Denture manufacturing scheduling control method, device, equipment and medium

Abstract

The application discloses a denture manufacturing scheduling control method, device, equipment and medium, and relates to the field of production and manufacturing. The method comprises the following steps: acquiring a batch order set; acquiring a preset process node sequence of each denture manufacturing order and a priority score based on multi-dimensional characteristics based on order information of the batch order set to obtain an order priority set queue; the higher the priority score is, the higher the urgency of completing the corresponding denture manufacturing order is; establishing an MDCN decision tree based on process types and corresponding resources; and matching corresponding resources for each preset process node based on preset rule constraints and the MDCN decision tree to obtain a feasible scheme set, wherein the feasible scheme set comprises at least one feasible scheme, and each feasible scheme comprises resources matched by each preset process node and operation time of the resources; and performing optimization screening on the feasible scheme set to obtain a target scheme. The application can improve the synergy between resources and the comprehensive utilization rate of resources.

Description

Technical Field

[0001] This application relates to the field of production scheduling technology, and in particular to a method, device, equipment and medium for scheduling control of denture manufacturing. Background Technology

[0002] With the continuous and in-depth penetration and widespread application of digital technology in the field of oral healthcare, the dental prosthesis manufacturing industry is rapidly developing and transforming towards digitalization and high efficiency at an unprecedented pace.

[0003] A typical intelligent manufacturing unit for dentures usually consists of multiple key components and heterogeneous resources, covering the entire process from design to finished product. Management and requirements of the manufacturing process ensure a safe, comfortable, and durable user experience for patients.

[0004] In existing resource scheduling methods, each module operates independently, lacking global coordination and making it difficult to guarantee the comprehensive utilization rate of resources. Summary of the Invention

[0005] The purpose of this application is to provide a method, device, equipment and medium for scheduling and controlling the manufacturing of dentures, which can improve the synergy between resources and the comprehensive utilization rate of resources.

[0006] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a method for scheduling and controlling the manufacturing of dentures, including: Obtain a batch order set; wherein, the batch order set includes order information corresponding to several denture manufacturing orders; Based on the order information, a preset process node sequence and a priority score based on multi-dimensional features are obtained for each denture manufacturing order; the denture manufacturing orders are sorted from high to low based on the priority scores to obtain an order priority set queue; wherein, the preset process node sequence includes multiple preset process nodes; the higher the priority score, the higher the urgency of completing the corresponding denture manufacturing order; An MDCN decision tree is established based on the process type and the corresponding resources. Based on preset rule constraints and the MDCN decision tree, the corresponding resources are matched for the preset process node to obtain a set of feasible solutions. The set of feasible solutions includes at least one feasible solution, and the feasible solution includes the resources matched for each preset process node and the operation time of the resources. The set of feasible solutions is optimized and filtered to obtain a target solution, which is one of the feasible solutions.

[0007] Optionally, the step of establishing an MDCN decision tree based on process type and corresponding resources, and matching the corresponding resources for the preset process node based on preset rule constraints and the MDCN decision tree to obtain a set of feasible solutions, includes: A preset process node priority sequence is obtained based on the preset process node sequence and the order priority queue; wherein, the preset process node priority sequence includes the preset process nodes corresponding to all the denture manufacturing orders; Based on the order information, obtain the process requirement information of the preset process node; The MDCN decision tree is constructed using process type as internal node, resource corresponding to the process type as leaf node, and capability parameters of the resource as attributes of the leaf node. The preset process node sequence, expected operation time, and process requirement information are input into the MDCN decision tree. The feasible solution set is obtained by traversing the MDCN decision tree based on preset constraint rules and the input items. The feasible solution set includes obtaining multiple feasible solutions, and each feasible solution includes the resource matched to each preset node and the operation time window of the resource.

[0008] Optionally, the preset constraint rules include: For each of the preset process nodes, the resource parameters and the required parameters can be matched; For each denture manufacturing order, the end time of each preset process node in the preset process node sequence is later than the start time of the previous preset process node. For each of the aforementioned resources, the corresponding job duration must be less than the duration of the corresponding available time window; The time to complete the last preset process node for each denture manufacturing order is less than the expected completion time of the denture manufacturing order.

[0009] Optionally, the preset constraint rules also include process interruptibility constraints, interruption recovery constraints, and resource preemption constraints; The interruptibility constraint of the process includes: for a preset process node that allows interruption, the operation time window can be divided into multiple non-overlapping sub-windows; The interruption recovery constraint includes: the operation duration after interruption recovery includes a preset skill recovery compensation time; The resource preemption constraint includes: when a high-priority order with a priority score higher than the threshold arrives, the interruptible process of the low-priority order is allowed to be interrupted and the corresponding resources are released.

[0010] Optionally, the order information includes order identifier, order type, user identifier, order delivery time, and order amount; the order type includes denture type and denture material; and the step of obtaining a preset process node sequence and a priority score based on multi-dimensional features for each denture manufacturing order based on the order information includes: For each of the aforementioned denture manufacturing orders, a preset sequence of process nodes is obtained based on the denture type and the denture material; The multidimensional features of the denture manufacturing order are obtained based on the order identifier, the order type, the user identifier, the order delivery time, the order amount, and the corresponding preset process node sequence; wherein, the multidimensional features include one or more of the following: delivery urgency feature, customer level value feature, material scarcity index feature, process complexity score feature, quality risk level feature, equipment health status dependence feature, and order economic value density feature. The weights corresponding to the multidimensional features are obtained by learning the weights of the multidimensional features based on the reinforcement learning model. The priority score is obtained based on the multidimensional features and their corresponding weights.

[0011] Optionally, when establishing an MDCN decision tree based on process type and corresponding resources, and matching the corresponding resources for the preset process node based on preset rule constraints and the MDCN decision tree to obtain a set of feasible solutions, the method further includes: Based on historical rework data, the single-batch rework probability of each preset process node is obtained; in the set of feasible solutions, the total expected rework cost is calculated for each feasible solution, and the total expected rework cost = Σ(rework probability of each preset process node × rework time of that node × corresponding resource unit time cost + rework material loss cost). When an actual rework event is detected, an incremental fast rescheduling is immediately triggered based on the current order priority set queue. The rework process is then inserted as a high-priority task into the idle window of the current scheduling sequence or preempted for interruptible resources.

[0012] Optionally, obtaining the multidimensional features of the denture manufacturing order based on the order identifier, the order type, the user identifier, the order delivery time, the order amount, and the corresponding preset process node sequence includes: The delivery urgency is determined based on the order delivery time and the current time. Obtain the corresponding customer category and customer historical cooperation data based on the user identifier, and obtain the customer level value based on the customer category and customer historical cooperation data; Based on the denture material, obtain the corresponding material parameters, and based on the material parameters, obtain the material scarcity attribute; wherein, the material parameters include the current inventory, safety stock threshold, current market price and base price, and the current delivery date and average historical delivery date of the material obtained from the supplier; The standard deviation and average working time of each preset process node are obtained according to the preset process node sequence, and the process complexity score is obtained according to the standard deviation and average working time. A preset quality risk coefficient is obtained based on the denture material; a preset process risk coefficient is obtained based on the preset process nodes; and a preset severity coefficient is obtained based on the denture type. The quality risk level is then determined based on the quality risk coefficient, the process risk coefficient, and the severity coefficient. Obtain the corresponding equipment set according to the preset process node, and obtain the equipment health status dependency based on the equipment set; The total processing time is obtained based on the denture type and denture material, and the economic benefits of the order are obtained based on the order amount and the total processing time. The delivery urgency, customer level value, material scarcity index, process complexity score, quality risk level, equipment health status dependence, and order economic value density are normalized respectively to obtain the delivery urgency feature, customer level value feature, material scarcity index feature, process complexity score feature, quality risk level feature, equipment health status dependence feature, and order economic value density feature.

[0013] Optionally, the reinforcement learning model is a reinforcement learning-driven weight generation model, which includes a deep neural network module and a policy learning module; wherein: The current state S_t of the deep neural network module is composed of statistical features, load rate vector, order scheduling effect, and the multidimensional features. The deep neural network module outputs an action A_t based on the input current state S_t, and the action A_t includes a weight vector composed of the weights corresponding to each of the multidimensional features; The policy learning module is used to update the network parameters of the deep neural network module based on the current state S_t, the action A_t, and the reward R_t.

[0014] Optionally, the step of optimizing and filtering the set of feasible solutions to obtain the target solution includes: For each feasible solution, obtain the comprehensive delay index, comprehensive utilization rate, and comprehensive resource utilization rate of the feasible solution; obtain a comprehensive score based on the comprehensive delay index, comprehensive utilization rate, and comprehensive resource utilization rate, and select the feasible solution with the highest comprehensive score as the target solution.

[0015] Secondly, this application provides a denture manufacturing scheduling and control device, comprising: The acquisition module is used to acquire a batch order set; wherein, the batch order set includes order information corresponding to several denture manufacturing orders; The processing module is used to obtain a preset process node sequence and a priority score based on multi-dimensional features for each denture manufacturing order based on the order information; sort the denture manufacturing orders from high to low based on the priority scores to obtain an order priority set queue; wherein, the preset process node sequence includes multiple preset process nodes; the higher the priority score, the higher the urgency of completing the corresponding denture manufacturing order; The matching module is used for: An MDCN decision tree is established based on the process type and the corresponding resources. Based on preset rule constraints and the MDCN decision tree, the corresponding resources are matched for the preset process node to obtain a set of feasible solutions. The set of feasible solutions includes at least one feasible solution, and the feasible solution includes the resources matched for each preset process node and the operation time of the resources. The set of feasible solutions is optimized and filtered to obtain a target solution, which is one of the feasible solutions.

[0016] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the denture manufacturing scheduling control method described above.

[0017] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the denture manufacturing scheduling control method described above.

[0018] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the denture manufacturing scheduling control method described above.

[0019] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a method, device, equipment, and medium for scheduling and control of denture manufacturing. By acquiring a batch of orders and calculating priority scores based on multi-dimensional feature information, it solves the problem of difficulty in quantifying and assessing the urgency of orders, and realizes scientific sorting and dynamic queue management of denture manufacturing orders based on their urgency. Through the step MDCN decision tree model and multi-objective optimization algorithm, it solves the problems of process resource allocation conflicts and difficulty in selecting the best solution in production scheduling, and realizes the rapid generation of feasible scheduling schemes and selection of the optimal objective solution under complex constraints, effectively improving resource utilization efficiency and order delivery timeliness. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 An application environment diagram of a denture manufacturing scheduling and control method provided in an embodiment of this application; Figure 2 A schematic flowchart illustrating a denture manufacturing scheduling and control method according to an embodiment of this application; Figure 3 for Figure 2 A detailed flowchart of step 202; Figure 4 for Figure 2 A detailed flowchart of step 203; Figure 5 This is a schematic diagram of an interruptible process resource preemption process provided in an embodiment of this application; Figure 6 This is a schematic diagram of a rapid rescheduling process for rework events provided in an embodiment of this application; Figure 7 A schematic diagram of the functional modules of a denture manufacturing scheduling and control device provided in an embodiment of this application; Figure 8 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] It should be noted that the terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented, for example, in orders other than those illustrated or described herein.

[0024] It should be noted that "at the time of..." in the embodiments of this application can be either at the instant when a certain situation occurs, or for a period of time after the occurrence of a certain situation. The embodiments of this application do not make specific limitations on this.

[0025] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0026] The denture manufacturing scheduling control method provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, terminal 102 is the dentist's port. After diagnosing a patient, medical staff input specific denture manufacturing orders and order information through terminal 102. The resource pool is used to obtain real-time information on all resources used for denture manufacturing. Server 101 obtains denture manufacturing orders and order information through terminal 102, and obtains real-time information on manufacturing resources through the resource pool (or related management equipment, which connects to various devices or terminals through interfaces to obtain real-time resource data). Through the denture manufacturing scheduling and control method provided in this application, resources in the resource pool are scheduled to complete the corresponding orders, thereby improving the synergy between resources and the overall utilization rate of resources.

[0027] In one exemplary embodiment, such as Figure 2 As shown, a method for scheduling and controlling the manufacturing of dentures is provided. This method is executed by a computer device, specifically a terminal or server, or both. In this embodiment, the method is applied to... Figure 1 Taking server 101 as an example, the explanation includes the following steps 201 to 204. Wherein: Step 201: Obtain a batch order set; wherein, the batch order set includes order information corresponding to several denture manufacturing orders; Specifically, order information includes one or more of the following: order identifier, order type, user identifier, and order delivery time. Specifically, the order identifier is used to indicate the order, and each order has a unique order identifier; As one example, the order type includes tooth position information, size information, denture type, and denture material; As one embodiment, tooth position information includes tooth number, used to indicate which tooth was manufactured; As one example, the dimensional information may include basic information such as crown height and root length, which is used as a reference for material selection during manufacturing and as basic information for subsequent manufacturing processes when sent to CAD nodes.

[0028] Specifically, the user identifier is used to identify the corresponding user, and each user has a unique identifier; Step 202: Obtain the preset process node sequence and priority score based on multi-dimensional features for each denture manufacturing order based on the order information; sort the denture manufacturing orders from high to low based on the priority scores to obtain the order priority set queue; wherein, the preset process node sequence includes multiple preset process nodes; the higher the priority score, the higher the urgency of completing the corresponding denture manufacturing order; Specifically, different order types correspond to different preset process node sequences, which are used to indicate the manufacturing sequence of different process nodes and between different process nodes; Specifically, the multidimensional feature information represents the urgency of the order in different dimensions. Based on the priority obtained from the multidimensional information, the higher the priority score, the higher the urgency of completing the corresponding denture manufacturing order. Step 203: Establish an MDCN decision tree based on the process type and corresponding resources. Match corresponding resources for preset process nodes based on preset rule constraints and the MDCN decision tree to obtain a set of feasible solutions. The set of feasible solutions includes at least one feasible solution, and each feasible solution includes the resources matched to each preset process node and the operation time of the resources. Specifically, based on the set of feasible solutions obtained from the MDCN decision tree model, resources that can work are allocated to each preset process node, ensuring that the order of operations of the preset process nodes in each denture manufacturing order remains unchanged and that the urgency requirement for order completion is met, thereby achieving reasonable resource allocation.

[0029] Step 204: Optimize and filter the set of feasible solutions to obtain the target solution, which is one of the feasible solutions.

[0030] By implementing steps 201 to 204 above, the urgency of orders can be quantified and ranked by acquiring batch order information and extracting multi-dimensional feature priority scores, which are then considered in resource allocation. By constructing an order priority set queue and combining it with the MDCN (Multi-dimensional Constraint Nested Decision Tree) model for resource matching, the problems of resource conflicts and scheduling chaos under multiple orders and multiple nodes are solved. This achieves systematic integration and dynamic matching of resources required by each process node, improving resource synergy and overall resource utilization.

[0031] In another exemplary embodiment of this application, such as Figure 3 As shown, step 202 includes the following steps 301 to 304: Step 301: Obtain the preset process node sequence for each denture manufacturing order based on the denture type and denture material; Specifically, order types include tooth position information, tooth size, denture type, and denture material; As one example, denture types include primary classification and secondary classification; The primary categories include: fixed restoration, implantation restoration, and active restoration. Among them, the secondary categories of fixed restorations include: single crowns, bridges, complete dentures, invisible dentures, partially removable dentures, and various types of brux pads; Among them, the secondary categories corresponding to planting restoration include: single crown, screw-fixed crown, bridge body, abutment, guide plate, planting pole clip, etc. Among them, the secondary categories corresponding to active restorations include: single crowns, bridges, complete dentures, invisible dentures, partially removable dentures, and various types of brux pads; Specifically, denture materials include a variety of materials such as metals (e.g., ordinary steel crowns, nickel-chromium alloys, cobalt-chromium alloys, etc.), ceramics (e.g., zirconium oxide, cast porcelain, porcelain-fused-to-metal, etc.), and synthetic resins (e.g., thermosetting resins, self-curing resins, light-curing resins, etc.).

[0032] Specifically, based on the type and material of the denture, the corresponding preset process node sequence is obtained through a preset process sequence mapping table; the preset process node sequence includes multiple preset process nodes arranged in the order of process manufacturing. Specifically, based on the characteristics of the manufacturing process, the order of the preset process nodes cannot be changed. Furthermore, considering the flexibility of manufacturing different types of dentures, different denture types and denture materials are mapped to different preset process node sequences.

[0033] As one embodiment, the preset process nodes include CAD design nodes, 3D printing nodes, machining nodes, porcelain baking nodes, polishing / finishing nodes, glazing / dyeing nodes, and sintering nodes; the process node sequence includes multiple preset process nodes arranged in the order of process manufacturing.

[0034] Specifically, in most cases, the preset process node sequence includes: CAD design node - 3D printing node - grinding / finishing node - glazing / dyeing node - sintering node; or CAD design node - machining node - grinding / finishing node - glazing / dyeing node - sintering node.

[0035] Specifically, from a process perspective, the choice between 3D printing and machine cutting is determined by material properties, precision requirements, cost-effectiveness, and other factors. For example, zirconia is a high-hardness ceramic that cannot be directly printed (or the printing cost is extremely high). Zirconia is typically cut using a cutting block (pre-sintered block) followed by sintering.

[0036] Furthermore, if the material is metal, there should also be a porcelain-baking joint between the polishing / finishing joint and the glazing / staining joint.

[0037] Furthermore, in some special types of dentures, such as complete dentures and removable dentures, sintering is usually not required. Invisible dentures, bruxism pads, and implant clasps, which need to be made of transparent materials or use metal parts, may not require glazing / staining nodes. Furthermore, the preset process nodes also include quality inspection nodes. Quality inspection nodes are generally set as the last node in the preset process node sequence. Quality inspection nodes include manual quality inspection nodes and / or AI quality inspection nodes. Depending on the type and material of the denture, some orders only need to go through one of the manual quality inspection nodes or the AI ​​quality inspection nodes to complete the quality inspection, while others need both manual and AI quality inspection nodes.

[0038] Based on the above considerations, the preset process node sequence can be determined through a preset mapping table.

[0039] Specifically, each preset process node in the preset process node sequence includes process type information and estimated operation time; the process type information is used to indicate that the node is one of the above-mentioned CAD design, 3D printing, grinding / finishing, glazing / dyeing, sintering nodes, etc.

[0040] Step 302: Obtain multi-dimensional features of the denture manufacturing order based on the order identifier, order type, user identifier, order delivery time, order amount, and corresponding preset process node sequence; the multi-dimensional features include one or more of the following: delivery urgency feature, customer level value feature, material scarcity index feature, process complexity score feature, quality risk level feature, equipment health status dependence feature, and order economic value density feature; As one embodiment, step 302 specifically includes the following steps 401 to 408: Step 401: Obtain the delivery urgency based on the order delivery time and the current time; Specifically, the calculation is performed according to the following formula: x_{i1}=1 / (T_{due,i}-T_{now}+ε); Where x_{i1} represents the urgency of the delivery of the i-th order, T_{due,i} represents the promised delivery time of the i-th order, and T_{now} represents the current time; ε is a very small positive number (such as 1e-5) to prevent the denominator from being zero or negative (when the order has expired).

[0041] Specifically, the less urgent the delivery date (the closer to the delivery date), the higher the priority score will be, showing a non-linear growth relationship, which is consistent with the understanding of "urgent" orders in production.

[0042] Step 402: Obtain the corresponding customer category and customer historical cooperation data based on the user identifier, and obtain the customer level value based on the customer category and customer historical cooperation data; Specifically, by using the user identifier, the corresponding customer category and historical cooperation data can be obtained from the database based on the corresponding storage interface; Specifically, customer categories include: VIP users, regular users, and new users, with different user level weights corresponding to different customer categories; As one example, customer level value is obtained according to the following formula: x_{i2}=w_{vip} δ_{vip}+w_{regular} δ_{regular}+w_{new} δ_{new}+S(C_i); Where x_{i2} represents the customer level value, δ_{vip}, δ_{regular}, and δ_{new} represent indicator functions for YIP users, regular users, and new users, respectively. When the customer category is the corresponding category, the corresponding indicator function is 1, otherwise it is 0; w_{vip}, w_{regular}, and w_{new} represent the preset weights corresponding to the customer categories, which are determined according to business rules. Wherein, S(C_i) represents the customer's historical value, which is obtained based on the customer's historical data. The customer's historical data includes one or more of the following: average annual order amount, payment on time rate, and years of cooperation. It can be obtained by weighted summation. For example, S(C_i) is obtained through the following formula: S(C_i)=θ1 (Average annual order amount) + θ2 (Payment timeliness rate) + θ3 (Term of cooperation); Wherein, θ1, θ2 and θ3 represent the weights corresponding to the average annual order amount, timely payment rate and cooperation period, respectively.

[0043] Specifically, customer level value combines classification weights with continuous value scores, transforming customer importance from a simple label into a continuous value score, providing more refined differentiation for scheduling.

[0044] Step 403: Obtain the corresponding material parameters based on the denture material, and obtain the material scarcity attributes based on the material parameters; wherein, the material parameters include the current inventory, safety stock threshold, current market price and base price, and the current delivery date and average historical delivery date of the material obtained from the supplier; Specifically, order types include tooth position information, tooth size, denture type, and denture material; Specifically, based on the material of the denture, a request is sent to the resource pool management system to obtain real-time material parameters, including inventory factors, price factors, and supply delivery date factors; As one implementation method, the scarcity property of materials is obtained through the following formula: x_{i3}=(1-I_{current} / I_{safe}) α1+(P_{market} / P_{base}-1) β1+(L_{lead} / L_{avg}) γ1; Where (1-I_{current} / I_{safe}) represents the inventory factor, (P_{market} / P_{base}-1) represents the price factor, (L_{lead} / L_{avg}) represents the supply delivery time factor, and α1, β1, γ1 are the weight coefficients of the inventory factor, price factor, and supply delivery time factor, respectively, satisfying α1+β1+γ1=1.

[0045] Wherein, I_{current} represents the current inventory of the materials required for the order in the online side warehouse, I_{safe} represents the safety stock threshold for the material, P_{market} represents the current market price of the material, P_{base} represents the current base price of the material, L_{lead} represents the current lead time of the material obtained from the supplier, and L_{avg} represents the average historical lead time of the material obtained from the supplier. All this information is stored in the database.

[0046] Specifically, the scarcity attribute of a material can be dynamically quantified by real-time inventory levels, market supply fluctuations, and procurement difficulty, thus assessing the ease and risk of acquiring that material. The higher the index, the scarcer the material.

[0047] Step 404: Obtain the standard deviation of working time and the average working time of each preset process node according to the preset process node sequence, and obtain the process complexity score based on the standard deviation of working time and the average working time. Specifically, each order type also includes tooth position information and tooth size. Based on the tooth position information, tooth size, denture type, and denture material, a baseline complexity can be obtained to quantify the complexity of that type of order. Specifically, for each preset process node in the preset process node sequence, the standard deviation and average working hours of the working hours are updated in real time in the resource pool database based on historical data.

[0048] As one implementation method, the process complexity score can be obtained using the following formula: x_{i4}=BaseComplexity(OrderType_i)+Σ_{step}Time_{std}(step) / Time_{avg}(step); Where x_{i4} represents the process complexity score, OrderType_i represents the order type, BaseComplexity(OrderType_i) represents the preset baseline complexity corresponding to the order type, Time_{std}(step) represents the standard deviation of the time for each preset process node in historical data, and Time_{avg}(step) represents the average time for the corresponding preset process node.

[0049] Specifically, the baseline score reflects the overall difficulty, while the coefficient of variation of working hours (standard deviation / mean) reflects the uncertainty of the order's performance in production. The higher the uncertainty, the larger the buffer that needs to be reserved during scheduling, and the higher the complexity score.

[0050] Step 405: Obtain a preset quality risk coefficient based on the denture material, a preset process risk coefficient based on the preset process nodes, and a preset severity coefficient based on the denture type; obtain the quality risk level based on the quality risk coefficient, process risk coefficient, and severity coefficient. Specifically, the quality risk coefficient is a preset value based on the denture material; it represents the probability that dentures made of that material may have quality problems; the severity coefficient is a preset value obtained based on the denture type; it represents the probability that factors related to that denture type may cause quality problems; and the process risk coefficient is the degree of quality risk that a preset process node may cause due to the manufacturing process.

[0051] Specifically, the quality risk level is obtained according to the following formula: x_{i5}=max(R_{material},R_{process}) Severity(OrderType_i); Where x_{i5} represents the quality risk coefficient, R_{material} represents the quality risk coefficient of the material used (e.g., resin shrinkage risk is higher than zirconia, preset based on experience, obtained from the corresponding mapping table). R_{process} represents the risk coefficient of the core process type step (e.g., sintering process risk is higher than cutting process, preset based on experience, obtained from the corresponding mapping table). Severity(OrderType_i) represents the severity coefficient of the order type (e.g., the cost of failed anterior aesthetic restoration is higher than that of posterior restoration).

[0052] Step 406: Obtain the corresponding equipment set according to the preset process node, and obtain the equipment health status dependency based on the equipment set.

[0053] Specifically, it can be obtained using the following formula: x_{i6}=1-min_{jinEligibleResources}(HealthScore_j); Where x_{i6} represents the equipment health status dependency, and EligibleResources represents the set of equipment processing this order process. (HealthScore_j) represents the real-time health score of the j-th device, which is a comprehensive value obtained based on IoT data such as vibration, temperature, and spindle accuracy, ranging from [0,1]).

[0054] Specifically, the real-time health score is obtained through the interfaces of each device. It is a comprehensive value obtained by the controller of each device based on its own IoT data such as vibration, temperature, and spindle accuracy, and ranges from [0,1].

[0055] Step 407: Obtain the total processing time based on the denture type and material, and obtain the order economic benefits based on the order amount and total processing time; Specifically, for each process node, the preset processing time for each process node is obtained through a preset process processing time mapping table based on tooth position information, denture type, and denture material; the total processing time is obtained based on the preset processing time corresponding to each preset process node.

[0056] Specifically, each preset process node includes a corresponding preset process processing time mapping table. The preset process processing time mapping table is used to indicate the estimated processing time for that process node to complete the corresponding tooth position information, tooth size, denture type and denture material.

[0057] Specifically, the economic benefits of an order are obtained using the following formula: x_{i7}=(OrderValue_i) / (EstimatedProcessTime_i); Where x_{i7} represents the economic benefit of the order, OrderValue_i represents the order amount, and EstimatedProcessTime_i is the total processing time.

[0058] Step 408: Normalize the delivery urgency, customer level value, material scarcity index, process complexity score, quality risk level, equipment health status dependence, and order economic value density to obtain the delivery urgency features, customer level value features, material scarcity index features, process complexity score features, quality risk level features, equipment health status dependence features, and order economic value density features. Specifically, the original feature values ​​x_{ij} of the above dimensions have different dimensions and ranges. Max-min normalization or Z-score standardization is required to map them to a uniform interval (e.g., [0,1]) to ensure the fair application of the weights.

[0059] Specifically, the following processes are performed: normalizing the delivery urgency to obtain delivery urgency characteristics; normalizing the customer level value to obtain customer level value characteristics; normalizing the material scarcity index to obtain material scarcity index characteristics; normalizing the process complexity score to obtain process complexity score characteristics; normalizing the quality risk level to obtain quality risk level characteristics; normalizing the equipment health status dependence to obtain equipment health status dependence characteristics; and normalizing the order economic value density to obtain order economic value density characteristics.

[0060] Step 303: Based on the reinforcement learning model, perform weight learning on the multi-dimensional features to obtain the weights corresponding to the multi-dimensional features; Specifically, the reinforcement learning model is a reinforcement learning-driven weight generation model, comprising an environment interaction module, a deep neural network module, and a policy learning module; among which: The environment interaction module is used to obtain real-time data of resources in the resource pool and obtain statistical features G (Queue), load rate vector H (Resource), order scheduling effect M (History) and reward R_t based on the real-time data; Statistical features G (Queue), load rate vector H (Resource), order scheduling effect M (History), and multi-dimensional features are used to form the current state S_t of the deep neural network module; The deep neural network module outputs action A_t based on the current input state S_t. Action A_t includes a weight vector composed of the weights corresponding to each multi-dimensional feature. The policy learning module updates the network parameters of the deep neural network module based on the current state S_t, action A_t, and reward R_t.

[0061] Furthermore, the state S_t is defined as: S_t=[G(Queue),H(Resource),M(History),X(PendingOrders)] Wherein, G(Queue) represents the statistical characteristics in the current batch order set Queue, such as the statistical characteristic function of the order quantity distribution and the statistical characteristic function of the type distribution. H(Resource) is a vector consisting of the real-time load rate of the equipment corresponding to each preset process node; M (History) represents the order scheduling effect within a preset historical period, which is obtained and updated in real time based on data from the resource pool (such as the average latency rate of the past hour, or the overall equipment utilization rate, U).

[0062] X(PendingOrders) is a vector composed of the multidimensional features obtained in the above steps; Specifically, the statistical characteristics of the Queue, the real-time load rate of the corresponding devices, and the order scheduling effect are all obtained in real time based on the resource pool data, and corresponding functions can be set to perform the calculations. For example, the real-time load rate can be obtained by monitoring the operational status of resources in each resource pool: For resource pools with process-type structures (such as AI design pool, cutting equipment pool, and sintering furnace pool), continuously track the number of resources currently executing tasks (i.e., the number of busy resources) and the total number of resources configured in that resource pool (such as the number of equipment units and workstations). The formula for calculating the real-time load rate is: Load rate = Number of currently busy resources / Total number of resources in the resource pool; This value ranges from 0 to 1, where 0 indicates all are idle and 1 indicates all are busy. For example, if there are 5 cutting machines in the pool and 3 are currently processing dentures, then the real-time load factor is 3 / 5 = 0.6.

[0063] M() represents the scheduling performance over a past period, transformed into a set of numerical vectors that the model can process. The goal is to enable the reinforcement learning model to perceive the effectiveness of the current scheduling strategy and make adjustments in the next decision.

[0064] First, a fixed time window needs to be defined, such as "the past 1 hour" or "the past 2 hours". The length of this window can be set according to the actual scheduling frequency and manufacturing rhythm.

[0065] The system extracts all completed processes and order data within the time window from the scheduling log and calculates the following core metrics: Average order delay rate D: The number of orders whose actual completion time exceeds the order delivery time within the statistical time window, divided by the total number of orders completed within that window. This metric reflects the on-time delivery rate of orders. For example, if 20 orders were completed in the past hour, and 5 of them were delayed, then the average delay rate is 5 / 20 = 0.25.

[0066] Equipment utilization rate U: The total actual processing time of all major equipment (such as cutting equipment, sintering furnaces, grinding stations, etc.) within the statistical time window, divided by the total available time of these equipment. This indicator reflects the efficiency of production resource utilization. For example, if in the past hour, 5 cutting machines theoretically had a usable time of 5 hours, but the actual processing time was 4 hours, then the equipment utilization rate is 4 / 5 = 0.8.

[0067] All calculated performance indicators are concatenated into a numerical vector in a fixed order, which serves as the output of the M(·) function. For example: M(History) = [D, U]; If the average delay rate D is too high, it means that the current strategy is causing too many order delays. The model will learn to adjust the weights so that subsequent decisions are more biased towards urgent orders. If the overall equipment utilization rate U is too low, it means that there are too many idle resources. The model will try to make orders occupy resources more compactly.

[0068] Furthermore, the action A_t is defined as: A_t=W_t=[w_{t1},w_{t2},...,w_{tn}]; Where w_{t1}, w_{t2}, ..., w_{tn} represent the weights of each feature in X(PendingOrders), and the sum of all weights is defined as 1.

[0069] The deep neural network module π_θ is used to implement A_t=π_θ(S_t).

[0070] The reward is defined as a reward calculated based on real-time data from the resource pool over a preset historical period, using the following formula: R_t =-[α3·(C_{actual}-C_{target})+β3·D_{actual}+γ3·(1-Q_{actual})] Where C_{actual}, D_{actual}, and Q_{actual} are the actual total cost, order delay rate, and average quality pass rate in the past, respectively, and α3, β3, and γ3 are the weights of (C_{actual}-C_{target}), D_{actual}, and (1-Q_{actual}), respectively, and α3+β3+γ3=1.

[0071] Specifically, the Monte Carlo policy gradient (REINFORCE) algorithm is used to update the network parameters θ, with the goal of maximizing the expected cumulative reward. The specific update formula is as follows: θ←θ+η·∇_θlogπ_θ(A_t|S_t)·R_t; This formula establishes a strict mathematical relationship between S_t (state), A_t (action / weight), and R_t (reward), driving the model to learn a strategy for selecting the optimal weight action A_t under a specific production state S_t.

[0072] Step 304: Obtain priority scores based on multidimensional features and corresponding weights, and sort the orders from high to low based on the priority scores to obtain an order priority set queue; where the higher the priority score, the higher the urgency of completing the corresponding denture manufacturing order. Specifically, the priority score is obtained by weighted summation of multidimensional features and their corresponding weights.

[0073] In another exemplary embodiment of this application, such as Figure 4 As shown, step 203 includes the following steps 501-503: Step 501: Obtain the preset process node priority sequence based on the preset process node sequence and the order priority queue; wherein, the preset process node priority sequence includes the preset process nodes corresponding to all denture manufacturing orders; Specifically, firstly, taking a dental prosthesis manufacturing order as a unit, all preset process nodes are arranged according to the order priority queue. For the same dental prosthesis manufacturing order, the preset process nodes are arranged according to the order in the corresponding preset process node sequence. For example: The priority sequence of a certain order is {order 1, order 2, order 3}; The preset process node sequence corresponding to Order 1 is {Process 1, Process 2, Process 3}; The preset process node sequence corresponding to Order 2 is {Process 1, Process 2, Process 4}; The preset process node sequence corresponding to order 3 is {process 1, process 2}; Based on this step, the preset process node priority sequence obtained is {1_process1, 1_process2, 1_process3, 2_process1, 2_process2, 2_process4, 3_process1, 3_process2}; Among them, "1_", "2_" and "3_" are used to indicate order 1, order 2 and order 3 respectively, and are generally referred to by the corresponding order identifier.

[0074] In this context, the same identifier in each preset process node sequence represents the same process. For example, "Process 1" represents a CAD design node. Generally, preset process nodes of the same type are identified by the same preset process identifier.

[0075] Step 502: Obtain the process requirements information for each preset process node based on the order information; The order type also includes tooth position information, tooth size, denture type and denture material. The required parameters for each preset process node can be obtained according to the corresponding mapping table. Specifically, order types include tooth position information, tooth size, denture type, and denture material. Denture types include primary and secondary categories. Specifically, for each order, the parameters are determined based on tooth position information, tooth size, denture type and material, or the requirements of each preset process node. As one embodiment, depending on the corresponding preset process, if the preset process node sequence includes a corresponding preset process node, the required parameters of the corresponding preset process node can be obtained according to the order information. Each preset process node includes one or more of the following required parameters: Requirements for CAD design nodes: one or more types of software used, and one or more skills possessed by the designer; The software types include one or more such as EXO and 3shape, and the skills possessed by the designers include one or more such as dental anatomy knowledge, materials science knowledge, aesthetic design and simulation restoration design. Requirements for machining nodes: material type, material quantity, number of machining axes, and cutting accuracy; Specifically, the material type corresponds to the denture material, the material usage is obtained by calculation or estimation based on tooth position information and denture size information, or by preset based on range values, and the number of machining axes and cutting accuracy are obtained based on tooth position information, denture size information and denture type (for example, by looking up a preset mapping table).

[0076] Requirements for 3D printing nodes: printing space, material type, material usage, and supported file formats; Specifically, the material type corresponds to the denture material, the material usage is obtained by calculation or estimation based on tooth position information and denture size information, or by preset based on a range value, and the supported file formats are obtained based on the denture type and denture material (e.g., a preset mapping table, obtained by looking up the table).

[0077] Requirements for grinding / finishing nodes: skill level requirements for processing personnel / quality requirements that processing personnel should achieve in their work / grinding / finishing skills that processing personnel need to possess.

[0078] Specifically, for example, the required proficiency level of processing personnel is reflected through levels (e.g., low, medium, and high), indicating the quality of the processing work completed by the personnel in previous operations. The proficiency level of each processing personnel is obtained based on their usual performance evaluations. The required proficiency level of processing personnel is based on the denture size information, denture type, and denture material. The grinding / finishing skills that processing personnel need to possess include one or more of the following: rough grinding, shaping, fine grinding, and polishing. Requirements for ceramic bonding nodes: supported equipment models, ceramic bonding type, ceramic powder usage, and sintering program type; Specifically, the supported equipment models, porcelain types, and porcelain powder usage are all obtained through the denture size and denture type (e.g., a preset mapping table, obtained by looking up the table). The porcelain program types include different program numbers, which are preset according to the porcelain powder type and denture type (e.g., a preset mapping table, obtained by looking up the table). Each program number corresponds to preset program parameters such as porcelain temperature, heating rate, and drying time. Requirements for glazing and dyeing nodes: skills required of operators; Specifically, the required skills include one or more of the following: glazing, porcelain layering, simulated aesthetic restoration, color matching, and shaping of porcelain. Specifically, the process includes: glazing: applying a glaze layer to the denture surface and sintering it to achieve a smooth, natural color and biocompatibility. porcelain layering: layering porcelain powder onto the inner crown of the denture to restore its basic color and shape. Aesthetic restoration: using personalized staining and texture depiction techniques in the anterior aesthetic zone to simulate the color, translucency, and texture characteristics of natural teeth. Color matching assistance for clinicians: adjusting the denture color based on clinically provided color photos or shade guides to ensure harmony with adjacent teeth. Refining the porcelain body shape: fine-tuning the shape of the porcelain layer before or after glazing to ensure a natural denture shape and accurate occlusal contact.

[0079] Requirements for sintering nodes: consumable type, program type; Specifically, consumable types include silicon molybdenum rods or silicon carbide rods; program types include one or more of the following: maximum temperature and heating rate; program types include different program numbers used to indicate different program parameters (such as different sintering curves, maximum temperature, heating rate, vacuum degree, etc.).

[0080] Quality inspection node requirements parameters: quality inspection capability; Specifically, AI quality inspection capabilities include one or more of the following: process inspection capability, finished product verification capability, rework judgment capability, tool usage capability, and mastery of industry standards (such as ISO13485). AI quality inspection capabilities include one or more of the following: data inspection capabilities, visual inspection capabilities, image scanning and imaging capabilities, automatic report generation capabilities, and inspection speed capabilities.

[0081] Step 503: Build an MDCN decision tree using process type as internal nodes, resources corresponding to the process type as leaf nodes, and resource capability parameters as attributes of the leaf nodes. Input the preset process node sequence, estimated operation time, and process requirement information into the MDCN decision tree. Based on preset constraint rules and input items, traverse the MDCN decision tree to obtain a set of feasible solutions. The set of feasible solutions includes obtaining multiple feasible solutions, and each feasible solution includes the resources matched to each preset node and the operation time window of the resources. Specifically, MDCN (Multi-level Decision Tree) is a multi-level decision tree used here to match resources for process tasks under multiple constraints and generate feasible scheduling solutions.

[0082] As one example, the internal nodes correspond to the process types of the preset process types, including CAD design type, 3D printing type, machining type, porcelain baking type, polishing / finishing type, glazing / staining type and sintering type; As another embodiment, the internal nodes also include manual quality inspection type and AI quality inspection type; Specifically, the preset process nodes are instances corresponding to the internal types. For example, if an order includes a CAD design node, it means that the process type of that node is the CAD design type. Specifically, each node in the preset process node sequence includes an identifier for the corresponding process type and a corresponding estimated operation time, both of which are obtained according to step 301; Specifically, the resources corresponding to each process type are used as leaf nodes of that process type, and the capability parameters of each resource are used as attributes of that leaf node; the capability parameters of the process and the requirement parameters obtained in step 502 are in one-to-one correspondence. Specifically, the resources corresponding to different process types can be equipment or personnel, both of which are obtained in real time through the resource pool; for example, the resource corresponding to CAD design is personnel, and the corresponding resources can be designer 001, designer 002, etc.; for example, the resource corresponding to cutting processing is equipment, and the corresponding resources are cutting equipment 001, cutting equipment 002, etc.; each resource is identified by a corresponding resource identifier. Furthermore, each resource also includes the remaining available time window obtained in real time from the resource pool.

[0083] The preset constraint rules include: Resource matching rules: For each preset process node, the resource parameters and the required parameters can be matched; Sequence constraint rule: For each denture manufacturing order, the end time of each preset process node in the preset process node sequence is later than the start time of the previous preset process node. First time constraint rule: For each resource, the corresponding job duration must be less than the duration of the corresponding available time window; The second time constraint rule is: for each denture manufacturing order, the time to complete the last preset process node is less than the estimated completion time of the denture manufacturing order; Furthermore, the preset constraint rules also include process interruptibility constraints, interruption recovery constraints, and resource preemption constraints; such as Figure 5 This is a flowchart based on process interruptibility constraints, interruption recovery constraints, and resource preemption constraints in this application.

[0084] Process interruptibility constraint: For preset process nodes that allow interruption, the operation time window can be divided into multiple non-overlapping sub-windows; Interruption recovery constraint: The duration of the operation after interruption recovery includes a preset skill recovery compensation time; Resource preemption constraint: When a high-priority order with a priority score above the threshold arrives, the interruptible process of the low-priority order is allowed to be interrupted and the corresponding resources are released.

[0085] Optionally, the estimated completion time of the order = the order delivery time + the preset allowed delay time, or: the estimated completion time of the order = the order delivery time. (1 + delay coefficient).

[0086] As one example, the list of interruptible processes includes: CAD design nodes, grinding / finishing nodes, glazing / dyeing nodes, and manual quality inspection nodes; the list of non-interrupted processes includes: 3D printing nodes, machining nodes, porcelain baking nodes, sintering nodes, and AI quality inspection nodes. The recommended value for skill recovery compensation time is 5%-15% of the remaining time of the original process; the specific value can be calibrated based on historical production data.

[0087] As one example, to ensure a one-to-one correspondence between parameter requirements and resource parameters, the resource parameters for different process types include: Resource parameters for CAD design nodes: one or more types of software used, and one or more skills possessed by the designer; The software types include one or more such as EXO and 3shape, and the skills possessed by the designers include one or more such as dental anatomy knowledge, materials science knowledge, aesthetic design and simulation restoration design. Resource parameters for machining nodes: material type, material allowance, number of machining axes, and cutting accuracy; Resource parameters for 3D printing nodes: printing space, material type, material allowance, and supported file formats; Resource parameters for polishing / refining nodes: the required proficiency level of the processing personnel and the polishing / refining skills that the processing personnel need to possess.

[0088] Resource parameters for the ceramic firing node: supported equipment models, ceramic firing type, ceramic powder usage, and sintering program type; each sintering program type corresponds to preset ceramic firing temperature, heating rate, drying time, and other program parameters; Resource parameters for glazing and staining nodes: skills required for operators; including one or more of the following: glazing, porcelain stacking, simulated aesthetic restoration, color matching, and shaping of porcelain body; Resource parameters of sintering nodes: consumable type and program type; the program type is used to indicate different program parameters (such as different sintering curves, maximum temperature, heating rate, vacuum degree, etc.).

[0089] Resource parameters for quality inspection nodes: quality inspection capacity; Specifically, AI quality inspection capabilities include one or more of the following: process inspection capability, finished product verification capability, rework judgment capability, tool usage capability, and mastery of industry standards (such as ISO13485). AI quality inspection capabilities include one or more of the following: data inspection capabilities, visual inspection capabilities, image scanning and imaging capabilities, automatic report generation capabilities, and inspection speed capabilities.

[0090] Furthermore, the method of this application also includes a method for rework risk scheduling, specifically including: obtaining the single-batch rework probability of each preset process node based on historical rework data statistics; calculating the total expected rework cost for each feasible solution in the set of feasible solutions, wherein the total expected rework cost = Σ(rework probability of each preset process node × rework time of that node × corresponding resource unit time cost + rework material loss cost); when an actual rework event is detected, immediately triggering incremental fast rescheduling based on the current order priority set queue, and prioritizing the rework process as a high-priority task by inserting it into the idle window of the current scheduling sequence or preempting interruptible resources.

[0091] As one example, such as Figure 6 The specific steps for implementing rework risk scheduling are as follows: Step 601: Establish a historical rework database to record the number of reworks, reasons for reworks, rework time, and material losses for each process node; Step 602: Update the single-batch rework probability P(step) for each process node quarterly = total historical rework batches of that node / total historical production batches of that node; Step 603: Calculate the total expected rework cost for each feasible solution C_rework=Σ[P(step)×(T_rework(step)×C_hour(step)+C_material(step))], where T_rework(step) is the average rework time for that node, C_hour(step) is the unit labor cost of resources for that node, and C_material(step) is the average rework material loss cost for that node; Step 604: Add the total expected rework cost as the fourth evaluation dimension to the comprehensive scoring formula, with a suggested weight of 0.1-0.2; Step 605: When the quality inspection node detects an actual rework event, it immediately generates a rework task and assigns it a higher priority than ordinary orders. Step 606 triggers incremental fast rescheduling, which only adjusts the currently unstarted and interruptible processes, while retaining the scheduling results of the started and uninterruptible processes. Step 607: Prioritize inserting rework tasks into the idle window of the current scheduling sequence. If there is no suitable idle window, preempt the interruptible resources of low-priority orders.

[0092] In another exemplary embodiment of this application, step 204 specifically includes the following steps: For each feasible solution, obtain the comprehensive delay index, comprehensive utilization rate, and comprehensive resource utilization rate of the feasible solution; obtain a comprehensive score based on the comprehensive delay index, comprehensive utilization rate, and comprehensive resource utilization rate; select the feasible solution with the highest comprehensive score as the target solution.

[0093] Specifically, for each order in a feasible solution, obtain the weighted delay metric for each order, and then obtain the weighted delay metric for each feasible solution based on the weighted delay metric for each order. Furthermore, the average of the weighted delay metrics of all orders can be taken as the weighted delay metric of the feasible solution; alternatively, the weighted average can be selected as the weighted delay metric of the feasible solution. Where, delay duration = end time - delivery time, and weighted delay index = Σ(order weight × delay duration), where order weight can be determined based on... Specifically, for each feasible solution, the overall utilization rate is obtained by comparing the sum of the busy times of all resources with the sum of the available times of all resources. The total available time of all resources = number of resources × total duration of scheduling cycle; Specifically, in each feasible solution, all idle time periods of all resources are obtained, and the minimum duration of the idle time period or the average duration of all idle time periods is taken as the comprehensive utilization rate.

[0094] Furthermore, as evaluation indicators, the larger the resource redundancy R_i, the smaller the weighted delay index D_i, and the larger the comprehensive utilization rate U_i, the better the performance of the corresponding feasible solution. Since the three indicators have different dimensions, they need to be unified to the same scale (e.g., [0,1]). Assume that in the set of all feasible solutions: R_max and R_min are the maximum and minimum values ​​of resource redundancy in all solutions, respectively; D_max and D_min are the maximum and minimum values ​​of weighted delay in all solutions, respectively; and U_max and U_min are the maximum and minimum values ​​of comprehensive utilization rate in all solutions, respectively.

[0095] Furthermore, the corresponding standardized formula is: For the positive indicators R and U: R_i' = R_i - R_min / R_max - R_min; U_i' = U_i - U_min / U_max - U_min; For the negative indicator (D): D_i' = D_max - D_i / D_max - D_min; For the negative index (C_rework): C_rework_i'=(C_rework_max-C_rework_i) / (C_rework_max-C_rework_min); Thus, all standardized values ​​fall within the range [0,1], with larger values ​​being better. Based on actual production needs, weights are assigned to the three indicators, for example: resource redundancy weight w_R, weighted delay weight w_D, overall utilization weight w_U, and rework cost weight w_C, satisfying (w_R+w_D+w_U+w_C=1). The comprehensive score for each feasible solution is: Si=w_R R_i'+w_D D_i'+w_U U_i'+w_CC_rework_i'.

[0096] Specifically, the feasible solution with the largest S_i is selected as the final scheduling scheme. If multiple solutions have the same highest score, the scheme that completes earliest is selected.

[0097] This application provides a method for scheduling and controlling the manufacturing of dental prostheses, which includes the following technical effects: By using the denture type and denture material in the order type, the preset process node sequence corresponding to each order is obtained, realizing the standardized association between order attributes and manufacturing process; From basic information such as order identifier, order type, user identifier, order delivery time, and order amount, multi-dimensional features such as delivery urgency, customer level value, material scarcity index, process complexity score, quality risk level, equipment health status dependence, and order economic value density are extracted to form a comprehensive characterization of order value.

[0098] By using reinforcement learning models to learn weights for multidimensional features, the traditional fixed weights or manual weighting methods are changed, enabling priority scores to be automatically adjusted according to real-time changes in the production environment, thus improving the adaptability of scheduling decisions.

[0099] The current state is composed of statistical features, load rate vectors, order scheduling effects, and multi-dimensional features. The deep neural network module outputs weight vectors as actions based on the state, and the policy learning module updates network parameters based on rewards, forming a continuously optimizing learning mechanism. This achieves collaborative optimization of multi-dimensional feature weights, supports continuous feedback and iterative improvement of scheduling effects, and provides a decision-making basis for multi-objective scheduling optimization. Denture manufacturing orders are sorted by order priority scores to form an order priority queue. Based on this, a preset process node priority sequence is generated to ensure that high-urgency orders receive priority processing in production scheduling, thereby improving the timeliness of order delivery and customer satisfaction.

[0100] A decision tree for MDCN is constructed, with process type as internal nodes, resources as leaf nodes, and resource capability parameters as attributes. Taking a preset process node sequence, estimated operation time, and process requirements as input, the decision tree is traversed to match resources that meet the requirements for each process node and generate feasible operation time windows, thereby ensuring accurate matching between process requirements and resource capabilities.

[0101] This includes constraints on resource capacity matching, process sequence, time window, and order delivery date, ensuring that the generated feasible solutions meet actual production requirements in terms of resource capacity, process sequence, time arrangement, and delivery time, and avoiding resource conflicts and process errors.

[0102] The addition of process interruptibility constraints, interruption recovery constraints, and resource preemption constraints enables the scheduling system to flexibly respond to urgent orders, maximizing the use of idle and interruptible resources without affecting the normal production of uninterruptible processes, thereby improving the system's response speed and resource utilization.

[0103] By introducing a rework risk pre-calculation and rapid rescheduling mechanism, rework costs are incorporated into the solution evaluation system in advance, enabling proactive selection of low-risk production paths. When actual rework occurs, production plans are quickly adjusted through incremental rescheduling to minimize the impact of rework on the overall production schedule.

[0104] The three core evaluation dimensions extracted from feasible solutions are comprehensive delay indicators, comprehensive utilization rate, and comprehensive resource utilization rate. These dimensions reflect the solution's delivery assurance capability, overall output efficiency, resource utilization effectiveness, and quality risk cost, respectively, enabling a comprehensive and multi-dimensional quantitative evaluation of feasible solutions.

[0105] In summary, the method of the present invention can effectively address the dynamic and complex scheduling needs in denture manufacturing, and achieve intelligent resource collaboration and production efficiency optimization.

[0106] Based on the same inventive concept, this application also provides a denture manufacturing scheduling control device for implementing the denture manufacturing scheduling control method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more denture manufacturing scheduling control device embodiments provided below can be found in the limitations of the denture manufacturing scheduling control method described above, and will not be repeated here.

[0107] In one exemplary embodiment, such as Figure 7 As shown, a denture manufacturing scheduling and control device is provided, comprising: The acquisition module is used to acquire a batch of orders; the batch of orders includes order information corresponding to several denture manufacturing orders. The processing module is used to obtain the preset process node sequence and priority score based on multi-dimensional features for each denture manufacturing order based on order information; sort the denture manufacturing orders from high to low based on the priority score to obtain the order priority queue; wherein, the preset process node sequence includes multiple preset process nodes; the higher the priority score, the higher the urgency of completing the corresponding denture manufacturing order; The matching module is used for: An MDCN decision tree is established based on the process type and the corresponding resources. Based on the preset rule constraints and the MDCN decision tree, the corresponding resources are matched for the preset process nodes to obtain a set of feasible solutions. The set of feasible solutions includes at least one feasible solution, and each feasible solution includes the resources matched for each preset process node and the operation time of the resources. The set of feasible solutions is optimized and filtered to obtain the target solution, which is one of the feasible solutions.

[0108] As an optional implementation, the matching module is specifically used for: The preset process node priority sequence is obtained based on the preset process node sequence and the order priority queue; wherein, the preset process node priority sequence includes the preset process nodes corresponding to all denture manufacturing orders; Obtain process requirement information for preset process nodes based on order information; An MDCN decision tree is constructed using process type as internal nodes, resources corresponding to the process type as leaf nodes, and resource capability parameters as attributes of the leaf nodes. The preset process node sequence, estimated operation time, and process requirement information are input into the MDCN decision tree. Based on preset constraint rules and input items, the MDCN decision tree is traversed to obtain a set of feasible solutions. The set of feasible solutions includes obtaining multiple feasible solutions, and each feasible solution includes the resources matched to each preset node and the operation time window of the resources.

[0109] As an optional implementation, the preset constraint rules in the matching module include: For each preset process node, the resource parameters and the required parameters can be matched; For each denture manufacturing order, the end time of each preset process node in the preset process node sequence is later than the start time of the previous preset process node. For each resource, the corresponding job duration must be less than the duration of the corresponding available time window; The time to complete the last preset process node for each denture manufacturing order is less than the estimated completion time of the denture manufacturing order.

[0110] As an optional implementation, the preset constraint rules in the matching module also include: Process interruptibility constraints, interruption recovery constraints, and resource preemption constraints; The interruptibility constraint of the process includes: for a preset process node that allows interruption, the operation time window can be divided into multiple non-overlapping sub-windows; The interruption recovery constraint includes: the operation duration after interruption recovery includes a preset skill recovery compensation time; The resource preemption constraint includes: when a high-priority order with a priority score higher than the threshold arrives, the interruptible process of the low-priority order is allowed to be interrupted and the corresponding resources are released.

[0111] As an optional implementation, the order information includes order identifier, order type, user identifier, order delivery time, and order amount; the order type includes denture type and denture material; the processing module specifically includes: For each denture manufacturing order, a preset sequence of process nodes is obtained based on the denture type and material. The multidimensional features of a denture manufacturing order are obtained based on the order identifier, order type, user identifier, order delivery time, order amount, and corresponding preset process node sequence. Among them, the multidimensional features include one or more of the following: delivery urgency feature, customer level value feature, material scarcity index feature, process complexity score feature, quality risk level feature, equipment health status dependence feature, and order economic value density feature. The weights corresponding to multi-dimensional features are obtained by learning weights based on reinforcement learning models; Priority scores are obtained based on multidimensional features and their corresponding weights.

[0112] As an optional implementation, the matching module is further used for: The probability of rework per batch for each preset process node is obtained based on historical rework data statistics. In the set of feasible solutions, the total expected rework cost is calculated for each feasible solution. The total expected rework cost = Σ(rework probability of each preset process node × rework time of that node × corresponding resource unit time cost + rework material loss cost). When an actual rework event is detected, an incremental fast rescheduling is immediately triggered based on the current order priority set queue. The rework process is then inserted as a high-priority task into the idle window of the current scheduling sequence or preempted for interruptible resources.

[0113] As an optional implementation, the processing module is further used for: Determine the urgency of delivery based on the order delivery time and the current time; Obtain the corresponding customer category and customer history cooperation data based on the user identifier, and obtain the customer level value based on the customer category and customer history cooperation data; Based on the denture material, obtain the corresponding material parameters, and based on the material parameters, obtain the material scarcity attributes; among them, the material parameters include the current inventory, safety stock threshold, current market price and base price, and the current delivery date and average historical delivery date of the material obtained from the supplier; The standard deviation and average working time of each preset process node are obtained according to the preset process node sequence, and the process complexity score is obtained according to the standard deviation and average working time. A preset quality risk coefficient is obtained based on the denture material; a preset process risk coefficient is obtained based on the preset process nodes; and a preset severity coefficient is obtained based on the denture type. The quality risk level is then determined based on the quality risk coefficient, process risk coefficient, and severity coefficient. Obtain the corresponding equipment set based on the preset process nodes, and obtain the equipment health status dependency based on the equipment set; The total processing time is determined based on the type and material of the denture, and the economic benefits of the order are determined based on the order amount and the total processing time. Normalize the delivery urgency, customer level value, material scarcity index, process complexity score, quality risk level, equipment health status dependence, and order economic value density to obtain the delivery urgency characteristics, customer level value characteristics, material scarcity index characteristics, process complexity score characteristics, quality risk level characteristics, equipment health status dependence characteristics, and order economic value density characteristics.

[0114] As an optional implementation, the processing module is further used for: The current state S_t of the deep neural network module is composed of statistical features, load rate vector, order scheduling effect and multi-dimensional features. The deep neural network module outputs action A_t based on the current input state S_t. Action A_t includes a weight vector composed of the weights corresponding to each multi-dimensional feature. The policy learning module is used to update the network parameters of the deep neural network module based on the current state S_t, action A_t, and reward R_t.

[0115] As an optional implementation, the matching module is further used for: For each feasible solution, obtain the comprehensive delay index, comprehensive utilization rate, and comprehensive resource utilization rate of the feasible solution; obtain a comprehensive score based on the comprehensive delay index, comprehensive utilization rate, and comprehensive resource utilization rate, and select the feasible solution with the highest comprehensive score as the target solution.

[0116] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 8 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When the computer program is executed by the processor, it implements a denture manufacturing scheduling and control method.

[0117] Those skilled in the art will understand that Figure 8 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0118] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0119] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0120] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0121] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0122] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0123] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0124] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0125] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. In summary, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for scheduling and controlling the manufacturing of dentures, characterized in that, The denture manufacturing scheduling and control method includes: Obtain a batch order set; wherein, the batch order set includes order information corresponding to several denture manufacturing orders; Based on the order information, a preset process node sequence and a priority score based on multi-dimensional features are obtained for each denture manufacturing order; the denture manufacturing orders are sorted from high to low based on the priority scores to obtain an order priority set queue; wherein, the preset process node sequence includes multiple preset process nodes; the higher the priority score, the higher the urgency of completing the corresponding denture manufacturing order; An MDCN decision tree is established based on the process type and the corresponding resources. Based on preset rule constraints and the MDCN decision tree, the corresponding resources are matched for the preset process node to obtain a set of feasible solutions. The set of feasible solutions includes at least one feasible solution, and the feasible solution includes the resources matched for each preset process node and the operation time of the resources. The set of feasible solutions is optimized and filtered to obtain a target solution, which is one of the feasible solutions.

2. The denture manufacturing scheduling and control method according to claim 1, characterized in that, The step of establishing an MDCN decision tree based on process type and corresponding resources, and matching the corresponding resources for the preset process node based on preset rule constraints and the MDCN decision tree to obtain a set of feasible solutions, includes: A preset process node priority sequence is obtained based on the preset process node sequence and the order priority queue; wherein, the preset process node priority sequence includes the preset process nodes corresponding to all the denture manufacturing orders; Based on the order information, obtain the process requirement information of the preset process node; The MDCN decision tree is constructed using process type as internal node, resource corresponding to the process type as leaf node, and capability parameters of the resource as attributes of the leaf node. The preset process node sequence, expected operation time, and process requirement information are input into the MDCN decision tree. The feasible solution set is obtained by traversing the MDCN decision tree based on preset constraint rules and the input items. The feasible solution set includes obtaining multiple feasible solutions, and each feasible solution includes the resource matched to each preset node and the operation time window of the resource.

3. The denture manufacturing scheduling and control method according to claim 2, characterized in that, The preset constraint rules include: For each of the preset process nodes, the resource parameters and the required parameters can be matched; For each denture manufacturing order, the end time of each preset process node in the preset process node sequence is later than the start time of the previous preset process node. For each of the aforementioned resources, the corresponding job duration must be less than the duration of the corresponding available time window; The time to complete the last preset process node for each denture manufacturing order is less than the expected completion time of the denture manufacturing order.

4. The denture manufacturing scheduling and control method according to claim 2, characterized in that, The preset constraint rules also include process interruptibility constraints, interruption recovery constraints, and resource preemption constraints; The interruptibility constraint of the process includes: for a preset process node that allows interruption, the operation time window can be divided into multiple non-overlapping sub-windows; The interruption recovery constraint includes: the operation duration after interruption recovery includes a preset skill recovery compensation time; The resource preemption constraint includes: when a high-priority order with a priority score higher than the threshold arrives, the interruptible process of the low-priority order is allowed to be interrupted and the corresponding resources are released.

5. The denture manufacturing scheduling and control method according to claim 1, characterized in that, The order information includes order identifier, order type, user identifier, order delivery time, and order amount; Order types include denture type and denture material. The process involves obtaining a preset sequence of process nodes and a priority score based on multi-dimensional features for each denture manufacturing order based on the order information; this includes: For each of the aforementioned denture manufacturing orders, a preset sequence of process nodes is obtained based on the denture type and the denture material; The multidimensional features of the denture manufacturing order are obtained based on the order identifier, the order type, the user identifier, the order delivery time, the order amount, and the corresponding preset process node sequence; wherein, the multidimensional features include one or more of the following: delivery urgency feature, customer level value feature, material scarcity index feature, process complexity score feature, quality risk level feature, equipment health status dependence feature, and order economic value density feature. The weights corresponding to the multidimensional features are obtained by learning the weights of the multidimensional features based on the reinforcement learning model. The priority score is obtained based on the multidimensional features and their corresponding weights.

6. The denture manufacturing scheduling and control method according to claim 1, characterized in that, When establishing an MDCN decision tree based on process type and corresponding resources, and matching the corresponding resources for the preset process node based on preset rule constraints and the MDCN decision tree to obtain a set of feasible solutions, the method further includes: Based on historical rework data, the single-batch rework probability of each preset process node is obtained; in the set of feasible solutions, the total expected rework cost is calculated for each feasible solution, and the total expected rework cost = Σ(rework probability of each preset process node × rework time of that node × corresponding resource unit time cost + rework material loss cost). When an actual rework event is detected, an incremental fast rescheduling is immediately triggered based on the current order priority set queue. The rework process is then inserted as a high-priority task into the idle window of the current scheduling sequence or preempted for interruptible resources.

7. The denture manufacturing scheduling and control method according to claim 5, characterized in that, The process of obtaining the multidimensional features of the denture manufacturing order based on the order identifier, the order type, the user identifier, the order delivery time, the order amount, and the corresponding preset process node sequence includes: The delivery urgency is determined based on the order delivery time and the current time. Obtain the corresponding customer category and customer historical cooperation data based on the user identifier, and obtain the customer level value based on the customer category and customer historical cooperation data; Based on the denture material, obtain the corresponding material parameters, and based on the material parameters, obtain the material scarcity attribute; wherein, the material parameters include current inventory, safety stock threshold, current market price and base price, current delivery date and average historical delivery date of the material obtained from the supplier; The standard deviation and average working time of each preset process node are obtained according to the preset process node sequence, and the process complexity score is obtained according to the standard deviation and average working time. A preset quality risk coefficient is obtained based on the denture material; a preset process risk coefficient is obtained based on the preset process nodes; and a preset severity coefficient is obtained based on the denture type. The quality risk level is then determined based on the quality risk coefficient, the process risk coefficient, and the severity coefficient. Obtain the corresponding equipment set according to the preset process node, and obtain the equipment health status dependency based on the equipment set; The total processing time is obtained based on the denture type and denture material, and the economic benefits of the order are obtained based on the order amount and the total processing time. The delivery urgency, customer level value, material scarcity index, process complexity score, quality risk level, equipment health status dependence, and order economic value density are normalized respectively to obtain the delivery urgency feature, customer level value feature, material scarcity index feature, process complexity score feature, quality risk level feature, equipment health status dependence feature, and order economic value density feature.

8. The denture manufacturing scheduling and control method according to claim 7, characterized in that, The reinforcement learning model is a reinforcement learning-driven weight generation model, which includes a deep neural network module and a policy learning module; wherein: The current state S_t of the deep neural network module is composed of statistical features, load rate vector, order scheduling effect, and the multidimensional features. The deep neural network module outputs an action A_t based on the input current state S_t, and the action A_t includes a weight vector composed of the weights corresponding to each of the multidimensional features; The policy learning module is used to update the network parameters of the deep neural network module based on the current state S_t, the action A_t, and the reward R_t.

9. The denture manufacturing scheduling and control method according to claim 7, characterized in that, The step of optimizing and filtering the set of feasible solutions to obtain the target solution includes: For each feasible solution, obtain the comprehensive delay index, comprehensive utilization rate, and comprehensive resource utilization rate of the feasible solution; obtain a comprehensive score based on the comprehensive delay index, comprehensive utilization rate, and comprehensive resource utilization rate, and select the feasible solution with the highest comprehensive score as the target solution.

10. A denture manufacturing scheduling and control device, characterized in that, The denture manufacturing scheduling and control device includes: The acquisition module is used to acquire a batch order set; wherein, the batch order set includes order information corresponding to several denture manufacturing orders; The processing module is used to obtain a preset process node sequence and a priority score based on multi-dimensional features for each denture manufacturing order based on the order information; sort the denture manufacturing orders from high to low based on the priority scores to obtain an order priority set queue; wherein, the preset process node sequence includes multiple preset process nodes; the higher the priority score, the higher the urgency of completing the corresponding denture manufacturing order; The matching module is used for: An MDCN decision tree is established based on the process type and the corresponding resources. Based on preset rule constraints and the MDCN decision tree, the corresponding resources are matched for the preset process node to obtain a set of feasible solutions. The set of feasible solutions includes at least one feasible solution, and the feasible solution includes the resources matched for each preset process node and the operation time of the resources. The set of feasible solutions is optimized and filtered to obtain a target solution, which is one of the feasible solutions.

11. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the steps of the denture manufacturing scheduling control method according to any one of claims 1-9.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the denture manufacturing scheduling control method as described in any one of claims 1-9.

Images