereafter, the graph output unit 140 outputs the aggregate graph structure (simplified graph Ga or aggregate graph structure data Sa) generated by the node aggregating unit 130 through aggregation to the external output device 300 (S180). That is, the graph output unit 140 draws a simplified graph Ga based on the aggregate graph structure data Sa, and outputs the simplified graph Ga to the output device 300 such as a display. At this time, the graph output unit 140 may output the aggregate graph structure data Sa itself (for example, the adjacency matrix Sm as shown in FIG. 4) to the output device 300 together with or instead of the simplified graph Ga. In addition, the graph output unit 140 may record the aggregate graph structure data Sa in the data storage unit 200 or an external storage device (not shown) connected to the communication network, read the aggregate graph structure data Sa in response to a user instruction or the like, and output the simplified graph Ga to the output device 300.

[0100] Specifically, the graph output unit 140 draws figures corresponding to the nodes and connection lines representing edges connecting the nodes based on the aggregate graph structure data Sa. At this time, the thickness, color, etc. of the connection lines may be adjusted according to the logistics volume (e.g., the number of coils, etc.) between the nodes. In addition, the graph output unit 140 may be configured to output the graph structure data S before simplification or a graph G representing the graph structure data S.

[0101] The simplified graph Ga represented by the aggregate graph structure data Sa may be displayed on the output device 300 so that the aggregated contents of the nodes aggregated by the node aggregating unit 130 can be identified. As a specific example, Figs. 10 to 14 show display examples of the simplified graph Ga obtained by simplifying the graph G shown in Fig. 5.

[0102] For example, as shown in Fig. 10, the graph output unit 140 may display the aggregated contents of a node (aggregated node) Na, which aggregates multiple nodes in the simplified graph Ga. A process name is displayed at each node in Fig. 10. At this time, the process names of the multiple aggregated nodes are displayed at the aggregated node Na. It can be seen that the aggregated node Na in Fig. 10 is an aggregate of process AGV3 and process AGV5.

[0103] Also, the graph output unit 140 may display the shape of the aggregated node Na in the simplified graph Ga differently from the shapes of the other nodes, as shown in FIG. 11. In the simplified graph Ga in FIG. 11, the non-aggregated nodes are represented by ovals, and the aggregated nodes Na are represented by rectangles. Alternatively, the graph output unit 140 may display the color of the aggregated node Na in the simplified graph Ga differently from the colors of the other nodes, as shown in FIG. 12. In the simplified graph Ga in FIG. 12, the non-aggregated nodes are represented by white, and the aggregated nodes Na are represented by gray. In addition, the graph output unit 140 may display the line type of the border of the node, the size of the node, and the like, differently between the aggregated node Na and the non-aggregated nodes.

[0104] Furthermore, for example, when the cursor is placed on an aggregation node Na in the simplified graph Ga, the graph output unit 140 may display a pop-up showing the aggregated contents of the nodes aggregated in the aggregation node Na. For example, as shown in Fig. 13, when the simplified graph Ga is displayed on the output device 300, when the user moves the cursor 510 and places the cursor 510 on the aggregation node Na, a speech bubble 520 showing the process name of the aggregated node may be displayed.

[0105] In this way, by displaying the graph represented by the aggregated graph structure data Sa on the output device 300 so that the aggregated contents of the aggregated nodes can be identified, the user can grasp the graph information that is missing from the graph structure.

[0106] The logistics drawing processing method according to this embodiment has been described above. According to this embodiment, in a graph that represents performance data R using nodes and edges, nodes are aggregated based on the similarity between two nodes to generate aggregated graph structure data Sa. By aggregating nodes with high similarity in this way, that is, nodes with similar roles, it becomes possible to appropriately present a simplified graph Ga that is highly readable and does not lose the order or flow that represents logistics.

[0107] In the logistics drawing processing method based on the flowchart shown in FIG. 2, the graph structure data update process of step S160 is performed after two nodes are aggregated into one node, but the present invention is not limited to this example. For example, the graph structure data may not be updated, and the aggregation of nodes may be repeatedly performed based on the similarity calculated in step S120. In this case, after performing steps S100 to S140 in FIG. 2, the aggregation of two nodes is repeated in order until an aggregation end condition is satisfied, that is, when there is no combination of nodes that satisfies the aggregation condition (S150, S170). Then, when the aggregation end condition is no longer satisfied (S170: YES), the graph aggregation unit 130 updates the graph structure data that was performed in step S160 in FIG. 2, and generates aggregated graph structure data Sa. After that, the graph output unit 140 may output the aggregated graph structure data Sa (S180).

[0108] [3. Modifications] The logistics drawing device 100 according to an embodiment of the present invention may be a device that includes only the graph output unit 140 among the functional units included in the logistics drawing device 100 shown in FIG. 1. That is, the logistics drawing device 100 displays an aggregated graph structure in which at least a part of a plurality of nodes constituting a graph showing the actual data of logistics in a logistics network is aggregated, so that the aggregated contents of the nodes aggregated in the aggregated node of the aggregated graph structure can be identified. Such a logistics drawing device 100 may, for example, perform the process of step S180 in FIG. 2, and display a simplified graph Ga in which nodes of a logistics network are represented by equipment, storage areas, and transport equipment, and the flow of products between the nodes is represented by edges, by aggregating and simplifying the nodes, so that the equipment, storage areas, and transport equipment before aggregation aggregated in the aggregated node can be identified. In this way, by displaying the simplified graph Ga on the output device 300 so that the aggregated contents of the aggregated nodes can be identified, the user can grasp the information of the graph that is missing in the graph structure.

[0109] [4. Hardware configuration] A hardware configuration of the logistics drawing device 100 according to this embodiment will be described with reference to Fig. 14. Fig. 14 is a block diagram showing an example of a hardware configuration of an information processing device 900 that functions as the logistics drawing device 100 according to this embodiment.

[0110] The information processing device 900 includes a processor (CPU 901 in FIG. 14), a ROM 903, and a RAM 905. The information processing device 900 also includes a bus 907, an input I / F 909, an output I / F 911, a storage device 913, a drive 915, a connection port 917, and a communication device 919.

[0111] The CPU 901 functions as an arithmetic processing device and a control device. The CPU 901 controls all or part of the operations in the information processing device 900 in accordance with various programs recorded in the ROM 903, the RAM 905, the storage device 913, or the removable recording medium 925. The ROM 903 stores programs or arithmetic parameters used by the CPU 901. The RAM 905 temporarily stores programs used by the CPU 901 or parameters that change as appropriate during program execution. These are connected to each other by a bus 907 constituted by an internal bus such as a CPU bus.

[0112] The bus 907 is connected to an external bus such as a PCI (Peripheral Component Interconnect / Interface) bus via a bridge.

[0113] The input I / F 909 is an interface that accepts input from an input device 921, which is an operation means operated by a user, such as a mouse, a keyboard, a touch panel, a button, a switch, and a lever. The input I / F 909 is configured as, for example, an input control circuit that generates an input signal based on information input by a user using the input device 921 and outputs the signal to the CPU 901. The input device 921 may be, for example, a remote control device that uses infrared rays or other radio waves, or an external device 927 such as a PDA (Personal Digital Assistant) that supports the operation of the information processing device 900. A user of the information processing device 900 can operate the input device 921 to input various data to the information processing device 900 and instruct the information processing device 900 to perform processing operations.

[0114] The output I / F 911 is an interface that outputs input information to an output device 923 that can visually or audibly notify a user. The output device 923 may be, for example, a display device such as a CRT display device, a liquid crystal display device, a plasma display device, an EL display device, or a lamp. Alternatively, the output device 923 may be an audio output device such as a speaker and a headphone, a printer, a mobile communication terminal, a facsimile, or the like. The output I / F 911 instructs the output device 923 to output, for example, a processing result obtained by various processes executed by the information processing device 900. Specifically, the output I / F 911 instructs the display device to display the processing result by the information processing device 900 as text or an image. In addition, the output I / F 911 instructs the audio output device to convert an audio signal such as audio data instructed to be played into an analog signal and output it.

[0115] The storage device 913 is one of the storage units of the information processing device 900, and is a device for storing data. The storage device 913 is configured, for example, with a magnetic storage device such as a hard disk drive (HDD), a semiconductor storage device such as a solid state drive (SSD), an optical storage device, or a magneto-optical storage device. The storage device 913 stores the programs executed by the CPU 901, various data generated by the execution of the programs, various data acquired from the outside, and the like.

[0116] The drive 915 is a reader / writer for a recording medium, and is built in or externally attached to the information processing device 900. The drive 915 reads out information recorded in the attached removable recording medium 925 and outputs it to the RAM 905. The drive 915 can also write information to the attached removable recording medium 925. The removable recording medium 925 is, for example, a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory. Specifically, the removable recording medium 925 may be a CD medium, a DVD medium, a Blu-ray (registered trademark) medium, a CompactFlash (registered trademark) (CF), a flash memory, an SD memory card (Secure Digital memory card), or the like. The removable recording medium 925 may also be, for example, an IC card (Integrated Circuit card) equipped with a non-contact IC chip, or an electronic device, or the like.

[0117] The connection port 917 is a port for directly connecting a device to the information processing device 900. The connection port 917 is, for example, a Universal Serial Bus (USB) port, an IEEE 1394 port, a Small Computer System Interface (SCSI) port, an RS-232C port, etc. The information processing device 900 can directly obtain various data from an external device 927 connected to the connection port 917 and provide various data to the external device 927.

[0118] The communication device 919 is, for example, a communication interface configured with a communication device for connecting to a communication network 929. The communication device 919 is, for example, a wireless LAN (Local Area Network), Bluetooth (registered trademark), or an antenna for mobile communication. The communication device 919 may also be a wired LAN, a router for optical communication, a router for ADSL (Asymmetric Digital Subscriber Line), or a modem for various communications. The communication device 919 can transmit and receive signals, for example, between the Internet and other communication devices in accordance with a predetermined protocol, for example, TCP / IP. The communication network 929 connected to the communication device 919 is configured with a network connected by wire or wirelessly. For example, the communication network 929 is a wide area network (WAN) such as the Internet, a LAN, a mobile communication network, infrared communication, radio wave communication, or satellite communication.

[0119] The above shows an example of the hardware configuration of the information processing device 900. Each of the above-mentioned components may be configured using general-purpose members, or may be configured using hardware specialized for the function of each component. The hardware configuration of the information processing device 900 can be appropriately changed depending on the technical level at the time of implementing this embodiment.

[0120] Although the preferred embodiment of the present invention has been described in detail above with reference to the accompanying drawings, the present invention is not limited to such an example. It is clear that a person having ordinary knowledge in the technical field to which the present invention pertains can conceive of various modified or altered examples within the scope of the technical ideas described in the claims, and it is understood that these also naturally belong to the technical scope of the present invention.

[0121] For example, in the above embodiment, an example of simplifying a graph representing a logistics network has been described, but the present invention is not limited to such an example. For example, the logistics drawing processing method of the present invention can simplify a graph representing a network with a passing order and flow, such as a logistics network. Specifically, it is also possible to simplify a graph representing a network of roads, trains, airplanes, etc., or a business network representing the administrative process in a company, hospital, etc. [Explanation of symbols]

[0122] 100 Logistics drawing device 110 Data Acquisition Section 120 Graph structure acquisition unit 130 Node Aggregation Unit 140 Graph output section 200 Data storage unit 210 Performance Data Table 300 Output Device 510 Cursor 520 Speech bubble 900 Information processing device 907 Bus 913 Storage Device 915 Drive 917 Connection Port 919 Communication Equipment 921 Input Device 923 Output Device 925 Removable Recording Media 927 External equipment 929 Communication Network Na Aggregation Node G Graph (before simplification) Ga Simplified Graph P0 reference position R Performance data S graph structure data Sm adjacency matrix (graph structure data) Sa Aggregation graph structure data T P Probability Distribution Tables T Q Inflow and Outflow Table T KL KL Divergence Table T JS JS Divergence Table

Claims

1. A data acquisition unit that acquires performance data in a logistics network; a node aggregating unit that generates an aggregated graph structure by aggregating nodes in a graph structure that represents the performance data and indicates connections between processes that constitute the logistics network, based on a similarity between two nodes in the graph structure; a graph output unit that outputs the aggregate graph structure; A logistics drawing device comprising:

2. The node aggregation unit includes: The physical distance between the two nodes, The time distance between two nodes, The distance of the probability distribution of the inflow from the parent node to the child node, or Distance of probability distribution of outflow to child node from parent node, The logistics drawing device according to claim 1 , wherein the similarity is calculated based on at least one of the following:

3. 2. The logistics drawing device according to claim 1, wherein the node aggregating unit determines the similarity between two nodes based on whether or not their logistics positions in the graph structure are the same.

4. 4. The logistics drawing device according to claim 3, wherein the node aggregating unit aggregates nodes representing the same function among the nodes determined to be in the same location in terms of logistics.

5. The logistics drawing device according to claim 1 , wherein the node aggregating unit aggregates nodes of performance data of the same product type.

6. The graph output unit The logistics drawing device according to any one of claims 1 to 5, wherein the graph represented by the aggregated graph structure is displayed on an output device so that the aggregated contents of the nodes aggregated into the node after aggregation by the node aggregation unit can be identified.

7. The logistics drawing device according to claim 6 , wherein the graph output unit displays an aggregated content of the aggregated node in the aggregated node of the graph displayed on the output device.

8. A logistics drawing device that displays an aggregated graph structure that aggregates at least a portion of a plurality of nodes that constitute a graph representing actual logistics data in a logistics network, in such a manner that the aggregated contents of the nodes aggregated in the aggregated node of the aggregated graph structure can be identified.

9. A data acquisition step of acquiring performance data in a logistics network; a node aggregation step of generating an aggregated graph structure by aggregating nodes in a graph structure that represents the performance data and indicates connections between processes that constitute the logistics network, based on a similarity between two nodes in the graph structure; a graph output step of outputting the aggregate graph structure; A logistics drawing processing method comprising:

10. Computer, A data acquisition unit that acquires performance data in a logistics network; a node aggregating unit that generates an aggregated graph structure by aggregating nodes in a graph structure that represents the performance data and indicates connections between processes that constitute the logistics network, based on a similarity between two nodes in the graph structure; a graph output unit that outputs the aggregate graph structure; A program for functioning as a logistics drawing device.

Images