Learning model creation device, plan creation device, learning model creation method, plan creation method, and program

Abstract

To enable creating a highly accurate production plan and logistic plan with as little human intervention as possible.SOLUTION: A learning model creation device 110 acquires training data including learning model creation product data having a value of a manufacturing specification m for each of a plurality of products included in a production plan and a production order of the products in the production plan, and pass / fail determination data indicating pass / fail of the production plan. The learning model creation device 110 creates a learning model creation plan representation image 230 which is an image in which, based on the learning model creation product data, each of the plurality of manufacturing specifications m is associated with columns, the production order of the plurality of products is associated with rows, and the values of the manufacturing specifications m are associated with pixel values. The learning model creation device 110 performs supervised learning using the learning model creation plan representation image 230 for creating a learning model that inputs the plan representation image and outputs a degree of certainty regarding the pass / fail of the production plan.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a learning model creation device, a plan creation device, a learning model creation method, a plan creation method, and a program, and is particularly suitable for use in creating production plans or logistics plans. [Background technology]

[0002] The constraints that must be considered when creating production and logistics plans can be numerous and complex, or they may be tacit knowledge based on the plan creator's experience and intuition. In such cases, it is difficult to define the precise constraints necessary for computer-based plan creation, and the resulting plans are unlikely to satisfy the user. To address this problem, Patent Document 1 discloses a technique that learns the production instruction rules of a plan creator and selects the same production instruction rules for production plan data similar to the production plan data to which the learned production instruction rules were applied. Furthermore, the technique described in Patent Document 2 creates multiple allocation candidates based on certain rules and rules of thumb when selecting production lots to allocate to a production line. Then, the evaluation results based on rules of thumb for each of the multiple allocation candidates and the evaluation results predicted by the plan simulation are aggregated and the allocation candidates are ranked.

[0003] However, these technologies assume that the constraints to be handled are clearly defined in advance and cannot handle constraints based on so-called tacit knowledge. One technology that can handle constraints based on tacit knowledge is the technology described in Patent Document 3. The technology described in Patent Document 3 compares a provisional plan created with known constraints with the actual plan. Correlation analysis is performed between the actual plan (output) data, which is determined to have a similarity to the provisional plan below a predetermined standard (i.e., dissimilar), and the input data for that actual plan (data used to create the actual plan). Then, constraints based on tacit knowledge are generated based on the relationship between input and output that is determined to have a correlation of a predetermined strength or higher. [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 7-21263 [Patent Document 2] Japanese Patent Publication No. 2012-194712 [Patent Document 3] Patent No. 6467264 [Overview of the Initiative] [Problems that the invention aims to solve]

[0005] However, the technology described in Patent Document 3 requires a complex, multi-stage procedure that inevitably involves human intervention, including the creation of a provisional plan, comparison of the provisional plan with the actual plan, and correlation analysis of the provisional plan with a dissimilar actual plan.

[0006] This invention has been made in view of the above-mentioned problems, and aims to enable the creation of highly accurate production plans and logistics plans with as little human intervention as possible. [Means for solving the problem]

[0007] The present invention relates to a learning model creation apparatus that creates a learning model used to create a plan, which is a production plan or a logistics plan, and comprises: a training data acquisition means for acquiring training data which includes training data which is product data for each of a plurality of products included in the plan, having values ​​for a plurality of items including the specifications of the product and the processing order of the product in the plan, and good / bad judgment data indicating the goodness or badness of the plan; a learning model creation image creation means for creating a plan representation image for learning model creation, which is an image in which each of the plurality of items is associated with either a row or a column, the processing order of the plurality of products is associated with the other row or column, and the values ​​of the items are associated with pixel values, based on the training model creation product data included in the training data; and a learning model creation means for creating a learning model which takes the plan representation image as input and outputs a confidence level regarding the goodness or badness of the plan by performing supervised learning using the learning model creation plan representation image created by the learning model creation image creation means. The planning device of the present invention is a planning device that creates a plan using the learning model created by the learning model creation device, and comprises: a product data acquisition means for planning, which is product data different from the product data for creating the learning model; a plan image creation means for creating a plan representation image, which is a plan representation image, based on the product data for planning; a confidence calculation means for inputting the plan representation image created by the plan image creation means into the learning model and calculating a confidence level regarding the quality of the plan; and a learning model that uses the confidence level calculated by the confidence calculation means to perform the learning The confidence calculation means includes a plan representation image modification means for modifying the plan representation image for plan creation input into the model, and a plan creation means for creating the plan based on the modified plan representation image for plan creation when the plan representation image for plan creation modified by the plan representation image modification means satisfies predetermined calculation completion conditions for calculations to modify the plan representation image for plan creation. If the plan representation image for plan creation modified by the plan representation image modification means does not satisfy the calculation completion conditions, the confidence calculation means inputs the plan representation image for plan creation modified by the plan representation image modification means into the learning model and recalculates the confidence.

[0008] The present invention relates to a method for creating a learning model used to create a plan, such as a production plan or a logistics plan. The means of acquiring training data is A training data acquisition step to acquire training data which includes product data for each of the multiple products included in the plan, the product data having values ​​for multiple items including the specifications of the product and the processing order of the product in the plan, and training data which includes good / bad judgment data indicating the goodness or badness of the plan. The means for creating images for creating learning models is,A learning model creation image creation step, which creates a planned representation image as a learning model creation image, based on the product data for creating the learning model included in the training data, by associating each of the multiple items with either a row or a column, and the processing order of the multiple products with the other row or column, and associating the values ​​of the items with pixel values. The means for creating a learning model is, The system includes a learning model creation step, which involves performing supervised learning using the learning model creation plan representation image created in the learning model creation step, thereby creating a learning model that takes the plan representation image as input and outputs a confidence level regarding the quality of the plan. The plan creation method of the present invention is The aforementioned A planning method for creating a plan using the learning model created by the learning model creation method, The means for acquiring product data for planning is, A planning product data acquisition step involves acquiring planning product data, which is different from the product data used for creating the learning model, The means for creating images for planning is, A planning image creation step, which creates a planning representation image for planning, based on the product data for planning, The confidence level calculation method is, A confidence calculation step involves inputting the plan representation image created in the plan creation step into the learning model and calculating the confidence level regarding the quality of the plan. The means for modifying the planned representation image is, A plan representation image modification step is performed to modify the plan representation image for plan creation input to the learning model based on the confidence level calculated in the confidence level calculation step, The means of creating a plan are The confidence calculation step includes, if the plan creation plan representation image corrected by the plan representation image correction step satisfies predetermined calculation completion conditions for calculations to correct the plan creation plan representation image, then creating the plan based on the corrected plan creation plan representation image, wherein if the plan creation plan representation image corrected by the plan representation image correction step does not satisfy the calculation completion conditions, the confidence calculation step inputs the plan creation plan representation image corrected by the plan representation image correction step into the learning model and recalculates the confidence.

[0009] A first example of the program of the present invention involves making a computer function as one of the means of the learning model creation apparatus. A second example of the program of the present invention involves making a computer function as one of the means of the planning apparatus. [Effects of the Invention]

[0010] According to the present invention, highly accurate production plans and logistics plans can be created with minimal human intervention. [Brief explanation of the drawing]

[0011] [Figure 1] This figure shows an example of the functional configuration of a learning model creation device and a plan creation device. [Figure 2] This figure shows an example of source images for designing a learning model and data for determining whether the model is good or bad. [Figure 3] This is a conceptual diagram illustrating an example of a CNN. [Figure 4] This diagram illustrates an example of a convolution operation. [Figure 5] This diagram provides a detailed explanation of an example of a convolution operation. [Figure 6] This diagram illustrates an example of processing in the convolutional layer and pooling layer. [Figure 7] This figure shows an example of an overview of how to create a plan. [Figure 8] This is a flowchart illustrating one example of a method for creating a learning model. [Figure 9] This is a flowchart illustrating one example of a plan creation method. [Modes for carrying out the invention]

[0012] One embodiment of the present invention will be described below with reference to the drawings. Figure 1 shows an example of the functional configuration of the learning model creation device 110 and the planning device 120. The hardware of the learning model creation device 110 and the planning device 120 is realized by using an information processing device equipped with, for example, a processor, main memory, auxiliary memory, and various interfaces. In this embodiment, the case in which the learning model creation device 110 and the planning device 120 are separate devices is illustrated. However, the learning model creation device 110 and the planning device 120 may be realized in a single device. Below, an example of the learning model creation device 110 and the planning device 120 of this embodiment will be described in the order of the learning model creation device 110 and the planning device 120.

[0013] [Learning model creation device 110] The learning model creation device 110 creates a learning model used to create a plan, such as a production plan or a logistics plan. In this embodiment, an example is given of the learning model creation device 110 creating a learning model used to create a production plan when producing a product. An example of a production plan is a hot rolling schedule (a schedule that includes the rolling order of slabs (steel materials) to be hot-rolled). However, the production plan is not limited to a hot rolling schedule as long as it includes the production order of the products. The learning model creation device 110 may also create a learning model used to create a logistics plan instead of a production plan. When creating a learning model used to create a logistics plan, in the following description, for example, the production order can be replaced with the transport order. The learning model can be any learning model (machine learning model) created by supervised learning, but in this embodiment, an example is given of the learning model being a CNN (Convolutional Neural Network) (however, the CNN algorithm itself is modified from a known CNN algorithm so that the calculations described later are performed, and is different from a known CNN algorithm).

[0014] The learning model creation device 110 includes a training data acquisition unit 111, a learning model creation image creation unit 112, a learning model creation unit 113, and a storage unit 114.

[0015] <Teacher data acquisition unit 111> The training data acquisition unit 111 acquires training data that includes product data for creating a learning model and good / bad judgment data corresponding to the product data for creating the learning model. The product data for creating the learning model is product data for each of the multiple products included in the plan, and the product data includes values ​​for multiple items, including product specifications, and the processing order of the products in the plan.

[0016] In this embodiment, we illustrate a case where multiple items in the product data for creating a learning model are the product's size, material, quality, and delivery date. The product's size, material, quality, and delivery date are the product's specifications. Note that the multiple items may also include items other than product specifications. For example, the multiple items may include the product's production conditions (e.g., the processes performed when producing the product, the equipment used when producing the product, and the materials used when producing the product). Furthermore, in this embodiment, we illustrate a case where it is a production plan, so the processing order is the production order.

[0017] Thus, in this embodiment, the product data includes the values ​​of multiple items, including specifications, for each product included in a particular production plan, and the production order of that product in that production plan. The planning device 120 (product data acquisition unit 121 for planning) described later also acquires product data. In the following description, in order to distinguish it from the product data acquired by the planning device 120, the product data acquired by the training data acquisition unit 111 will be referred to as product data for learning model creation, as needed. Also, in the following description, the multiple items, including specifications, will be referred to as manufacturing specifications, as needed.

[0018] The quality judgment data indicates whether a plan is good or bad. In this embodiment, we will illustrate a case where "1" is set as the quality judgment data for a good plan and "0" is set for a bad plan. A good plan is a plan that does not require modification, for example, a plan that was used in actual production. Alternatively, a good plan may be a plan that was used when products were produced according to plan. A bad plan is a plan that requires modification (the opposite of a good plan), for example, a plan that was drafted but not used in actual production, or a plan that was used when products were not produced according to plan.

[0019] In this embodiment, we illustrate a case where the good / bad judgment data includes a confidence level indicating the likelihood of a good plan and a confidence level indicating the likelihood of a bad plan. Furthermore, in this embodiment, we illustrate a case where the confidence level takes a value between 0 and 1. A confidence level of "1" indicates the highest confidence level, and a confidence level of "0" indicates the lowest confidence level. In the training data, it is identified whether the product data for creating the learning model is based on a good plan or a bad plan. Therefore, if the product data for creating the learning model is based on a good plan, the good / bad judgment data corresponding to that product data will include data indicating a confidence level of "1" for a good plan and a confidence level of "0" for a bad plan. On the other hand, if the product data for creating the learning model is based on a bad plan, the good / bad judgment data corresponding to that product data will include data indicating a confidence level of "0" for a good plan and a confidence level of "1" for a bad plan. In other words, the pass / fail judgment data corresponding to the product data used to create the learning model includes data with a value of "0" and data with a value of "1".

[0020] The training data acquisition unit 111 acquires multiple training data sets as the training data described above. Preferably, the training data acquired by the training data acquisition unit 111 includes multiple training data sets containing product data for creating a learning model with a good design, and multiple training data sets containing product data for creating a learning model with a poor design. Furthermore, from the viewpoint of training the learning model (CNN in this embodiment) with high accuracy, it is preferable that the number of training data sets acquired by the training data acquisition unit 111 that include product data for creating a learning model with a good design, and the number of training data sets that include product data for creating a learning model with a poor design, are both large. However, if the number of training data sets acquired by the training data acquisition unit 111 is too large, the computational load will increase. The number of training data sets acquired by the training data acquisition unit 111 is determined appropriately from this perspective.

[0021] As a method for acquiring training data, at least one of the following is adopted: receiving training data transmitted from an external device, reading training data stored on a portable storage medium, and inputting training data through input operations of the user interface of the learning model creation device 110 by an operator.

[0022] <Image creation unit 112 for creating learning models> The learning model creation image creation unit 112 creates a learning model planning representation image based on the learning model creation product data included in the training data acquired by the training data acquisition unit 111.

[0023] The plan representation image is an image in which each of the multiple items included in the product data is associated with either a row or a column, the processing order of the multiple products included in the product data is associated with the other row or column, and the value of each item is associated with a pixel value. The plan creation device 120 (plan creation image creation unit 122), which will be described later, also creates plan representation images. In the following description, in order to distinguish it from the plan representation image created by the plan creation device 120, the plan representation image created by the learning model creation image creation unit 112 will be referred to as the learning model creation plan representation image as needed.

[0024] In this embodiment, we illustrate a case where each of the multiple manufacturing specifications (items) included in the product data for creating the learning model is associated with a column, and the production order of the multiple products included in the product data for creating the learning model is associated with a row.

[0025] Figure 2 shows an example of a plan representation source image 210 and pass / fail judgment data 220 for creating a learning model. The plan representation image is created by treating the product data as an image and converting the values ​​of each manufacturing specification (item) included in the product data into pixel values. The plan representation source image refers to the image that serves as the basis for the plan representation image. The plan creation device 120 (plan creation image creation unit 122), which will be described later, also creates a plan representation source image. To distinguish it from the plan representation source image created by the plan creation device 120, the plan representation source image created by the learning model creation image creation unit 112 will be referred to as the learning model creation plan representation source image as needed.

[0026] In this embodiment, multiple categories are set for each manufacturing specification (size, material, quality, delivery date) included in the product data for creating the learning model. Then, the value of each manufacturing specification included in the product data for creating the learning model is classified into one of the multiple categories set for that manufacturing specification. In the example shown in Figure 2, six categories from 1 to 6 are set for size. Here, a larger number indicating the size category represents a larger size. Another example is set of six categories from A to F for material. Here, categories with letters closer together alphabetically represent similar materials. Another example is set of five categories from 1 to 5 for quality. Here, a larger number indicating the quality category represents a higher quality product. Another example is set of five categories from 1 to 5 for delivery date. Here, a larger number indicating the delivery date category represents a later delivery date (more time available).

[0027] Figure 2 illustrates a case where the production plan includes 12 products. The learning model creation unit 112 identifies which of the aforementioned categories each of the 12 products' manufacturing specifications (size, material, quality, delivery date) belongs to. As described above, in this embodiment, each of the multiple manufacturing specifications included in the learning model creation product data is associated with a column, and the production order of the multiple products included in the learning model creation product data is associated with a row. Therefore, in Figure 2, the manufacturing specifications (size, material, quality, delivery date) of the 12 products included in the production plan are associated with the columns of the learning model creation plan representation source image 210. Furthermore, the production order of the 12 products included in the production plan is associated with a row.

[0028] In the example shown in Figure 2, the production order of the products is assigned to each pixel in ascending order, starting from the top row of the learning model creation plan representation image 210. Therefore, for example, the product that will be produced earliest among the 12 products included in the production plan is assigned to the pixel in the first row of the learning model creation plan representation image 210. In this embodiment, v is a variable that identifies the row of the learning model creation plan representation image 210, and the learning model creation plan representation image 210 is from row v1 (=1st row) to v max Assume that there are up to row 12.

[0029] Also, in the example shown in FIG. 2, six categories of size, six categories of material, five categories of quality, and five categories of delivery date are assigned to each pixel in order from the left of the columns of the planning expression source image 210 for learning model creation. Therefore, for example, the pixel value of the pixel corresponding to the category to which the value of each manufacturing specification (size, material, quality, delivery date) of the product belongs is made different from the pixel values of the pixels corresponding to other categories of the same manufacturing specification. For example, a bit pattern is used in which the pixel value of the pixel corresponding to the category to which the value of each manufacturing specification (size, material, quality, delivery date) of the product belongs is set to "1", and the pixel values of other categories of the same manufacturing specification are set to "0". In FIG. 2, pixels with a pixel value of "1" are shown in black, and pixels with a pixel value of "0" are shown in white (this shall be the same in each figure after FIG. 3. This corresponds to one-hot encoding). For example, in FIG. 2, among the 12 products included in the production plan, the categories of the product, material, quality, and delivery date that are produced earliest are 5, C, 4, and 5, respectively, as shown in the first row.

[0030] In the present embodiment, a variable for identifying the columns of the planning expression source image 210 for learning model creation is denoted as h, and a variable for identifying the manufacturing specification is denoted as m. However, the variable m shall not be assigned to the delivery date. Also, the values of m for size, material, and quality are set to 1, 2, and 3, respectively. And as shown in FIG. 2, in the planning expression source image 210 for learning model creation, the region corresponding to the manufacturing specification m exists from column h1 m to column h max m shall be assumed. Therefore, the region corresponding to the size (m = 1) exists from column h1 1 (= the first column) to column h max 1 (= the sixth column). Also, the region corresponding to the material (m = 2) exists from column h1 2 (= the seventh column) to column h max 2 (= the twelfth column). The region corresponding to the quality (m = 3) exists from column h1 3 (= the thirteenth column) to column h max 3There are up to 17 columns. For example, h1 1 This corresponds to the symbol in Figure 2 where a 1 is placed above the 1 in h1. This method of notation is h1 1 The same applies to other symbols, and they shall also be represented in the same way in each figure and each formula.

[0031] In this embodiment, the portion of the source image 210 for creating the learning model to which the variable m has been added in this manner is designated as the learning model creation plan representation image 230 (let the number of m values ​​be M). In this embodiment, the case in which the variable m is not added to the delivery date is shown as an example, but the variable m may also be added to the delivery date. In this case, the source image 210 for creating the learning model to create

[0032] As mentioned above, in this embodiment, the good / bad judgment data 220 includes a confidence level for a good plan and a confidence level for a bad plan. An example is given where "1" indicates the highest confidence level and "0" indicates the lowest confidence level. In the training data, it is identified whether the product data for creating the learning model is a good plan or a bad plan. In Figure 2, a confidence level of "1" for a good plan and a confidence level of "0" for a bad plan is indicated as "Good: 1 Bad: 0". In other words, the product data for creating the learning model from which the plan representation image 230 shown in Figure 2 was created is a good plan.

[0033] In the following explanation, the confidence level Y1 that the plan represented by the plan representation image input to the input layer 310 is a good plan will be referred to as the confidence level Y1 for a good plan, as needed. Similarly, the confidence level Y2 that the plan represented by the plan representation image input to the input layer 310 is a poor plan will be referred to as the confidence level Y2 for a bad plan, as needed.

[0034] In this embodiment, the learning model creation image creation unit 112 creates the learning model creation plan representation source image 210 and the learning model creation plan representation image 230 as described above. Note that the learning model creation plan representation source image 210 and the learning model creation plan representation image 230 are not limited to those shown in Figure 2. For example, when the pixel values ​​of the learning model creation plan representation source image 210 and the learning model creation plan representation image 230 are represented by 8 bits, the values ​​of each of the multiple manufacturing specifications m may be classified into one of the 256 grayscale values.

[0035] Furthermore, in this embodiment, an example was given in which each of the multiple manufacturing specifications included in the product data for creating the learning model is associated with a column (horizontal direction), and the production order of the multiple products included in the product data for creating the learning model is associated with a row (vertical direction). However, it is also possible to associate each of the multiple manufacturing specifications included in the product data for creating the learning model with a row, and the production order of the multiple products included in the product data for creating the learning model with a column.

[0036] <Learning model creation unit 113, memory unit 114> The learning model creation unit 113 performs supervised learning using the learning model creation plan representation image 230 to create a learning model that takes the plan representation image as input and outputs a confidence level regarding the quality of the plan, and stores it in the storage unit 114. As mentioned above, this embodiment exemplifies the case where the learning model is a CNN.

[0037] Figure 3 is a conceptual diagram illustrating an example of a CNN. In the example shown in Figure 2, the CNN comprises an input layer 310, a convolutional layer 320, a pooling layer 330, and an output layer 340. This embodiment illustrates a case where the CNN has the configuration shown in Figure 3. However, the configuration of the CNN is not limited to the configuration shown in Figure 3, and other known CNN configurations may be adopted. For example, in Figure 3, for the sake of simplicity of explanation and notation, the case where there is one convolutional layer 320 and one pooling layer 330 is illustrated, but the CNN may have multiple sets of convolutional layers 320 and pooling layers 330. Also, in this embodiment, the case where the neurons of the output layer 340 and each component of the pooling layer 330 are fully connected is illustrated, illustrating a case where the CNN does not have a fully connected layer, but the CNN may have a fully connected layer.

[0038] In a CNN, an image is input to the input layer 310. In this embodiment, we illustrate the case where a design representation image 230 for creating a learning model is input to the input layer 310. In the convolutional layer 320, a convolution operation is performed on the image using kernels (filters). Figure 4 is a diagram illustrating an example of the overview of the convolution operation. In the example shown in Figure 4, the kernels used in the convolution operation are set individually for one of several items (manufacturing specifications m) assigned to the columns of the learning model planning representation image 230. Specifically, Figure 4 illustrates the case where kernel 411 is set for size, kernel 412 is set for material, and kernel 413 is set for quality.

[0039] Here, let M be the number of items (manufacturing specifications m) included in the learning model planning representation image 230. In the example shown in Figure 2, M = 3. The same kernel may be set for items (manufacturing specifications m) that are 2 or more but less than M in the learning model planning representation image 230. In the example shown in Figure 2, for example, the same kernel may be set for size and material, the same kernel may be set for size and quality, the same kernel may be set for material and quality, or the same kernel may be set for size, material and quality.

[0040] In the learning model creation planning representation image 230, when setting the same kernel for items (manufacturing specification m) between 2 and M, it is possible to set the number of kernels applied to each item (manufacturing specification m) to one, or to set multiple kernels to a single item (manufacturing specification m). In the example shown in Figure 2, for example, one kernel may be set for size and material and another kernel may be set for quality, or one kernel may be set for size and material and another kernel may be set for quality and material. As described above, when setting the same kernel for items (manufacturing specification m) between 2 and M in the learning model creation planning representation image 230, in the following explanation, items between 2 and M (for example, size and material in the example shown in Figure 2) can be considered as a single item (manufacturing specification m).

[0041] Furthermore, Figure 4 illustrates a case where the size of kernels 411 to 413 set for a single item (manufacturing specification m) in the longitudinal direction (i.e., row direction (horizontal direction)) of the rows corresponding to the processing order (production order) of multiple products in the learning model creation planning representation image 230 is the same as the size of the area corresponding to that single item (manufacturing specification m) in the learning model creation planning representation image 230 in the same direction. That is, as illustrated with reference to Figure 2, in the learning model creation planning representation image 230, the production order of multiple products is associated with rows, and multiple items (manufacturing specification m) are associated with columns. Therefore, the row direction (horizontal direction) size of kernel 411 set for size is the same as the row direction size of the area corresponding to size in the learning model creation planning representation image 230. Similarly, the row direction (horizontal direction) sizes of kernels 412 and 413 set for material and quality are the same as the row direction sizes of the areas corresponding to material and quality in the learning model creation planning representation image 230, respectively.

[0042] Furthermore, Figure 4 illustrates the case where the size of the columns corresponding to multiple items (manufacturing specifications m) in the learning model creation planning representation image 230 for kernels 411 to 413 is adjusted to a size corresponding to the items (manufacturing specifications m) for which kernels 411 to 413 are set. For example, it may be preferable that the material of the product produced after a product made of a certain material is a specific material (for example, in Figure 4, the product produced after a product made of material F is made of material C). In such cases, a production plan that includes many relationships where the material of the product produced after a product made of a certain material is a good production plan. Therefore, to make it easier to extract such features, the column-direction (vertical direction) size of the kernel 412 set for the material is reduced.

[0043] Furthermore, for example, if high-quality and low-quality products are distributed in the production order, it may become necessary to process the low-quality products in the same way as the high-quality products, potentially increasing the production cost of the low-quality products. In such cases, a production plan in which products of the same quality are produced consecutively is a good production plan. Therefore, to make it easier to extract such characteristics, the size of the column direction (vertical direction) of kernel 413 set for the material is increased.

[0044] Furthermore, for example, if the size of adjacent products changes abruptly in the production order, it may become difficult to produce the products. In other words, in such cases, a production plan that ensures the size changes gradually is a good production plan. Therefore, in order to extract such characteristics, the column-direction (vertical) size of kernel 411 set for the size is set to a medium size.

[0045] Also, in FIG. 4, a case is illustrated where kernels 411 to 413 set for each item (each manufacturing specification m) are slid only within the region corresponding to the item (manufacturing specification m) in the plan expression image 230 for learning model creation. Specifically, in FIG. 4, kernel 411 is slid only within the region corresponding to the size in the plan expression image 230 for learning model creation. Similarly, kernels 412 and 413 are slid only within the regions corresponding to the material and quality, respectively, in the plan expression image 230 for learning model creation. More specifically, in FIG. 4, a case is illustrated where the direction in which kernels 411 to 413 are slid is only the direction of the column corresponding to a plurality of items (manufacturing specifications m) (the column direction (vertical direction)) (refer to the white arrow lines shown below kernels 411 to 413 in FIG. 4).

[0046] However, it is not always necessary to set kernels 411 to 413 as shown in FIG. 4. For example, the size of the kernel in the row direction set for one item (manufacturing specification m) may be smaller than the size in the row direction of the region corresponding to the one item (manufacturing specification m) in the plan expression image 230 for learning model creation. In such a case, the kernel set for the one item (manufacturing specification m) is slid, for example, in the row direction (horizontal direction) and the column direction (vertical direction) within the region corresponding to the item (manufacturing specification m) in the plan expression image 230 for learning model creation.

[0047] FIG. 5 is a diagram for explaining an example of a convolution operation in detail. In the present embodiment, in the plan expression image 230 for learning model creation, the pixel value of the v-th row and h-th column of the product specification m is represented as x v,h m Here, when representing the pixel value x v,h m it is assumed that a numerical value is set for h for each region to which kernels 411 to 413 are applied in the plan expression image 230 for learning model creation. In the present embodiment, kernel 411 is set in the region corresponding to the size in the plan expression image 230 for learning model creation. Therefore, for example, the pixel value x v,h mWhen representing this, Figure 2 shows the h1 region corresponding to the size (m=1). 1 The value of h, which indicates a column, will be "1", h max 1 The value of h, which indicates the column, will be "6". Also, the pixel value x v,h m When expressing this, the h1 region corresponding to the material (m=2) 2 The value of h, which indicates a column, will be "1", h max 2 The value of h, which indicates the column, will be "6". Also, the pixel value x v,h m When expressing this, the h1 region corresponding to quality (m=3) 3 The value of h, which indicates a column, will be "1", h max 3 The value of h, which indicates the column, will be "5". In this way, x shown in Figure 5 v,h m This is defined as the pixel value of the planning representation image 230 for creating the learning model. Also, the weights of kernels 411 to 413 are set w l,s m_fc This is how it is written. Here, m_f c This is a symbol that identifies the convolutional layer 320 for manufacturing specification m. In the examples shown in Figures 3 to 5, one convolutional layer 320 is created for each manufacturing specification m, so f c = 1. When multiple convolutional layers 320 are created for one manufacturing specification m, f c f can take values ​​greater than or equal to 2 (Note that in the formula described later, f c The number F c (This is how it is written.) l is the position of an element in the feature map that constitutes the convolutional layer 320, and is a number (row number) that identifies the position of the element in the direction corresponding to the column direction (vertical direction) of the image input to the input layer 310, and s is the position of an element in the feature map that constitutes the convolutional layer 320, and is a number (column number) that identifies the position of the element in the direction corresponding to the row direction (horizontal direction) of the image input to the input layer 310. Also, the number of elements in the column direction (vertical direction) of kernels 411 to 413 is k r m This is expressed as k, where k is the number of elements in the row (horizontal direction) of kernels 411-413. c m This is how it is written.

[0048] Then, in the convolution operation in the convolutional layer 320, the calculation of the following formula (1) is performed, and u l,s m_fc is calculated.

[0049]

Equation

[0050] FIG. 6 is a diagram for explaining an example of processing in the convolutional layer 320 and the pooling layer 330. In FIG. 6, for the sake of notation, only the processing of the region of size (m = 1) in the planned representation image 230 for creating a learning model will be described. Also, in FIG. 6, for the planned representation image 230 for creating a learning model shown in FIG. 2, the processing when the kernels 411 to 413 are set as shown in FIGS. 4 and 5 is exemplified. When the kernels 411 to 413 are set as shown in FIGS. 4 and 5, k c m =h max m Therefore, h max m -k c m +1 = 1. In this case, as shown in FIG. 6, the image of the region corresponding to one manufacturing specification m in the planned representation image 230 for creating a learning model becomes a single-column image by the convolution operation in the convolutional layer 320. In FIG. 3, as a result of applying the kernel 411 to the region of size (m = 1) in the planned representation image 230 for creating a learning model, the calculated value u 1,1 1_1 of the convolution operation is shown in gray to be stored in the element 321 of the feature map. Similarly, as a result of applying the kernel 412 to the region of material (m = 2) in the planned representation image 230 for creating a learning model, the calculated value u 1,1 2_1 of the convolution operation is shown in gray to be stored in the element 322 of the feature map, and as a result of applying the kernel 413 to the region of quality (m = 3) in the planned representation image 230 for creating a learning model, the calculated value u 1,1 3_1However, it is indicated in gray that it will be stored in element 323 of the feature map.

[0051] Also, in equation (1), d l+q-1 This is shown in Figure 2, d v This corresponds to the example shown in Figure 2, where v takes values ​​from 1 to 12, so in Figure 5, d v d1~d 12 This is how it is written. d v (d1~d 12 ) is the value of the delivery date category shown in Figure 2. As mentioned above, the larger the value of the delivery date category, the further away the delivery date is (in other words, the smaller the value of the delivery date category, the closer the delivery date is). In Figure 2, for example, the delivery date category value of the product with production order "6" is "1", while the delivery date category value of the products with production order "1" to "5" is "4" or "5". That is, in the example shown in Figure 2, the planning representation image 230 for creating the learning model is created based on a production plan that includes products that are produced before products with later delivery dates and products with closer delivery dates. Therefore, in this embodiment, products that are produced before products with closer delivery dates are considered to be of high importance from the perspective of manufacturing specifications m, and the weight of the pixel value corresponding to such products is increased in the planning representation image 230 for creating the learning model. Thus, in this embodiment, as shown in equation (1), the weight w set in kernels 411 to 413 p,q m against d l+q-1 By multiplying by x, the pixel values ​​of the planning representation image 230 for creating the learning model are obtained. v,h m (x in equation (1)) l+q-1,s+r-1 m The weight for ) is set to a value corresponding to the delivery date of the product that corresponds to that pixel value.

[0052] In the convolution operation, for example, a bias (threshold) is added to the right-hand side of equation (1), but the bias (threshold) itself is implemented using well-known techniques for CNNs. Therefore, the notation and detailed explanation of the bias (threshold) are omitted here.

[0053] Then, in the convolutional layer 320, the calculated value u of the convolutional operation obtained as described above l,s m_fc is given to u of the activation function σ c (u) as shown in the following equation (2), and the calculated value c of the convolutional layer 320 l,s m_fc is calculated.

[0054]

Equation

[0055] In the example shown in FIG. 6, the calculated values u of the convolutional operation 1,1 1_1 , ···, u 8,1 1_1 respectively indicate that the calculated values c of the convolutional layer 320 1,1 1_1 , ···, c 8,1 1_1 are calculated. In FIG. 3, for the elements 321, 322, 323 of the feature map, c 1,1 1_1 , c 1,1 2_1 , and c 1,1 3_1 obtained by giving u c (u) of the activation function σ 1,1 1_1 , c 1,1 2_1 , and c 1,1 3_1 are stored. Similarly, for the elements 324, 325, 326 of the feature map shown in FIG. 3, c 2,1 1_1 , c 2,1 2_1 , and c 2,1 3_1 obtained by giving u c (u) of the activation function σ 2,1 1_1 , c 2,1 2_1 , and c 2,1 3_1 are stored. Note that the activation function σ cFor (u), you can use a well-known function used in CNNs, such as the sigmoid function.

[0056] The processing in the convolutional layer 320 is performed as described above. The processing in the pooling layer 330 and the output layer 340 is implemented using the same processing as in known CNNs. Figure 3 shows the calculated value c of the convolutional layer 320 for a region of size (m=1) of the design representation image 230 for creating the learning model. l,s 1_fc The two elements 321 and 324 of the feature map that store the data are compressed (downsampled) and stored in the output table (compressed feature map elements) 331 of the pooling layer 330, as shown in gray (see also feature map elements 321, 324, and 331 in Figure 6). Similarly, the calculated values ​​c of the convolutional layer 320 for the material (m=2) region of the planned representation image 230 for creating the learning model are shown. l,s 2_fc The two feature maps 322 and 325 that store the data are compressed and stored in the output table 332 of the pooling layer 330, as shown in gray. The calculated values ​​c of the convolutional layer 320 for the quality (m=3) region of the design representation image 230 for creating the learning model. l,s 3_fc The grayed-out section shows that two elements 323 and 326 of the feature map that store the data are compressed and stored in the output table 333 of the pooling layer 330. Compression of multiple elements of the feature map is achieved, for example, by calculating the maximum or average value of the values ​​stored in those elements.

[0057] In this embodiment, the calculated value of the pooling layer 330 obtained in this manner is p i,j m_fp This is how it is written. i is the position in the output table of the pooling layer 330, and is a number (row number) that identifies the position in the direction corresponding to the column direction (vertical direction) of the image input to the input layer 310, and j is the position in the output table of the pooling layer 330, and is a number (column number) that identifies the position in the direction corresponding to the row direction (horizontal direction) of the image input to the input layer 310. m_f pThis is a symbol that identifies the pooling layer 330 for manufacturing specification m. In this embodiment, one pooling layer 330 is created for each manufacturing specification m, so f p = 1. When multiple pooling layers 330 are created for one manufacturing specification m, f p f can take values ​​greater than or equal to 2 (Note that in the formula described later, f p The number F p (This is how it is written.) In Figure 6, the calculated value c of the convolutional layer 320 for a region of size (m=1) of the design representation image 230 for creating the learning model. l,s 1_fc For every two elements of the feature map that stores the calculated value c 1,1 1_1 ~c 2,1 1_1 , c 3,1 1_1 ~c 4,1 1_1 , c 5,1 1_1 ~c 6,1 1_1 , c 7,1 1_1 ~c 8,1 1_1 Compress it, size (m=1), f p The calculated value p stored in output tables 331, 334, 335, and 336 of pooling layer 330 for =1 1,1 1_1 , p 2,1 1_1 , p 3,1 1_1 , p 4,1 1_1 This shows that each component of the pooling layer 330 is calculated accordingly.

[0058] The learning model creation unit 113 creates a CNN in which the processing in the convolutional layer 320 described above is performed by supervised learning using the learning model creation plan representation image 230 and good / bad judgment data 220 created by the learning model creation image creation unit 112, and stores it in the storage unit 114. Specifically, in the example CNN configuration shown in Figure 3, the learning model creation unit 113 calculates parameters as learning parameters for the CNN, including the weights w and bias (threshold) in the convolutional layer 320, and the weights w and bias (threshold) used when fully connecting the pooling layer 330 and the output layer 340. The learning method for the CNN is implemented using known methods such as the backpropagation method, so a detailed explanation of the learning method for the CNN is omitted.

[0059] [Planning device 120] The planning device 120 creates a production plan or a logistics plan using the learned model created by the learning model creation device 110. As described above, this embodiment illustrates the case in which the learning model creation device 110 creates a learned model used to create a production plan when producing products such as steel plates. Therefore, this embodiment illustrates the case in which the planning device 120 creates a production plan.

[0060] The planning device 120 includes a product data acquisition unit 121 for planning, an image creation unit 122 for planning, a confidence level calculation unit 123, a planning representation image correction unit 124, and a planning unit 125.

[0061] Figure 7 shows an example of an overview of the plan creation method in the plan creation device 120. In Figure 7, a different plan representation image from the training model creation plan representation image 230 included in the training data is input to the CNN created by the learning model creation device 110. Then, the output layer 340 outputs the confidence level Y1 for a good plan and the confidence level Y2 for a bad plan. The plan creation device 120 calculates the confidence error, which is the error of the confidence level Y1 for a good plan and the confidence level Y2 for a bad plan against their respective target values. As mentioned above, in this embodiment, a confidence level of "1" indicates the highest confidence level, and a confidence level of "0" indicates the lowest confidence level. Therefore, in this embodiment, we illustrate the case where the target value of the confidence level Y1 for a good plan is "1" and the target value of the confidence level Y2 for a bad plan is "0".

[0062] The planning device 120 modifies the plan representation image input to the CNN based on the good plan confidence error, which is the error of the confidence level Y1 for a good plan against a target value, and the bad plan confidence error, which is the error of the confidence level Y2 for a bad plan against a target value. The planning device 120 repeatedly modifies the plan representation image in this way until the calculation termination condition is met. As the calculation termination condition, for example, one of the following conditions may be adopted: the confidence error is below a threshold, the difference between the calculated representation image and the previous value is below a predetermined value, or the number of iterations of the calculation is a predetermined value.

[0063] <Product data acquisition unit 121 for plan creation> The product data acquisition unit 121 for planning acquisition acquires product data for planning, which is different from the product data for creating a learning model that is included in the training data acquired by the training data acquisition unit 111. The product data for planning acquisition, like the product data for creating a learning model, has the values ​​of multiple items, including specifications, for each product included in a certain production plan, and the production order of that product in that production plan. In order to distinguish it from the product data (product data for creating a learning model) included in the training data acquired by the learning model creation device 110 (training data acquisition unit 111), the product data acquired by the product data acquisition unit 121 for planning acquisition is referred to as product data for planning acquisition as necessary. The product data for planning acquisition has the values ​​of multiple items (manufacturing specification m), including specifications, for each product included in the production plan to be modified, and the production order of that product in that production plan. Note that the planning device 120 does not need to acquire pass / fail judgment data 220. Furthermore, the product data for creating a learning model acquired by the product data acquisition unit 121 for planning only needs to be one set (that is, the product data for creating a learning model acquired by the product data acquisition unit 121 for planning only needs to be the product data from one production plan).

[0064] <Image creation unit for planning 122> The production plan image creation unit 122 creates a production plan representation source image and a production plan representation image based on the production plan data for plan creation obtained by the production plan data acquisition unit 121. The method for creating the production plan representation source image and the production plan representation image is the same as the method for creating the learning model representation source image 210 and the learning model representation image 230. Therefore, in the following description, in order to distinguish from the representation source image (learning model representation image 230) created by the learning model creation device 110 (learning model image creation unit 112), the representation image created by the production plan image creation unit 122 is referred to as the production plan representation image as necessary. Also, in the production plan device 120, the CNN created by the learning model creation device 110 is used. Therefore, in the following, the same symbols as those described while referring to FIGS. 2 to 6, including formulas as necessary, are used for the description. Also, for convenience of explanation, it is assumed that the production plan representation image is the image shown in FIG. 2 etc., and the production plan representation image is also described with the symbol 230 as necessary.

[0065] <Confidence calculation unit 123> The confidence calculation unit 123 calculates the confidence in the quality of the production plan represented by the production plan representation image 230 by inputting the production plan representation image 230 created by the production plan image creation unit 122 into the learning model (input layer 310 of the CNN) stored in the storage unit 114. In the present embodiment, the case where the confidence calculation unit 123 calculates the confidence Y1 for a good plan and the confidence Y2 for a bad plan is illustrated.

[0066] <Plan representation image correction unit 124> The plan representation image correction unit 124 corrects the production plan representation image 230 input into the input layer 310 of the learning model (CNN) by the confidence calculation unit 123 based on the confidence (confidence Y1 for a good plan and confidence Y2 for a bad plan) calculated by the confidence calculation unit 123. In the present embodiment, the case where the plan representation image correction unit 124 includes an evaluation value calculation unit 124a, an error partial derivative value calculation unit 124b, a correction unit 124c, and a determination unit 124d is illustrated.

[0067] <<Evaluation Value Calculation Unit 124a>> The evaluation value calculation unit 124a calculates the value of the evaluation function E based on the confidence levels (confidence level Y1 for a good plan and confidence level Y2 for a bad plan) calculated by the confidence level calculation unit 123. The evaluation function E includes a confidence error (as a parameter that defines the function), which is the error of the confidence level with respect to the target value. As described above, in this embodiment, we will exemplify a case where the confidence error is the error of the confidence level Y1 for a good plan with respect to the target value 1, and the error of the confidence level Y2 for a bad plan with respect to the target value 0, and we will exemplify a case where the evaluation function E is given by the following equation (3).

[0068]

number

[0069] In equation (3), (1-Y1) represents the good design confidence error, and (0-Y2) represents the bad design error confidence. Therefore, the smaller the absolute values ​​of the good design confidence error (1-Y1) and the bad design confidence error (0-Y2), the smaller the value of the evaluation function E. In this embodiment, we illustrate a case where the design representation image 230 input to the input layer 310 of the learning model (CNN) is modified so that the value of the evaluation function E is small. Alternatively, the design representation image 230 input to the input layer 310 of the learning model (CNN) may be modified so that the value of the evaluation function E is large. In this case, for example, the evaluation function E can be obtained by multiplying the entire right-hand side of equation (3) by (-1). In this embodiment, since the value of the evaluation function E may be used in equation (17) described later, we illustrate a case where the value of the evaluation function E is calculated in the evaluation value calculation unit 124a. However, in this embodiment, the value of the evaluation function E itself is not used in the calculation to modify the design representation image 230. Therefore, if, for example, equation (17) described later is not used, the plan representation image correction unit 124 does not need to have an evaluation value calculation unit 124a.

[0070] <<Error Partial Differential Value Calculation Unit 124b>> The error partial derivative calculation unit 124b uses the confidence levels calculated by the confidence level calculation unit 123 (confidence level Y1 for a good plan and confidence level Y2 for a bad plan) to calculate the error partial derivative value ∂E / ∂x v,h m Calculate the partial derivative value of the error ∂E / ∂x v,h m The value of the evaluation function E is the pixel value x of the plan representation image 230 used for plan creation. v,h m This is the value obtained by partial differentiation with respect to . As mentioned above, in this embodiment, the confidence calculated by the confidence calculation unit 123 is the confidence Y1 for a good plan and the confidence Y2 for a bad plan, and the confidence error is the confidence error for a good plan (1-Y1) and the confidence error for a bad plan (0-Y2). In this embodiment, the error partial derivative calculation unit 124b is exemplified as having a first partial derivative calculation unit 124b1, a second partial derivative calculation unit 124b2, and a partial derivative synthesis unit 124b3.

[0071] <<<Partial Differential Value Calculation Unit 124b1>>> The first partial derivative calculation unit 124b1 calculates the confidence level (confidence level Y1 for a good plan and confidence level Y2 for a bad plan) calculated by the confidence level calculation unit 123 and the calculated value u input to the output layer 340. k 0 Using this, the value of the evaluation function E is input to the output layer 340 as the calculated value u k 0 The first partial derivative value ∂E / u is the value obtained by taking the partial derivative with respect to . k 0 Calculate.

[0072] In this embodiment, the processing in the convolutional layer 320 is performed using the aforementioned equations (1) and (2). The following describes an example of the calculations performed in the pooling layer 330 and the output layer 340.

[0073] First, this embodiment illustrates a case where the CNN does not have a fully connected layer. In this embodiment, in the pooling layer 330, the calculated value (component) p of the pooling layer 330 is calculated using equation (4a) or (4b) below. i,j m_fp The following example illustrates the case where the neurons of the output layer 340 and each component of the pooling layer 330 are fully connected. For the sake of notation, here we illustrate the case where the 2x2 elements of the feature map constituting the convolutional layer 320 are compressed into a single element. When the calculation of equation (4a) is performed, the pooling layer 330 is a so-called max pooling layer. When the calculation of equation (4b) is performed, the pooling layer 330 is a so-called average pooling layer.

[0074]

number

[0075] Next, in this embodiment, the calculated value p stored in each output table of the pooling layer 330 is calculated by the following equation (5). i,j m_fp (Each component of pooling layer 330) Weighted linear sum u k 0 However, the calculated value u is input to the output layer 340. k 0 An example of how it is calculated is given below. k is a parameter for identifying neurons in the output layer 340, and represents a value between 1 and the number of neurons in the output layer 340 (the number of calculated values ​​output from the output layer 340) (in the following description, k will be referred to as the neuron identification parameter k of the output layer 340 as needed). In this embodiment, the confidence level Y1 for a good plan and the confidence level Y2 for a bad plan are output from the output layer 340, so "1" and "2" are set as the neuron identification parameter k of the output layer 340 (i.e., u k 0 As shown in equation (5) below, u1 0 and u2 0 (It will become...)

[0076]

number

[0077] Here, J m This represents the number of output tables of the pooling layer 330, and indicates the number of columns in the direction corresponding to the row direction (horizontal direction) of the image input to the input layer 310. m w is the number of output tables of the pooling layer 330, and indicates the number of rows in the direction corresponding to the column direction (vertical direction) of the image input to the input layer 310. i,j_k m_fp This is the calculated value (each component of the pooling layer 330) p stored in the output tables of the neurons in the output layer 340 and the pooling layer 330. i,j m_fp These are the weights used when and are fully connected. As mentioned above, in this embodiment, "1" and "2" are set as parameters k that identify neurons in the output layer 340 (i.e., w i,j_k m_fp As in equation (5), w i,j_1 m_fp and w i,j_2 m_fp (This is the result). In reality, a bias (threshold) is added to the right-hand side of equation (5), but the bias (threshold) itself is implemented using well-known techniques for CNNs. Therefore, the notation and detailed explanation of the bias (threshold) are omitted here.

[0078] Next, in this embodiment, we will illustrate a case in which the confidence level Y1 for a good plan and the confidence level Y2 for a bad plan are calculated in the output layer 340 using the following equation (6).

[0079]

number

[0080] Here, σ 0 (u) represents the activation function. Note that the activation function σ used in the convolutional layer 320 is shown. c (u) is the same as the activation function σ used in the output layer 340. 0(u) can also be any well-known function used in CNNs, such as the sigmoid function.

[0081] In this embodiment, an example of the processing in the first partial derivative calculation unit 124b1 and the second partial derivative calculation unit 124b2 will be explained by illustrating the case where the above calculations are performed in the convolutional layer 320, the pooling layer 330, and the output layer 340. Therefore, the confidence calculation unit 123 calculates the plan representation image 230 (pixel value x) for plan creation to be modified. v,h m By inputting the following into the input layer 310, the following calculated values ​​are obtained. Specifically, the confidence calculation unit 123 calculates the pixel value x of the plan representation image 230 for plan creation. v,h m The calculated value u of the convolution operation in the convolution layer 320 is obtained using l,s m_fc To calculate the value of the convolution operation u l,s m_fc The calculated value c of the convolutional layer 320 is obtained using this method. l,s m_fc To calculate the calculated value c of the convolutional layer 320 l,s 3_fc The calculated values ​​(components) of the pooling layer 330 are obtained using p i,j m_fp To calculate the calculated values ​​(each component) of the pooling layer 330 p i,j m_fp The weighted linear sum of the calculated values ​​u is input to the output layer 340. k 0 The calculation is performed as follows, and the calculated value u k 0 The calculations Y1 and Y2 (output values ​​of output layer 340) are calculated using the above method, and these steps are performed sequentially in this order.

[0082] Pixel value x of the plan representation image 230 for plan creation v,h m The partial derivative of the evaluation function E is ∂E / ∂x v,h m This is the pixel value x of the plan representation image 230 for plan creation. v,h m This represents the change in the evaluation function E in relation to the change in . ∂E / ∂xv,h m x becomes "0" v,h m However, the value of the evaluation function E is minimized by x v,h m This represents ∂E / ∂x. Therefore, ∂E / ∂x v,h m x v,h m If you modify this, the pixel values ​​x of the plan representation image 230 for plan creation will decrease so that the value of the evaluation function E decreases. v,h m This can be modified. That is, the pixel value x of the plan representation image 230 for plan creation. v,h m The amount of correction is ∂E / ∂x v,h m It can be determined based on ∂E / ∂x v,h m This is expressed by the chain rule in equation (7) below. In other words, based on the evaluation function E, the pixel values ​​x of the plan representation image 230 for plan creation v,h m The corrective gradient for correcting this is ∂E / ∂x v,h m is ∂E / u k 0 and ∂u k 0 / ∂x v,h m It can be expressed as the product of two derivatives for (k=1, 2). Therefore, the former ∂E / u k 0 The first partial derivative value calculation unit 124b1 calculates the latter ∂u k 0 / ∂x v,h m This is calculated by the second partial derivative calculation unit 124b2.

[0083]

number

[0084] (7) The ∂E / ∂u1 on the right side of equation (7) 0 , ∂E / ∂u2 0 This can be expressed as shown in equation (8) below.

[0085]

number

[0086] Here, {σ 0 (u1 0 )}' is σ 0 (u1 0 ) to u1 0 We show that this is the value obtained by differentiating with respect to {σ 0 (u2 0 )}' is σ 0 (u2 0 ) to u2 0 This indicates that it is the value obtained by differentiating with respect to . Note that this notation (expressing ' as a differential operator) is the same for other expressions as well. Also, as mentioned above, u1 0 , u2 0 This is the calculated value p stored in each output table of the pooling layer 330, which is input from the pooling layer 330 to the output layer 340. i,j m_fp This is the calculated weighted linear sum of each component of the pooling layer 330.

[0087] As mentioned above, Y1, Y2, u1 on the right side of equation (8) 0 , u2 0 This refers to the pixel value x of the plan representation image 230 (for plan creation) to be modified. v,h m The confidence level is calculated by the confidence calculation unit 123 by inputting the following to the input layer 310: σ. 0 The functional form of is predetermined. Therefore, the first partial derivative value calculation unit 124b1 calculates the first partial derivative value ∂E / u1 from equation (8). 0 , ∂E / u2 0 It is possible to calculate this.

[0088] <<<Second Partial Differential Value Calculation Unit 124b2>>> The second partial derivative calculation unit 124b2 inputs the design representation image 230 for design creation to the input layer 310 via the confidence calculation unit 123, and uses the results of the calculation performed by the trained CNN to input the calculated value u to the output layer 340. k0 (The weighted linear sum of each component of the pooling layer 330 is input from the pooling layer 330 to the output layer 340. k 0 ) Pixel value x of the planning representation image 230 for planning creation v,h m The second partial derivative value ∂u is the value obtained by partially differentiating with respect to . k 0 / ∂x v,h m Calculate.

[0089] In this embodiment, the second partial derivative calculation unit 124b2 inputs the plan creation plan representation image 230 to the input layer 310 via the confidence calculation unit 123, thereby calculating the value of the second calculation, which is performed upstream of the CNN (input layer 310), before the first calculation, which is performed downstream of the CNN (output layer 340), and assigns the calculated value to the pixel value x of the plan creation plan representation image 230. v,h m The value obtained by partially differentiating with respect to ∂u l,s m_fc / ∂x v,h m , ∂c l,s m_f / ∂x v,h m , ∂p i,j m_fp / ∂x v,h m Using this, the calculated value c from the first calculation is obtained. l,s m_fc , p i,j m_fp u k 0 Pixel value x of the plan representation image 230 for plan creation v,h m The value obtained by partially differentiating with respect to ∂c l,s m_fc / ∂x v,h m , ∂p i,j m_fp / ∂x v,h m , ∂u k 0 / ∂x v,h m The calculation involves performing the following steps in this order to obtain the second partial derivative value ∂u k0 / ∂x v,h m Let's take an example of how to calculate ∂u. l,s m_fc / ∂x v,h m Using ∂c l,s m_fc / ∂x v,h m Calculate the calculated ∂c l,s m_f / ∂x v,h m Using ∂p i,j m_fp / ∂x v,h m Calculate the calculated ∂p i,j m_fp / ∂x v,h m Using ∂u k 0 / ∂x v,h m Let's take an example of how to calculate ∂u. This calculation is done by ∂u k 0 / ∂x v,h m Until the calculation is performed (i.e., the pixel value x of the plan representation image 230 for plan creation) v,h m The first calculation, which involves partial differentiation, yields the calculated value c. l,s m_fc , p i,j m_fp u k 0 However, the calculated value u is input to the output layer 340. k 0 (until it becomes) the pixel value x v,h m The first calculation, which involves partial differentiation, yields the calculated value c. l,s m_fc , p i,j m_fp u k 0 This is done by changing the position of the CNN where the result is calculated, from the input layer 310 to the output layer 340.

[0090] As mentioned above, the pixel value x of the plan representation image 230 for plan creation. v,h m The expression shown in equation (7), ∂E / ∂xv,h m If modified accordingly, the pixel values ​​x of the plan representation image 230 for plan creation will decrease in order to reduce the value of the evaluation function E. v,h m This can be corrected. (7) The right-hand side of equation ∂E / u1 0 , ∂E / u2 0 The (first partial derivative) is calculated by the first partial derivative calculation unit 124b1. Therefore, the second partial derivative calculation unit 124b2 calculates ∂u on the right side of equation (7). k 0 / ∂x v,h m The second partial derivative value ∂u k 0 / ∂x v,h m It is calculated as follows.

[0091] As mentioned above, the confidence calculation unit 123 calculates the plan representation image 230 (pixel value x) for plan creation that is subject to correction. v,h m ) is input to the input layer 310, u l,s m_fc , c l,s m_fc , p i,j m_fp u k 0 These are calculated sequentially in this order. In this embodiment, the second partial derivative calculation unit 124b2 uses these values ​​to calculate ∂u as follows: l,s m_fc / ∂x v,h m , ∂c l,s m_fc / ∂x v,h m , ∂p i,j m_fp / ∂x v,h m , ∂u k 0 / ∂x v,h m Let's illustrate the case where these are calculated sequentially in this order.

[0092] First, ∂u l,s m_fc / ∂x v,h mFrom equation (9), which is a transformation of equation (1), it can be expressed as equation (10) below.

[0093]

number

[0094] The second partial derivative calculation unit 124b2 calculates ∂u according to equation (10). l,s m_fc / ∂x v,h m The following is calculated: Here, the weights w of kernels 411-413 used in the convolution operation in convolution layer 320 are calculated. l,s m_fc (w v-q+1,h-r+1 m_fc ) is calculated by the learning model creation device 110 (learning model creation unit 113). Therefore, according to equation (10), ∂u l,s m_fc / ∂x v,h m It is possible to calculate this.

[0095] Next, ∂c l,s m_fc / ∂x v,h m is, c l,s m_fc =σ c (u l,s m_fc )(σ c (*) is the activation function of the convolutional layer, and it can be expressed as shown in equation (11) below.

[0096]

number

[0097] The second partial derivative calculation unit 124b2 calculates ∂c according to equation (11). l,s m_fc / ∂x v,h m The following is calculated: Here, the calculated value u of the convolution operation in the convolution layer 320. l,s m_fcThis is the plan representation image 230 for creating the plan to be modified (pixel value x v,h m The confidence score is calculated in the confidence calculation unit 123 by inputting the following to the input layer 310: σ. c The functional form of is predetermined, and its derivative can be calculated. Therefore, {σ c (u l,s m_fc )}' can be calculated, and by combining it with equation (10) and equation (11), ∂c l,s m_fc / ∂x v,h m It is possible to calculate this.

[0098] Next, ∂p i,j m_fp / ∂x v,h m This can be expressed by equation (12) below, derived from equation (4b). However, here we illustrate the case where the 2x2 elements of the feature map constituting the convolutional layer 320 (the elements (2i-1,2j-1), (2i,2j-1), (2i-1,2j), and (2i,2j)) are compressed into the output table of (i,j) of the pooling layer 330.

[0099]

number

[0100] The second partial derivative calculation unit 124b2 calculates ∂p according to equation (12). i,j m_fp / ∂x v,h m The calculated value c of the convolutional layer 320 is calculated here. l,s m_fc This is the plan representation image 230 for creating the plan to be modified (pixel value x v,h m The confidence level is calculated by the confidence level calculation unit 123 by inputting the following to the input layer 310. i,j m_fp The function form of (*) is predetermined, and its derivative can be calculated. Therefore, {p i,j m_fp (c l,sm_fc )}' can be calculated, and by combining it with equation (11) and equation (12), ∂p i,j m_fp / ∂x v,h m It is possible to calculate this.

[0101] Note that the calculated value p for pooling layer 330 i,j m_fp If it is calculated by equation (4a), then ∂p i,j m_fp / ∂x v,h m This is calculated using the following equation (13).

[0102]

number

[0103] Finally, ∂u k 0 / ∂x v,h m This can be expressed as equation (15) below, which is obtained by expressing equation (5) in terms of the neuron discrimination parameter k of the output layer 340 as shown in equation (14).

[0104]

number

[0105] The second partial derivative value calculation unit 124b2 calculates ∂u according to equation (15). k 0 / ∂x v,h m (k=1,2) is calculated. Here, the calculated values ​​(each component of the pooling layer 330) p are stored in the output tables of the neurons in the output layer 340 and the pooling layer 330. i,j m_fp The weights w used when fully connecting and i,j m_fp This is calculated by the learning model creation device 110 (learning model creation unit 113). Therefore, by combining it with equation (12) and equation (14), ∂u k 0 / ∂x v,hm It is possible to calculate this.

[0106] As described above, in this embodiment, the second partial differential value calculation unit 124b2 is w v-q+1,h-r+1 m_fc Using equation (10), ∂u l,s m_fc / ∂x v,h m Calculate and its ∂u l,s m_fc / ∂x v,h m and u l,s m_fc Using (11), ∂c l,s m_fc / ∂x v,h m Calculate and its ∂c l,s m_f / ∂x v,h m and c l,s m_fc Using equation (12) or (13), ∂p i,j m_fp / ∂x v,h m Calculate and its ∂p i,j m_fp / ∂x v,h m and w i,j m_fp Using equation (15), the second partial derivative value is ∂u k 0 / ∂x v,h m Let's take an example of calculating (k=1, 2).

[0107] That is, ∂c l,s m_fc / ∂x v,h m When calculating c l,s m_fc The calculation of u is performed upstream of the CNN (input layer 310 side) rather than the calculation of (first calculation). l,s m_fc The calculated value u obtained by the calculation (second calculation) l,s m_fc to x v,h m The value obtained by partially differentiating with respect to ∂ul,s m_fc / ∂x v,h m The formula used is (see equation (11)). ∂p i,j m_fp / ∂x v,h m When calculating p i,j m_fp c is calculated upstream of the CNN, rather than the calculation of (first calculation). l,s m_fc The calculated value c obtained by the calculation (second calculation) l,s m_fc to x v,h m The value obtained by partially differentiating with respect to ∂c l,s m_fc / ∂x v,h m The formula used is (see formula (12) or formula (13)). ∂u k 0 / ∂x v,h m When calculating u k 0 p is calculated upstream of the CNN, rather than the calculation of (first calculation). i,j m_fp The calculated value p (second calculation) is obtained by the calculation of (second calculation). i,j m_fp to x v,h m The value obtained by partially differentiating with respect to ∂p i,j m_fp / ∂x v,h m This is used (see equation (15)).

[0108] As described above, in this embodiment, the pixel values ​​x of the plan representation image 230 for plan creation input to the CNN input layer 310 are calculated by the confidence calculation unit 123. v,h m The calculated values ​​u in each layer when corrected l,s m_fc , c l,s m_fc , p i,j m_fp By propagating the change in x from the input layer 310 to the output layer 340 of the CNN, v,h m If you modify u k 0Change ∂u k 0 / ∂x v,h m This is calculated.

[0109] <<<Partial partial differential value synthesis unit 124b3>>> As described above, the first partial derivative calculation unit 124b1 calculates ∂E / u1 on the right side of equation (7). 0 , ∂E / u2 0 The (first partial derivative value) is calculated, and the second partial derivative value calculation unit 124b2 calculates ∂u1 on the right side of equation (7). 0 / ∂x v,h m , ∂u2 0 / ∂x v,h m The following is calculated. The partial derivative value synthesis unit 124b3 calculates the first partial derivative value ∂E / u1 0 , ∂E / u2 0 And the second partial derivative value ∂u1 0 / ∂x v,h m , ∂u2 0 / ∂x v,h m Using and , we obtain the partial derivative value of the error ∂E / ∂x by equation (7). v,h m (v=1,2,···,v max h=1,2,···,h max m Calculate ).

[0110] <<Correction part 124c>> The correction unit 124c calculates the error partial derivative value ∂E / ∂x calculated by the error partial derivative value calculation unit 124b (partial partial derivative value synthesis unit 124b3). v,h m The plan representation image 230 for plan creation input to the learning model is modified using this. In this embodiment, the modification unit 124c modifies the error partial derivative value ∂E / ∂x using the following equation (16). v,h m The source of the calculation is the plan representation image 230 for plan creation (the pixel value x of the plan representation image 230 to be modified). v,h m An example of how to modify )) is given.

[0111]

number

[0112] Here, (R) and (R-1) represent the number of iterations in the iterative calculation of modifying the plan representation image 230 for plan creation (R is an integer greater than or equal to 1). Also, η is the learning rate (pixel value x in one iterative calculation). v,h m This parameter determines the correction range. The larger the absolute value of the learning rate η, the lower the computational load, but the lower the computational accuracy (pixel value x). v,h m The resolution of the pixels also decreases. Conversely, the smaller the absolute value of the learning rate η, the lower the computational accuracy (pixel value x). v,h m The resolution (x) increases, but the computational load also increases. The learning rate η is set in advance from this perspective. However, the learning rate η is set based on the corrected pixel value x v,h m(R) However, pixel value x v,h m The range should be determined to be within a reasonable range.

[0113] <<Determination unit 124d>> The determination unit 124d determines whether a predetermined calculation termination condition is met for the calculation to modify the plan representation image 230 for plan creation. The calculation termination condition is not particularly limited, but in this embodiment, an example is given in which the determination unit 124d determines whether a calculation termination condition is met based on the plan representation image 230 for plan creation modified by the modification unit 124c. For example, the determination unit 124d determines that the calculation termination condition is met if the value of the evaluation function E is less than or equal to the threshold ε1, as shown in equation (17) below, and determines that the calculation termination condition is not met otherwise. In addition, the determination unit 124d determines, for example, the pixel value x of the plan representation image 230 for plan creation (the latest plan representation image 230 for plan creation modified by the modification unit 124c) when the number of iterations is R, as shown in equation (18) below. v,h m(R)And the pixel value x of the plan creation plan representation image 230 (the plan creation plan representation image 230 that was modified one iteration earlier by the modification unit 124c) when the number of iterations is R-1. v,h m(R-1) The calculation termination condition may be determined to be satisfied if the maximum absolute value of the difference between the two is less than or equal to the threshold ε2, and not satisfied if it is not. Alternatively, as mentioned above, the calculation termination condition may be that the number of iterations of the calculation is a predetermined value.

[0114]

number

[0115] If the confidence calculation unit 123 determines, based on the determination unit 124d, that the calculation termination condition is not met, it inputs the latest plan representation image 230 for plan creation, corrected by the correction unit 124c, to the input layer 310 as the new plan representation image 230 to be corrected. The confidence calculation unit 123 and the plan representation image correction unit 124 then repeat the process described above. The confidence calculation unit 123 and the plan representation image correction unit 124 repeat this process until the calculation termination condition is met. Then, if the determination unit 124d determines that the calculation termination conditions are met, the plan representation image modification unit 124 determines the latest plan representation image 230 for plan creation, modified by the modification unit 124c, as the final modified plan representation image 230 for plan creation.

[0116] <Planning Department 125> The planning unit 125 creates a production plan based on the modified planning representation image 230, if the plan representation image 230 modified by the planning representation image modification unit 124 satisfies the predetermined calculation completion conditions for the calculations to modify the plan representation image 230. In the following description, the plan representation image 230 will be referred to as the final modified plan representation image 230, as necessary.

[0117] In this embodiment, the pixel value x of the final revised plan creation plan representation image 230 v,h m The result of the calculation in equation (16) can take values ​​other than "0" and "1". Therefore, the planning unit 125 calculates the pixel value x of the final revised plan representation image 230 for planning. v,h m The maximum value is converted to "1", and all other values ​​are converted to "0", and this process is performed individually for each manufacturing specification m. In the following explanation, the pixel value x is converted in this way. v,h m The final modified plan representation image 230 for plan creation, which has been converted, is referred to as the optimal plan representation image 230 for plan creation, if necessary. For example, let's assume that the optimal plan representation image 230 for plan creation is the one shown in Figure 4. In Figure 4, the black areas, which indicate a pixel value of "1", specify what kind of products with what manufacturing specifications m (size, material, quality) should be manufactured and in what production order. In this embodiment, the plan creation device 120 creates the optimal plan representation image 230 for plan creation as a production plan.

[0118] When actually producing a product, the production plan is used in the following ways, for example: First, the planning unit 125 inputs data indicating the manufacturing specifications m for each of several products that need to be produced. Then, for each product included in the input data, the planning unit 125 identifies the product for which all manufacturing specification m values ​​correspond to pixels with a pixel value of "1" (black pixels) in the optimal planning representation image 230. This determines the production order for all products included in the input data. The planning unit 125 outputs information including the production order of the products determined in this way as information indicating the production plan. At least one of the following forms of output is used: display on a computer display, storage in an internal or external storage device of the planning device 120, and transmission to an external device.

[0119] Furthermore, the planning unit 125 may output an optimal planning representation image 230 for planning and present it to the plan creator. In this case, the plan creator determines the production order of each product by identifying the product for which all manufacturing specification m values ​​in the optimal planning representation image 230 correspond to pixels with a pixel value of "1" (black pixels). In this case, the planning device 120 may input and store information including the manufacturing specification m, delivery date, and production order of each product determined in this way.

[0120] [flowchart] Next, with reference to the flowchart in Figure 8, an example of a learning model creation method performed using the learning model creation device 110 of this embodiment will be described.

[0121] First, in step S801, the training data acquisition unit 111 acquires training data. The training data includes product data for creating a learning model and quality judgment data 220. The product data for creating a learning model includes the value of the manufacturing specification m for each product included in a certain production plan, and the production order of the product in that production plan. The quality judgment data 220 includes data indicating "1" or "0" as the confidence level Y1 for a good plan and the confidence level Y2 for a bad plan.

[0122] Next, in step S802, the learning model creation image creation unit 112 creates a learning model creation plan representation image 230 based on the learning model creation product data included in the training data in step S801.

[0123] Next, in step S803, the learning model creation unit 113 performs supervised learning using the learning model creation plan representation image 230 created in step S802 to create a learning model (CNN) that takes the plan representation image as input and outputs confidence levels for the quality of the plan (confidence level Y1 for a good plan and confidence level Y2 for a bad plan).

[0124] Next, in step S804, the learning model creation unit 113 stores the information of the learning model (CNN) created in step S803 in the storage unit 114.

[0125] Next, an example of a planning method performed using the planning device 120 of this embodiment will be described with reference to the flowchart in Figure 9. The flowchart in Figure 9 is executed after the learning model (CNN) information has been stored in the storage unit 114 according to the flowchart in Figure 8.

[0126] First, in step S901, the product data acquisition unit 121 for plan creation acquires product data for plan creation. The product data for plan creation includes, similar to the product data for learning model creation, the value of the manufacturing specification m for each product included in a particular production plan, and the production order of that product in that production plan.

[0127] Next, in step S902, the planning image creation unit 122 creates a planning representation image 230 based on the planning product data acquired in step S901. Next, in step S903, the confidence calculation unit 123 inputs the plan creation plan representation image 230 created in step S902 into the learning model (CNN input layer 310) stored in the memory unit 114 by executing the process shown in the flowchart of Figure 8, and calculates the confidence level (confidence level Y1 for a good plan and confidence level Y2 for a bad plan) regarding the quality of the production plan represented by the plan creation plan representation image 230.

[0128] Next, in step S904, the evaluation value calculation unit 124a uses the confidence levels calculated in step S903 (confidence level Y1 for a good plan and confidence level Y2 for a bad plan) to calculate the value of the evaluation function E using equation (3). Next, in step S905, the first partial derivative calculation unit 124b1 calculates the confidence level (confidence level Y1 for a good plan and confidence level Y2 for a bad plan) calculated in step S903 and the calculated value u1 input to the output layer 340. 0 , u20 Using this, from equation (8), the first partial derivative value ∂E / u1 0 , ∂E / u2 0 Calculate.

[0129] Next, in step S906, the second partial derivative value calculation unit 124b2 calculates ∂u according to equation (11). l,s m_fc / ∂x v,h m Using ∂c l,s m_fc / ∂x v,h m Calculate and ∂c by equation (13) or (14). l,s m_f / ∂x v,h m Using ∂p i,j m_fp / ∂x v,h m Calculate and ∂p by equation (15) i,j m_fp / ∂x v,h m Using ∂u k 0 / ∂x v,h m The calculation involves performing the following steps in this order: Calculating the second partial derivative value ∂u k 0 / ∂x v,h m Calculate.

[0130] Next, in step S907, the partial derivative value synthesis unit 124b3 combines the first partial derivative value ∂E / u1 calculated in step S905. 0 , ∂E / u2 0 And the second partial derivative value ∂u1 calculated in step S906 0 / ∂x v,h m , ∂u2 0 / ∂x v,h m Using and , we obtain the partial derivative value of the error ∂E / ∂x by equation (7). v,h m (v=1,2,···,v max h=1,2,···,h maxm Calculate ).

[0131] Next, in step S908, the correction unit 124c calculates the partial derivative value of the error ∂E / ∂x calculated in step S907. v,h m By performing the calculation in equation (16) using this method, the plan representation image 230 for plan creation is modified.

[0132] Next, in step S909, the determination unit 124d determines whether the calculation termination condition is met by performing the calculation of equation (18) or equation (19). If the calculation termination condition is not met as a result of this determination (NO in step S909), the process in step S903 is executed again. In this case, in the process in step S903, the plan creation plan representation image 230 modified in step S908 (not the plan creation plan representation image 230 created in step S902) is used as the plan creation plan representation image 230 to be modified.

[0133] On the other hand, if the calculation termination condition is met (YES in step S909), the process in step S910 is performed. In step S910, the planning unit 125 creates a production plan based on the plan representation image 230 for plan creation that was modified in step S908, immediately preceding the step in which it was determined that the calculation termination condition was met in step S909.

[0134] [summary] As described above, in this embodiment, the learning model creation device 110 acquires training data including product data for creating a learning model, which has the value of the manufacturing specification m for each of the multiple products included in the production plan and the production order of the products in the production plan, and quality judgment data indicating the quality of the production plan. Based on the product data for creating a learning model, the learning model creation device 110 creates a plan representation image 230 for creating a learning model, which is an image in which each of the multiple manufacturing specifications m corresponds to a column, the production order of the multiple products corresponds to a row, and the value of the manufacturing specification m corresponds to a pixel value. By performing supervised learning using the plan representation image 230 for creating a learning model, the learning model creation device 110 takes the plan representation image as input and outputs a confidence level for the quality of the production plan (confidence level Y1 for a good plan and confidence level Y2 for a bad plan).Therefore, it is possible to construct a learning model for creating highly accurate production plans that take tacit knowledge into consideration, with as little human intervention as possible. For example, in cases where there are many constraints and evaluation items to consider when creating a production plan, and these are not always clear, or in cases where the plan is based on tacit knowledge derived from the planner's experience and intuition, the learning model can be automatically trained to automatically extract the features necessary for creating a good production plan and create a production plan that reflects those features.

[0135] In this embodiment, the planning device 120 inputs the planning representation image 230 for planning into a trained model (CNN) to calculate the confidence level for the quality of the production plan (confidence level Y1 for a good plan and confidence level Y2 for a bad plan), and modifies the planning representation image 230 input into the trained model (CNN) based on the calculated confidence levels. Then, if the planning device 120 satisfies the predetermined calculation termination conditions based on the modified planning representation image 230 for planning, it creates a production plan based on the planning representation image 230. If the conditions are not met, it inputs the modified planning representation image 230 into the trained model to recalculate the confidence level for the quality of the production plan. Therefore, a highly accurate production plan that takes tacit knowledge into consideration can be created with as little human intervention as possible.

[0136] Furthermore, the embodiments of the present invention described above can be realized by a computer executing a program. A computer-readable recording medium on which the program is stored, and a computer program product such as the program itself, can also be applied as embodiments of the present invention. Examples of recording media include flexible disks, hard disks, optical disks, magneto-optical disks, CD-ROMs, magnetic tapes, non-volatile memory cards, ROMs, etc. Moreover, embodiments of the present invention may be realized by dedicated hardware such as an ASIC (Application Specific Integrated Circuit). Furthermore, the embodiments of the present invention described above are merely examples of how the invention can be implemented, and the technical scope of the invention should not be interpreted as being limited by them. In other words, the present invention can be implemented in various forms without departing from its technical concept or its main features. [Explanation of symbols]

[0137] 110 Learning Model Creation Device 111 Training Data Acquisition Unit 112 Image creation section for creating learning models 113 Learning Model Creation Section 114 Storage section 120 Planning device 121 Product data acquisition unit for planning 122 Image Creation Department for Planning 123 Confidence calculation part 124 Planned Representation Image Correction Section 124a Evaluation Value Calculation Unit 124b Error Partial Differential Value Calculation Unit 124b1 Partial Differential Value Calculation Unit (First Part) 124b2 Second Partial Differential Value Calculation Unit 124b3 Partial partial differential value synthesis unit 124c Correction section 124d Judgment section 125 Planning Department 210 Source image for design representation used to create the learning model 220 pass / fail judgment data 230 Plan Representation Source Images (Plan Representation Images for Creating Learning Models, Plan Representation Images for Creating Plans) 310 Input Layers 320 Convolutional Layers 321-326 Feature map elements that make up the convolutional layer 330 Pooling Layers 331-333 Output table of the pooling layer (elements of the compressed feature map) 340 output layers 411-413 Kernel (Filter)

Claims

1. A learning model creation device that creates a learning model used to create a plan, such as a production plan or a logistics plan, A means for acquiring training data that acquires training data including product data for each of the multiple products included in the plan, which is product data for creating a learning model having values ​​for multiple items including the specifications of the product and the processing order of the product in the plan, and quality judgment data indicating the quality of the plan, A learning model creation image creation means creates a learning model creation image, which is an image in which each of the multiple items is associated with either a row or a column, the processing order of the multiple products is associated with the other row or column, and the values ​​of the items are associated with pixel values, based on the product data for creating the learning model included in the training data. A learning model creation means that creates a learning model that takes the plan representation image as input and outputs a confidence level regarding the quality of the plan, by performing supervised learning using the learning model creation plan representation image created by the learning model creation image creation means, A learning model creation device equipped with the following features.

2. The learning model creation device according to claim 1, wherein the learning model is a CNN (Convolutional Neural Network).

3. Let M be the number of the aforementioned items. The kernel used during the convolution operation in the CNN is set individually for one of the items or for two or more items less than M in the learning model design representation image. The learning model creation apparatus according to claim 2, wherein the kernel set for one item or two or more items less than M is slid only within the region corresponding to the one item or two or more items less than M in the planned representation image.

4. The size in the longitudinal direction of the row or column of the kernel set for one item or two or more items less than M that corresponds to the processing order of the multiple products in the planned representation image is the same as the size in that direction of the area corresponding to the one item or two or more items less than M in the planned representation image. The learning model creation device according to claim 3, wherein the kernel set for one item or two or more items less than M is slid only in the longitudinal direction of the rows or columns corresponding to the plurality of items in the planned representation image.

5. A planning device that creates the plan using the learning model created by the learning model creation device according to any one of claims 1 to 4, A means for acquiring product data for planning, which is product data different from the product data for creating the learning model, A plan creation image creation means that creates a plan creation plan representation image, which is a plan representation image, based on the product data for plan creation, A confidence calculation means inputs the plan representation image created by the plan creation image creation means into the learning model and calculates a confidence level regarding the quality of the plan. A plan representation image modification means modifies the plan representation image for plan creation input to the learning model based on the confidence level calculated by the confidence level calculation means, When the plan creation plan representation image modified by the plan representation image modification means satisfies predetermined calculation completion conditions for calculations to modify the plan creation plan representation image, the plan creation means creates the plan based on the modified plan creation plan representation image. Equipped with, The confidence calculation means is a plan creation device that, if the plan creation plan representation image corrected by the plan representation image correction means does not satisfy the calculation termination condition, inputs the plan creation plan representation image corrected by the plan representation image correction means into the learning model and recalculates the confidence.

6. The aforementioned plan representation image modification means is An error partial derivative value calculation means calculates an error partial derivative value, which is the value obtained by partially differentiating the value of an evaluation function, which includes a confidence error (the error of the confidence value relative to a target value), with respect to the pixel values ​​of the plan representation image for plan creation, using the confidence level calculated by the confidence level calculation means. A correction means for correcting the plan representation image for plan creation input to the learning model using the error partial derivative value calculated by the error partial derivative value calculation means, A planning device according to claim 5, having the following features.

7. The planning device creates the plan using the CNN created by the learning model creation device according to any one of claims 2 to 4. The error partial derivative calculation means is: A first partial derivative calculation means calculates a first partial derivative value, which is the value obtained by partially differentiating the value of the evaluation function with respect to the calculated value input to the output layer of the CNN, using the confidence level calculated by the confidence level calculation means and the calculated value input to the output layer of the CNN. A second partial derivative calculation means calculates a second partial derivative value, which is the value obtained by partially differentiating the calculated value input to the output layer of the CNN with respect to the pixel value of the plan representation image for plan creation, using the results of calculations performed by the CNN by inputting the plan representation image for plan creation into the input layer of the CNN. A partial derivative value synthesis means for calculating the error partial derivative value using the second partial derivative value and the first partial derivative value, A planning device according to claim 6, having the following features.

8. The second partial derivative calculation means calculates the second partial derivative by using the value obtained by partially differentiating the value obtained by the second calculation, which is a calculation performed on the input layer side of the CNN than the first calculation, with respect to the pixel value of the plan creation plan representation image, and performing a calculation that includes changing the position of the CNN where the calculated value obtained by the first calculation, which is partially differentiated with respect to the pixel value of the plan creation plan representation image, is calculated from the input layer side to the output layer side of the CNN until the calculated value obtained by partially differentiating the value obtained by the first calculation, which is partially differentiated with respect to the pixel value of the plan creation plan representation image, is a calculated value input to the output layer of the CNN, thereby calculating the second partial derivative.

9. A method for creating a learning model used to create a plan, such as a production plan or a logistics plan, A training data acquisition step includes acquiring training data which is product data for each of the multiple products included in the plan, which is product data for creating a learning model having values ​​for multiple items including the specifications of the product and the processing order of the product in the plan, and good / bad judgment data indicating the goodness or badness of the plan, The learning model creation image creation means creates a learning model creation image as a learning model creation plan representation image, which is an image in which each of the multiple items is associated with either a row or a column, the processing order of the multiple products is associated with the other row or column, and the values ​​of the items are associated with pixel values, based on the learning model creation product data included in the training data. A learning model creation means performs supervised learning using the learning model creation plan representation image created in the learning model creation image creation step to create a learning model that takes the plan representation image as input and outputs a confidence level regarding the quality of the plan. A method for creating a learning model, comprising the following features.

10. A planning method for creating a plan using the learning model created by the learning model creation method described in claim 9, The product data acquisition means for plan creation includes a product data acquisition step for plan creation, which acquires product data for plan creation that is different from the product data for learning model creation, A planning image creation means creates a planning image, which is a planning representation image, based on the planning product data, and The confidence calculation means includes a confidence calculation step in which the plan representation image for plan creation created by the plan creation image creation step is input to the learning model and a confidence level is calculated regarding the quality of the plan, The plan representation image modification means modifies the plan representation image for plan creation input to the learning model based on the confidence level calculated in the confidence level calculation step, A planning creation means provides a planning creation step in which, when the planning representation image for planning creation modified by the planning representation image modification step satisfies predetermined calculation completion conditions for calculations to modify the planning representation image for planning creation, the plan is created based on the modified planning representation image for planning creation. Equipped with, The confidence calculation step is a plan creation method in which, if the plan creation plan representation image corrected by the plan representation image correction step does not satisfy the calculation termination condition, the plan creation plan representation image corrected by the plan representation image correction step is input into the learning model and the confidence is recalculated.

11. A program for causing a computer to function as one of the means of the learning model creation apparatus according to any one of claims 1 to 4.

12. A program for causing a computer to function as one of the means of the planning device according to any one of claims 5 to 8.

Images