Methods and Systems for Dynamic Optimization Configuration of Intelligent Manufacturing Service Chain
By employing a multi-level linkage response mechanism and a multi-objective optimization algorithm, production anomalies are identified and processed, addressing the lack of flexibility in centralized manufacturing service optimization configuration methods. This enables dynamic optimization configuration of the intelligent manufacturing service chain, improving the response speed and stability of production tasks.
Patent Information
- Application Number
- CN202310956038.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-08-01
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2043-08-01
AI Technical Summary
Existing centralized manufacturing service optimization methods are inflexible, lack autonomous decision-making power, and are unable to cope with the complexity of the manufacturing environment and the delay in handling anomalies, resulting in coordination difficulties and making it difficult to achieve dynamic optimization of the intelligent manufacturing service chain.
A multi-level linkage response mechanism is adopted, which identifies production anomalies through decision tree algorithm, combines multi-objective optimization algorithm and grey relational analysis, performs task matching and optimization configuration in manufacturing service alliance, and uses manufacturing service platform to reconfigure abnormal tasks to ensure rapid response and handling of production anomalies.
It enables timely and accurate responses to production anomalies, enhances the manufacturing service alliance's rapid response and autonomous decision-making capabilities, and ensures the stable and reliable operation of production tasks.
Smart Images

Figure CN117114381B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a method for dynamic optimization configuration of intelligent manufacturing service chains, and more particularly to a method for dynamic optimization configuration of intelligent manufacturing service chains based on production anomaly perception. Background Technology
[0002] In response to the characteristics of large-scale, customized, short R&D cycles, and high quality requirements in the development of complex products such as high-end equipment, service-oriented intelligent manufacturing models such as cloud manufacturing, social manufacturing, and shared manufacturing have been proposed and initially applied, supported by next-generation manufacturing information and management technologies such as the Industrial Internet of Things, cloud computing, and cyber-physical systems. These models aim to virtualize and service-encapsulate massive heterogeneous manufacturing resources and capabilities, and publish them as intelligent manufacturing services on manufacturing cloud platforms, thereby providing personalized production services for diverse manufacturing needs.
[0003] As a key component of the aforementioned intelligent manufacturing models, the optimized configuration of the intelligent manufacturing service chain primarily refers to the process of combining and optimizing intelligent manufacturing services for manufacturing tasks of different granularities. When a complex product task is submitted to the manufacturing cloud platform, the platform decomposes the complex task into sub-tasks of different granularities based on constraints such as the product structure tree and process route, and retrieves and generates a set of candidate manufacturing services that can meet the functional requirements of the sub-tasks. Based on this, and according to the service quality requirements of the service requester, the candidate manufacturing services are combined and optimized to ultimately obtain the optimal manufacturing service combination path (i.e., the intelligent manufacturing service chain) capable of executing the manufacturing task.
[0004] Currently, Professor Tao Fei and colleagues at Beijing University of Aeronautics and Astronautics are using particle swarm optimization and improved genetic algorithms to solve the service composition optimization problem in MGrid that minimizes time and cost. Huang has proposed a service composition optimization model that considers service quality indicators and solved it using a chaotic optimization algorithm. Wang has also used particle swarm optimization to solve the machining resource scheduling problem in manufacturing systems. Furthermore, in application scenarios that achieve optimized configuration of manufacturing service chains, improved ant colony optimization, swarm leader algorithms, imperialist competition algorithms, and social learning optimization have all achieved certain results in their respective fields.
[0005] Although the above methods use centralized strategies and models to obtain optimization results, and consider multiple parameters simultaneously to achieve efficient and high-quality management of manufacturing services, which has greatly promoted the application of service-oriented intelligent manufacturing models, they still face certain challenges in the dynamic optimization and configuration of the intelligent manufacturing service chain.
[0006] In addition, the above research also has the following problems:
[0007] 1) The centralized optimization configuration method transforms the multi-objective optimization problem of service composition selection into a single-objective optimization problem. In fact, it simplifies the complexity of the actual problem selection and lacks the flexibility to cope with changes in actual operation.
[0008] 2) The complexity of the manufacturing environment has not been fully considered, resulting in problems such as delays in handling exceptions and difficulties in coordination, and the service combination and optimization selection cannot be flexibly configured;
[0009] 3) Manufacturing service providers have difficulty having the autonomy to maintain their flexibility and sustainable competitiveness.
[0010] Therefore, there is an urgent need to develop a method and system for dynamic optimization configuration of intelligent manufacturing service chains to solve the above problems. Summary of the Invention
[0011] To address the technical problems of poor flexibility and weak autonomy in existing centralized manufacturing service optimization configuration methods, this invention proposes a dynamic optimization configuration method and system for intelligent manufacturing service chains. In a manufacturing service alliance, a multi-level linkage response mechanism provides more flexible options for service combinations and optimization in production and manufacturing. Multiple manufacturing service providers can be matched as candidate manufacturing service providers based on their own resources and the needs of sub-tasks at different granularities, thus gaining autonomy in decision-making and greater competitiveness.
[0012] The idea behind this invention is:
[0013] First, for anomalies in the production process within the enterprise's manufacturing workshop, a decision tree is constructed to analyze and identify the anomalies, and to further clarify and locate the causes of the anomalies.
[0014] Secondly, for abnormal tasks that cannot be effectively resolved within the enterprise, solutions are sought within the manufacturing service alliance by considering constraints such as service cost, completion cycle, and service quality indicators.
[0015] Then, for thorny anomalies that cannot be resolved in a single manufacturing service alliance, the anomaly tasks are reconfigured across multiple manufacturing service alliances by adding a reputation index for screening, and then grey relational analysis is used to determine the task implementation plan.
[0016] Finally, the Manufacturing Services Alliance uses the Manufacturing Services Platform to find effective solutions for tasks that cannot be completed independently.
[0017] By employing the multi-level linkage approach described above to address production anomalies, we can respond to abnormal events promptly and accurately, and provide effective solutions for production managers. This can greatly enhance the manufacturing service alliance's ability to respond quickly and handle such issues.
[0018] The technical solution adopted by this invention to solve its technical problem is as follows:
[0019] The method for dynamic optimization and configuration of intelligent manufacturing service chains is unique in that it includes the following steps:
[0020] Step 1: Based on the decision tree algorithm, identify internal production anomalies and locate the causes of the anomalies;
[0021] Step 2: Within a single enterprise in the Manufacturing Services Alliance, optimize and match several abnormal production tasks identified in Step 1, and determine whether the optimized and matched task information can meet the delivery time requirements. If yes, synchronize the optimized and matched task information to the designated processing unit; otherwise, proceed to Step 3.
[0022] Step 3: For tasks that cannot meet the delivery time requirements within a single enterprise in the manufacturing service alliance after optimization and matching in Step 2, further optimization and matching are performed among multiple enterprises in a single manufacturing service alliance to obtain a multi-objective optimization scheme. It is then determined whether the service providers in the multi-objective optimization scheme can meet the delivery time requirements. If so, the abnormal task is matched with the screened service providers, and they are assigned to process the corresponding abnormal task; otherwise, proceed to Step 4.
[0023] Step 4: For tasks that still cannot be delivered on time after optimization and matching in a single manufacturing service alliance in Step 3, reconfigure among multiple manufacturing service alliances to obtain the optimal service provider combination scheme. Determine whether the optimal service provider combination scheme can complete the abnormal task. If yes, match the optimal service provider combination scheme with the abnormal task and assign it to the corresponding abnormal production task; otherwise, proceed to Step 5.
[0024] Step 5: For tasks that are still difficult to complete after reconfiguration in Step 4 and have abnormal disturbances, the manufacturing service alliance to which the task belongs shall, as the task issuer, initiate the abnormal production task on the manufacturing service platform for reconfiguration.
[0025] Furthermore, step 1 specifically includes:
[0026] Step 1.1: Based on different production processes, the decision tree algorithm is used to analyze the historical data information of people, machines, materials, methods and environment in the production system. Based on the analysis results, a decision tree for production anomaly extraction and cause localization is obtained. Each path from the root node to the leaf node in the decision tree is the corresponding rule, thus forming a corresponding rule library for anomaly extraction and cause localization.
[0027] Step 1.2: Analyze and compare the real-time task implementation production progress, the deviation value from the initial plan, the processing time of each equipment, and the pass rate of the process with the rules in the anomaly extraction rule base constructed in Step 1.1. Following the goal of timely delivery of production tasks, identify the deviation value between the progress of each production task and the initial plan. When the deviation value between the progress of a production task and the initial plan exceeds the set threshold, it is judged as a production anomaly event.
[0028] Step 1.3 Matches the production anomaly event identified in Step 1.2 with the rules in the anomaly cause localization rule base constructed in Step 1.1 to locate the cause of the production anomaly event and transmit it to the production management system.
[0029] Furthermore, the specific method for optimizing task matching in step 2 is as follows:
[0030] Step 2.1: Parameter control for multi-objective optimization matching;
[0031] Step 2.1.1: Establish the objective function:
[0032] T = min(maxC) max ) = min(maxC i (i∈m)
[0033]
[0034]
[0035] In the formula: T is the completion time; C max C represents the maximum completion time for different tasks. i Let t be the completion time of task i, where i is the task number to be processed, i∈(1,2,...,m), and m is the total number of tasks; W is the machine load; j is the number of machines available within a certain enterprise in the manufacturing service alliance, j∈(1,2,...,n), and n is the total number of machines; k is the k-th subtask of task i, k∈(1,2,...,p), and p is the total number of subtasks; ijk X is the processing time of the k-th subtask of task i on machine j; ijk The value is 1 if the k-th subtask of task i is processed on machine j, and 0 otherwise; C is the processing cost; c ijk Let be the processing cost required to process the k-th subtask of task i on machine j; c1 is the load power of the machine; TG ijk c2 is the machine's end or completion time; c2 is the machine's no-load power.
[0036] Step 2.1.2: Formulating Constraints:
[0037] Constraints on processing time for a single subtask: B ijk +t ijk ≤WG ijk ;
[0038] Sequence constraints between subtasks of the same task: B ijk >WG i(k-i)j ;
[0039] Each subtask can only be processed on one machine at a time.
[0040] Subtasks cannot be interrupted once processing begins: WG ijk -B ijk =t ijk ;
[0041] The completion time of any subtask shall not exceed the maximum completion time: WG ijk ≤C max ;
[0042] In the formula: B ijk WG represents the start time of the k-th subtask of task i on machine j. ijk Let WG represent the completion time of the k-th subtask of task i on machine j; i(k-i)j Let be the processing time of the (k-1)th subtask of task i on machine j;
[0043] Step 2.2: Solve the objective function established in Step 2.1.1 using a multi-objective optimization algorithm;
[0044] Step 2.3: Obtain the optimization scheme for task implementation based on the optimal solution set of the multi-objective optimization algorithm, and perform optimization matching on the task.
[0045] Furthermore, the multi-objective optimization algorithm in step 2.2 is the NSGA-II multi-objective optimization algorithm.
[0046] Furthermore, step 2.3 involves drawing a corresponding task execution Gantt chart based on the optimal solution set of the multi-objective optimization NSGA-II algorithm to obtain the optimization scheme for task implementation.
[0047] Furthermore, the specific method for further optimizing the matching among multiple enterprises in a single manufacturing service alliance in step 3 is as follows:
[0048] Step 3.1: Establish an objective function for selecting service providers within a single manufacturing service consortium:
[0049]
[0050] minT′=min(maxC′ max)=min(maxC′ i (i∈m′)
[0051] Q = L1 + L2
[0052]
[0053]
[0054] In the formula: C′ represents the service cost; i represents the number of the task to be processed, i∈(1,2,...,m′), m′ represents the total number of tasks; j represents the number of the service provider available in the existing manufacturing service alliance, j∈(1,2,...,n′), n′ represents the total number of service providers; k represents the k-th subtask of task i, k∈(1,2,...,p′), p′ represents the total number of subtasks in task i; c′ ijk X represents the service cost of the k-th subtask of task i at service provider j. ijk The value is 1 if the k-th subtask of task i is served by service provider j, and 0 otherwise; minT′ is the minimum completion period; C′ max C′ represents the maximum completion time for different tasks. i L1 represents the completion time of task i; Q represents the service quality; L1 represents the task completion inspection pass rate, used as a process metric; L2 represents the rework rate, used as a result metric; L Q Indicates the number of items that passed inspection upon completion of the task; L T Indicates the total number of inspections required to complete the task; L t Indicates rework hours; L to Represents total production labor hours;
[0055] Step 3.2: Solve the objective function established in Step 3.1 using a multi-objective optimization algorithm to obtain the multi-objective optimization scheme.
[0056] Furthermore, the specific method for reconfiguring among multiple manufacturing service alliances in step 4 is as follows:
[0057] Step 4.1: Establish a model for optimizing the candidate set of manufacturing alliances:
[0058]
[0059]
[0060] In the formula: D represents historical performance; D Y The number of times a contract has been fulfilled on time in history; D T Z represents the total number of historical tasks; Z represents the on-time delivery rate; Z L To ensure on-time delivery; Z T This represents the total number of deliveries.
[0061] Step 4.2: Resource optimization;
[0062] Step 4.2.1: Construct service provider selection model indicators:
[0063] T = T pro +T trs
[0064] C″=C′+C t
[0065]
[0066]
[0067]
[0068]
[0069] In the formula: T is the service time; T pro The time a task is served by a service provider; T trs C represents the transportation time between service providers; C″ represents the service cost; C′ represents the processing cost; C t For transportation costs; j is the ID of the service provider available in the existing manufacturing service alliance, j∈(1,2,...,n′), n′ is the total number of service providers; k is the k-th subtask of task i, k∈(1,2,...,p′), p′ is the total number of subtasks; c′ ijk X represents the service cost of the k-th subtask of task i at service provider j. ijk The value is 1 if the k-th subtask of task i is served by service provider j, and 0 otherwise; d jj′ q represents the distance between different service providers; q represents the unit transportation price; T represents the distance between different service providers. ∝ S represents the total available runtime among service providers within the most recent time α; i For historical demanders' assessment of the sustainability of the manufacturing services alliance; n s The total number of times it was evaluated;
[0070] Step 4.2.2: Grey Relational Analysis
[0071] First, the service provider selection model indicators are normalized. Then, the relatively optimal grey relational degree matrix is calculated. Next, the selection indicator weights are determined. Finally, the maximum value of the grey weighted relational degree is calculated based on the relatively optimal grey relational degree matrix and the selection indicator weights.
[0072] Step 4.2.3: The service resource combination corresponding to the maximum value of the gray weighted correlation degree is taken as the optimal service provider combination scheme.
[0073] The intelligent manufacturing service chain dynamic optimization configuration system is unique in that it includes:
[0074] The enterprise internal production anomaly identification and location module is used to acquire real-time production information through intelligent sensing devices in the manufacturing process, construct the judgment criteria for production anomalies in the manufacturing service alliance based on the decision tree method, analyze and evaluate the key event information in the real-time acquired production tasks, diagnose production anomaly events in a timely manner, locate the cause of the anomaly and provide feedback.
[0075] The abnormal disturbance optimization configuration module within the manufacturing service alliance constructs an optimization configuration mathematical model based on the identification and location results of production abnormality information fed back by the enterprise's internal production abnormality identification and location module, and uses a multi-objective solution algorithm to obtain a better task allocation scheme.
[0076] The inter-manufacturing service alliance anomaly reconfiguration module is used to reconfigure tasks for which the service providers of a single manufacturing service alliance cannot meet the delivery time requirements across multiple manufacturing service alliances.
[0077] The manufacturing service platform conducts demand analysis and optimizes the allocation of tasks among qualified candidate service providers based on the process of candidate selection, task matching, task implementation tracking and monitoring, and task results and evaluation.
[0078] A computer-readable storage medium having a computer program stored thereon, characterized in that the computer program is used to execute the above-described intelligent manufacturing service chain dynamic optimization configuration method when run.
[0079] An electronic device, including a processor and a memory, wherein a computer program is stored in the memory, is characterized in that the computer program is executed by the processor to perform the above-mentioned intelligent manufacturing service chain dynamic optimization configuration method.
[0080] Compared with existing technologies, the advantages of this invention are:
[0081] This invention analyzes the production system and constructs a corresponding decision tree algorithm to promptly diagnose and quickly locate production anomalies caused by the complex and volatile manufacturing environment and frequent disturbances. Then, through a hierarchical linkage between the enterprise manufacturing workshop, a single manufacturing service alliance, multiple manufacturing service alliances, and the manufacturing service platform, it responds to production anomalies, achieving dynamic optimization of intelligent manufacturing services for handling anomalies during production. This invention provides a reference for those handling production anomalies to make autonomous decisions quickly; the proposed hierarchical linkage mechanism between the enterprise manufacturing workshop, a single manufacturing service alliance, multiple manufacturing service alliances, and the manufacturing service platform significantly reduces response time to anomalies, ensuring the stable and reliable operation of the production process. Attached Figure Description
[0082] Figure 1 This is a system architecture diagram of the optimized configuration method of the present invention.
[0083] Figure 2 This is the production anomaly analysis module used in the manufacturing workshop of an enterprise in this invention.
[0084] Figure 3 This is the dynamic optimization module for anomaly-oriented manufacturing service alliance in this invention.
[0085] Figure 4 This is the abnormal disturbance reconfiguration module between manufacturing service alliances in this invention.
[0086] Figure 5 This is a flowchart of the algorithm execution of the optimized configuration method of the present invention. Detailed Implementation
[0087] The present invention will now be described in further detail with reference to the accompanying drawings.
[0088] The purpose of the intelligent manufacturing service chain dynamic optimization configuration method of the present invention is to better assist production managers in enterprises to obtain production anomaly information in a timely and accurate manner, and to provide them with solution references.
[0089] First, a decision tree algorithm is constructed to analyze and identify production anomalies and quickly locate the cause of the anomalies.
[0090] Then, an optimal configuration mathematical model is constructed by constraining various indicators (service cost, completion cycle and service quality indicators). A multi-objective solution algorithm is used to solve the constructed optimal configuration mathematical model to obtain a better task allocation scheme in a single manufacturing service alliance.
[0091] Furthermore, based on the aforementioned task allocation scheme for dynamic optimization of production anomalies within a single alliance, for some tasks that still cannot meet the requirements, a reconfiguration solution under abnormal disturbances is provided through rapid response among manufacturing service alliances.
[0092] Finally, for some production anomaly tasks that are still difficult to complete, the Manufacturing Services Alliance, as the initiator, publishes the task on the Manufacturing Services Platform. The platform conducts a comprehensive requirements analysis on the received tasks, further clarifies the specific requirements of the tasks, and completes the allocation and implementation of the tasks.
[0093] The above-mentioned multi-level linkage solution enables optimized configuration for production anomalies.
[0094] like Figure 1 As shown, the intelligent manufacturing service chain dynamic optimization configuration method provided by the present invention specifically includes the following steps:
[0095] Step 1: Identify and locate internal production anomalies within the enterprise:
[0096] The analysis of internal production performance in this invention mainly includes analysis based on the production system, identification of production anomalies, and classification of production anomalies; the identification and location of internal production anomalies can be achieved by constructing a decision tree.
[0097] Step 1.1: Analysis based on internal production anomalies:
[0098] Based on existing manufacturing process information and different production processes, decision tree construction algorithms (such as C4.5 algorithm / CART algorithm) are used to analyze the historical production status data of production factors related to people, machines, materials, methods and environment in the production system. Based on the analysis results of historical data, a decision tree for production anomaly extraction and cause localization is obtained. Each path from the root node to the leaf node in the decision tree is the corresponding rule, thus forming a corresponding rule library for anomaly extraction and cause localization.
[0099] Step 1.2: Identification of Production Anomalies:
[0100] By sensing the production process through intelligent sensing devices, the production performance data information such as the real-time task implementation progress, the deviation value from the initial plan, the processing time of each device, and the pass rate of the process can be analyzed and compared with the matching of the rules in the anomaly extraction rule base constructed in step 1.1. Following the goal of timely delivery of production tasks, the deviation value between the progress of each production task and the initial plan is identified. When the deviation value between the progress of a production task and the initial plan is greater than the maximum acceptable offset value, it is judged as a production anomaly event.
[0101] Step 1.3: Locating the cause of production anomalies and transmitting production anomaly information:
[0102] The production anomaly event identified in step 1.2 is matched with the rules in the anomaly cause localization rule base constructed in step 1.1 to locate the cause of the production anomaly event. The production anomaly information is then promptly transmitted to the production management system to achieve collaborative information sharing and provide technical support for production managers to make rapid responses and decisions regarding production anomalies.
[0103] Step Two: Real-time acquisition, identification, and location of abnormal event information in production and manufacturing, and synchronization with the Manufacturing Service Alliance for global manufacturing resource information sharing. First, within a single enterprise within the Manufacturing Service Alliance, several abnormal production tasks resulting from abnormal disturbances are optimized and matched:
[0104] Step 2.1: Parameter control for multi-objective optimization matching, which consists of the following two steps:
[0105] Step 2.1.1 Establish the objective function
[0106] 1) Completion time: T = min(maxC) max ) = min(maxC i (i∈m)
[0107] In the formula: C max Indicates the maximum completion time for different tasks; C i Let represent the completion time of task i, where i represents the number of the task to be processed, i∈(1,2,...,m), and m represents the total number of tasks.
[0108] 2) Machine load:
[0109] In the formula: j represents the number of machines that a company within the manufacturing service alliance can provide, j∈(1,2,...,n), where n represents the total number of machines; k represents the k-th subtask of task i, k∈(1,2,...,p), where p represents the total number of subtasks; t ijk X represents the processing time of the k-th subtask of task i on machine j; ijk The value is 1 if the k-th subtask of task i is processed on machine j, and 0 otherwise.
[0110] 3) Processing cost:
[0111]
[0112] In the formula: c ijk c1 represents the processing cost required to process the k-th subtask of task i on machine j; c1 represents the load power of the machine; TG ijkc1 indicates the end or completion time of the machine; c2 indicates the no-load power of the machine.
[0113] Step 2.1.2 Formulate constraints, including
[0114] 1) Constraints on the processing time of a single subtask: B ijk +t ijk ≤WG ijk
[0115] 2) Sequence constraints between subtasks of the same task: B ijk >WG i(k-i)j
[0116] 3) Each subtask can only be processed on one machine at a time:
[0117] 4) Subtasks cannot be interrupted once processing begins: WG ijk -B ijk =t ijk
[0118] 5) The completion time of any subtask shall not exceed the maximum completion time: WG ijk ≤C max
[0119] In the formula: B ijk WG represents the start time of the k-th subtask of task i on machine j; ijk WG represents the completion time of the k-th subtask of task i on machine j; i(k-i)j This represents the end processing time of the (k-1)th subtask of task i on machine j.
[0120] Step 2.2: Execute the multi-objective optimization NSGA-II algorithm to solve the objective function established in Step 2.1.1.
[0121] The NSGA-II algorithm is executed using the candidate equipment information and the planned processing time of the candidate equipment as input data. When the population reaches the maximum number of iterations required during the execution of the NSGA-II algorithm, the final Pareto optimal solution set is output and the algorithm optimization process terminates.
[0122] Step 2.3: Assign tasks based on the output of the multi-objective optimization NSGA-II algorithm in Step 2.2.
[0123] The NSGA-II algorithm is executed to output the Pareto optimal solution set. Then, based on the Pareto optimal solution set, a corresponding task execution Gantt chart is drawn to obtain the optimization scheme for task implementation. The task is optimized and matched, and it is determined whether the optimized and matched task information can meet the delivery time requirements. If so, the optimized and matched task information is synchronized to the designated processing unit to carry out the task work; otherwise, proceed to step three.
[0124] Step 3: For tasks in the manufacturing service alliance where the service providers within a single enterprise cannot meet the delivery time requirements, further optimization matching is performed among multiple enterprises within the same manufacturing service alliance, based on the task information obtained after the optimization matching in Step 2.
[0125] Step 3.1: Establish an objective function for selecting service providers within a single manufacturing service alliance.
[0126] Based on the requirements of the task objectives, service cost, completion cycle, and service quality are used as the three indicators for selecting service providers within a single manufacturing service alliance:
[0127] 1) Service costs:
[0128] In the formula: i represents the number of the task to be processed, i∈(1,2,...,m′), m′ represents the total number of tasks; j represents the number of the service provider available in the existing manufacturing service alliance, j∈(1,2,...,n′), n′ represents the total number of service providers; k represents the k-th subtask of task i, k∈(1,2,...,p′), p′ represents the total number of subtasks in task i; c′ ijk X represents the service cost of the k-th subtask of task i at service provider j; ijk The value is 1 if the k-th subtask of task i is served by service provider j, and 0 otherwise.
[0129] 2) Completion period: minT′=min(maxC′) max )=min(maxC′ i (i∈m′)
[0130] In the formula: C′ max Indicates the maximum completion time for different tasks; C′ i Indicates the completion time of task i.
[0131] 3) Service quality: Q = L1 + L2
[0132]
[0133]
[0134] The above formula represents service quality by comprehensively considering process and outcome metrics, where: L1 is the task completion inspection pass rate, used as a process metric; L2 is the rework rate, used as an outcome metric; L Q Indicates the number of items that passed inspection upon completion of the task; L T Indicates the total number of inspections required to complete the task; L t Indicates rework hours; L to It represents the total production labor hours.
[0135] After identifying service provider resources, optimize and match abnormal tasks.
[0136] Step 3.2: Execute the multi-objective optimization NSGA-II algorithm to solve the objective function established in step 3.1.
[0137] Using known task processing service providers and abnormal task information as input data, the NSGA-II multi-objective optimization algorithm is executed. When the population reaches the maximum number of iterations required, the final Pareto optimal solution set is output, the multi-objective optimization scheme is obtained, and the algorithm ends.
[0138] Step 3.3: Optimization and Implementation of Abnormal Tasks
[0139] Based on the multi-objective optimization scheme obtained in step 3.2, determine whether the service providers in the multi-objective optimization scheme can meet the delivery time requirements of the abnormal task. If so, match the abnormal task with the selected service providers and assign them to process the corresponding abnormal task; otherwise, proceed to step four.
[0140] Step 4: For tasks that still cannot be delivered on time after optimization and matching within a single manufacturing service alliance in Step 3, reconfigure them across multiple manufacturing service alliances to ensure on-time delivery and improve customer satisfaction (see reference). Figure 4 ):
[0141] Step 4.1: Establish a candidate set optimization model for manufacturing alliances to select alliance resources.
[0142] Among numerous service resource alliances, manufacturing service alliances that meet the requirements are selected. This is mainly achieved by using the historical performance and on-time delivery rate of each manufacturing service alliance as a reference for reputation indicators, and establishing an optimal model for the candidate set of manufacturing service alliances to support the final selection of the candidate set.
[0143] 1) Historical performance of obligations:
[0144] In the formula: D Y Indicates the number of times a contract has been fulfilled on time throughout history; D T This indicates the total number of historical missions.
[0145] 2) On-time delivery rate:
[0146] In the formula: Z L Indicates the number of on-time deliveries; Z T Indicates the total number of deliveries;
[0147] Step 4.2: Resource optimization, mainly including three parts: optimization model indicators, grey relational analysis, and task implementation. The specific steps are as follows:
[0148] Step 4.2.1: Construct service provider selection model indicators
[0149] First, for different manufacturing service alliances, we construct service provider selection model indicators based on service time, service cost, service quality, availability, maintainability, and sustainability. We then further select multiple service providers from the candidate set of manufacturing service alliances and determine the appropriate implementation plan through grey relational analysis.
[0150] 1) Service time: T = T pro +T trs
[0151] In the formula: T pro Indicates the time the task is served by the service provider; T trs This indicates the transportation time between the various service providers.
[0152] 2) Service cost: C″=C′+C t
[0153]
[0154]
[0155] The service cost in the above formula takes into account both the processing cost C′ and the transportation cost C. t In the formula, i represents the number of the task to be processed, i∈(1,2,...,m′), and m′ represents the total number of tasks; j represents the number of the service provider that can provide services in the existing manufacturing service alliance, j∈(1,2,...,n′), and n′ represents the total number of service providers; k represents the k-th subtask of task i, k∈(1,2,...,p′), and p′ represents the total number of subtasks; c′ ijk X represents the service cost of the k-th subtask of task i at service provider j; ijk A value of 1 indicates that the k-th subtask of task i is served by service provider j, and a value of 0 indicates that the subtask is not served by service provider j; d jj′ q represents the distance between different service providers; q represents the unit transportation price.
[0156] 3) Service quality:
[0157] In the formula: Q s This indicates the number of times the service quality rating for completing a task was rated as "good"; Q t This indicates the total number of times the service has been evaluated.
[0158] 4) Availability:
[0159]
[0160] In the formula: T ∝ This represents the total runtime available from the service providers within the most recent time period α.
[0161] 5) Maintainability: The probability that the service provider can correctly maintain the system in the event of an anomaly:
[0162] In the formula: Y C Indicates the number of times unexpected maintenance was successfully performed; Y t This indicates the total number of unexpected events.
[0163] 6) Sustainability:
[0164]
[0165] In the formula: i represents the historical demanders who made the evaluation, i∈(1,2,...,n), n historical demanders, S i It is a historical demand-side assessment of the sustainability of the manufacturing services alliance, n s This indicates the total number of times the evaluation was conducted.
[0166] Step 4.2.2: Grey Relational Analysis
[0167] 1) Standardization of service provider selection model indicators
[0168] Since the selection indicators of different service provider selection models are of different qualities and have inconsistent dimensions, it is necessary to perform dimensionless processing on the initial parameter values of the selection indicators, that is, to normalize them to reduce the differences in the absolute values of the data. Among the service provider selection indicators mentioned above, service time and service cost are cost-type indicators, while service quality, availability, maintainability, and sustainability are benefit-type indicators. The specific normalization formula for the indicators is as follows:
[0169]
[0170]
[0171] In the formula: v ij These are the initial parameter values for the preferred index.
[0172] 2) Calculate the relative optimal grey relational degree matrix γij :
[0173]
[0174] In the formula: y ij ε is the normalized result of the indicator, and ε is the resolution coefficient, ε∈[0,1], generally taken as ε=0.5.
[0175] 3) Determine the weights of the preferred indicators:
[0176] First, the data of the preferred model indicators for each service provider are sampled and statistically processed to remove invalid statistical data. Then, the goodness-of-fit χ² test is applied to the overall statistical data sample obtained after removing invalid statistical data. 2 The hypothesis testing method is used to perform hypothesis testing, and finally the weights of the preferred index are adjusted.
[0177] Hypothesis testing:
[0178]
[0179]
[0180]
[0181] Where: χ 2 χ² of the sample 2 Value; α is the significance level, typically 0.05. It is the rejection region of the test statistic; υ i p represents the frequency of the sample observations in interval i; i ω′ represents the probability of the sample interval i in the assumed population distribution function; n represents the number of samples; n′ represents the number of preferred indicators, j = 1, 2, ..., n′; m′ is the number of groups into which the statistical data is divided; ω′ j ω is the initial value for the weight of the preferred index j; j This is the correction value for the weight of the preferred index j.
[0182] Weight adjustment: Optimization index matrix ω = [ω1, ω2, ..., ω] j ,...,ω′ n ],ω j This represents the adjustment value of the weight of the j-th preferred index.
[0183] 4) Calculate the grey weighted correlation degree u i :
[0184]
[0185] In the formula: u i Let be the grey weighted correlation degree between the i-th decision scheme and the ideal decision scheme;
[0186] Based on the above formula, according to the relative optimal grey relational degree matrix γ ij It can calculate the grey weighted correlation degree u i The maximum value.
[0187] Step 4.2.3: Task Implementation Phase
[0188] The gray weighted correlation degree u i The service resource combination corresponding to the maximum value is taken as the optimal service provider combination scheme. It is determined whether the optimal service provider combination scheme can complete the abnormal task. If so, the optimal service provider combination scheme is matched with the abnormal task, and it is specified to process the corresponding abnormal production task to complete the reconfiguration of the task under the abnormal disturbance. If not, proceed to step five.
[0189] Step 5: For abnormal production tasks that are still difficult to complete after reconfiguration in Step 4 and exhibit abnormal disturbances, the manufacturing service alliance to which the service provider affected by the abnormal disturbance belongs shall act as the task issuer and initiate the abnormal production task on the manufacturing service platform for reconfiguration.
[0190] Step 5.1: Task Requirements Analysis
[0191] In the manufacturing service platform, a requirements analysis is performed on the tasks published by the task publishers. This analysis covers the basic attributes, functional requirements, and non-functional requirements of the tasks. Specifically, it includes a comprehensive analysis of the task ID, the required processing methods, and the quality requirements. This helps to further clarify the specific requirements of the tasks and supports the smooth allocation of subsequent tasks.
[0192] Step 5.2: Task Allocation and Implementation Process
[0193] Based on the process of candidate selection, task matching, task implementation tracking and monitoring, and task results and evaluation, tasks are optimized and allocated among qualified candidate service providers. The task execution process is tracked and monitored in real time on the manufacturing service platform to complete the allocation and implementation of tasks. Service providers are evaluated on the manufacturing service platform based on the completed tasks.
[0194] Step 5.3: Updating and Feedback of Manufacturing Service Platform Data
[0195] The manufacturing service platform records evaluations of task completion results in real time, serving as historical data to support the re-selection of service providers and further enabling real-time feedback and updates of service provider information within the platform.
[0196] In summary, the above-mentioned multi-level linkage solution enables the manufacturing service alliance to generate greater economic benefits, thereby completing the allocation of abnormal production tasks.
[0197] In addition to the above-mentioned method for dynamic optimization and configuration of intelligent manufacturing service chains, the present invention also provides a system for dynamic optimization and configuration of intelligent manufacturing service chains, including: an internal production anomaly identification and location module, an anomaly disturbance optimization and configuration module within a manufacturing service alliance (including an internal anomaly disturbance optimization module and an anomaly disturbance optimization module within the manufacturing service alliance), an anomaly disturbance reconfiguration module between manufacturing service alliances, and a manufacturing service platform.
[0198] The enterprise-internal production anomaly identification and location module is used to acquire real-time production information through intelligent sensing devices during manufacturing. Based on the decision tree method, it constructs the judgment criteria for production anomalies in the manufacturing service alliance, analyzes and evaluates the key event information in the real-time acquired production tasks, diagnoses production anomalies in a timely manner, quickly locates the cause of the anomalies, and feeds back and synchronizes them to the manufacturing service alliance for global manufacturing resource information sharing. This better assists production managers in the manufacturing service alliance to obtain production anomaly information in a timely and accurate manner, providing them with a basis for decision-making.
[0199] The abnormal disturbance optimization configuration module within the manufacturing service alliance constructs an optimization configuration mathematical model based on the identification and location results of production abnormality information fed back by the enterprise's internal production abnormality identification and location module, and uses a multi-objective solution algorithm to obtain a better task allocation scheme.
[0200] The inter-manufacturing disturbance reconfiguration module, based on the dynamic optimization configuration scheme for production anomalies provided by the intra-manufacturing disturbance optimization configuration module, further optimizes and matches tasks for which the service providers provided by a single manufacturing service alliance cannot meet the delivery time requirements, expands available resources, and provides a reconfiguration solution for production anomalies by reconfiguring across multiple manufacturing service alliances.
[0201] For tasks that are difficult to complete independently and are subject to abnormal disturbances, the Manufacturing Service Alliance, with the help of the Manufacturing Service Platform, initiates production anomaly tasks that are difficult to complete as the task issuer. The platform then performs requirements analysis, candidate selection, task matching, task implementation tracking and control, and task results evaluation on the received tasks, thus completing the configuration of production anomaly tasks.
[0202] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when run, is used to execute and implement the intelligent manufacturing service chain dynamic optimization configuration method provided by the present invention.
[0203] In addition, the present invention also provides an electronic device, including a processor and a memory, wherein a computer program is stored in the memory, and the computer program is executed by the processor to execute and implement the intelligent manufacturing service chain dynamic optimization configuration method provided by the present invention.
Claims
1. A method for dynamically optimizing and configuring intelligent manufacturing service chains, characterized in that, Includes the following steps: Step 1: Based on the decision tree algorithm, identify internal production anomalies and locate the causes of the anomalies; Step 2: Within a single enterprise in the Manufacturing Services Alliance, optimize and match several abnormal production tasks identified in Step 1, and determine whether the optimized and matched task information can meet the delivery time requirements. If yes, synchronize the optimized and matched task information to the designated processing unit; otherwise, proceed to Step 3. Step 3: For tasks that cannot meet the delivery time requirements within a single enterprise in the manufacturing service alliance after optimization and matching in Step 2, further optimization and matching are performed among multiple enterprises in a single manufacturing service alliance to obtain a multi-objective optimization scheme. It is then determined whether the service providers in the multi-objective optimization scheme can meet the delivery time requirements. If so, the abnormal task is matched with the screened service providers, and they are assigned to process the corresponding abnormal task; otherwise, proceed to Step 4. Step 4: For tasks that still cannot be delivered on time after optimization and matching in a single manufacturing service alliance in Step 3, reconfigure among multiple manufacturing service alliances to obtain the optimal service provider combination scheme. Determine whether the optimal service provider combination scheme can complete the abnormal task. If yes, match the optimal service provider combination scheme with the abnormal task and assign it to the corresponding abnormal production task; otherwise, proceed to Step 5. Step 5: For tasks that are still difficult to complete after reconfiguration in Step 4 and have abnormal disturbances, the manufacturing service alliance to which the task belongs shall, as the task issuer, initiate the abnormal production task on the manufacturing service platform for reconfiguration.
2. The method for dynamic optimization configuration of intelligent manufacturing service chain according to claim 1, characterized in that: Step 1 is as follows: Step 1.1: Based on different production processes, the decision tree algorithm is used to analyze the historical data information of people, machines, materials, methods and environment in the production system. Based on the analysis results, a decision tree for production anomaly extraction and cause localization is obtained. Each path from the root node to the leaf node in the decision tree is the corresponding rule, thus forming a corresponding rule library for anomaly extraction and cause localization. Step 1.2: Analyze and compare the real-time task implementation production progress, the deviation value from the initial plan, the processing time of each equipment, and the pass rate of the process with the rules in the anomaly extraction rule base constructed in Step 1.
1. Following the goal of timely delivery of production tasks, identify the deviation value between the progress of each production task and the initial plan. When the deviation value between the progress of a production task and the initial plan exceeds the set threshold, it is judged as a production anomaly event. Step 1.3 Matches the production anomaly event identified in Step 1.2 with the rules in the anomaly cause localization rule base constructed in Step 1.1 to locate the cause of the production anomaly event and transmit it to the production management system.
3. The method for dynamic optimization configuration of intelligent manufacturing service chain according to claim 2, characterized in that: The specific method for optimizing task matching in step 2 is as follows: Step 2.1: Parameter control for multi-objective optimization matching; Step 2.1.1: Establish the objective function: In the formula: The completion time; The maximum completion time for different tasks; For the task The completion time, This is the number of the task to be processed. , The total number of tasks; For machine load; This refers to the number of machines that a company within a manufacturing services consortium can provide. , The total number of machines; For the task The Sub-tasks, , This represents the total number of subtasks; For the task The Sub-tasks in the machine Processing time in the process; For the task The Sub-tasks in the machine The value is 1 if the processing is above, and 0 otherwise; For processing costs; For the task The Sub-tasks in the machine The processing cost required for the above processing; This refers to the machine's load power. The end or completion time of the machine; This refers to the machine's no-load power. Step 2.1.2: Formulating Constraints: Constraints on the processing time of individual subtasks: ; Sequence constraints exist between subtasks of the same task: ; Each subtask can only be processed on one machine at a time. ; Subtasks cannot be interrupted once processing begins: ; The completion time of any subtask shall not exceed the maximum completion time. ; In the formula: For the task The Sub-tasks in the machine The start time of processing; To indicate the task The Sub-tasks in the machine End processing time; For the task The Sub-tasks in the machine End processing time; Step 2.2: Solve the objective function established in Step 2.1.1 using a multi-objective optimization algorithm; Step 2.3: Obtain the optimization scheme for task implementation based on the optimal solution set of the multi-objective optimization algorithm, and perform optimization matching on the task.
4. The method for dynamic optimization configuration of intelligent manufacturing service chain according to claim 3, characterized in that: The multi-objective optimization algorithm in step 2.2 is the NSGA-II multi-objective optimization algorithm.
5. The method for dynamic optimization configuration of intelligent manufacturing service chain according to claim 4, characterized in that: Step 2.3 involves drawing a corresponding Gantt chart of task execution based on the optimal solution set of the multi-objective optimization NSGA-II algorithm to obtain the optimization scheme for task implementation.
6. The method for dynamic optimization configuration of intelligent manufacturing service chain according to any one of claims 3-5, characterized in that: The specific method for further optimizing and matching among multiple enterprises in a single manufacturing service alliance in step 3 is as follows: Step 3.1: Establish an objective function for selecting service providers within a single manufacturing service consortium: In the formula: For service costs; This is the number of the task to be processed. ϵ(1,2,..., ), The total number of tasks; This refers to the service provider IDs available within the existing manufacturing services consortium. , Total number of service providers; For the task The Sub-tasks, , For the task Total number of neutron missions; For the task The Each task is handled by the service provider. Service costs; For the task The Each task is handled by the service provider. The value is 1 if the service is provided, and 0 otherwise. Minimum completion time; The maximum completion time for different tasks; Task Completion time; For service quality; The pass rate for task completion inspection is used as a process metric. The rework rate is used as a metric for the outcome. Indicates the number of items that passed inspection upon completion of the task; This indicates the total number of inspections required to complete the task. Indicates rework hours; Represents total production labor hours; Step 3.2: Solve the objective function established in Step 3.1 using a multi-objective optimization algorithm to obtain the multi-objective optimization scheme.
7. The method for dynamic optimization configuration of intelligent manufacturing service chain according to claim 6, characterized in that: The specific method for reconfiguring among multiple manufacturing service alliances in step 4 is as follows: Step 4.1: Establish a model for optimizing the candidate set of manufacturing alliances: In the formula: Historical performance record; The number of times historical contracts have been fulfilled on time; The total number of historical missions; To improve on-time delivery rate; To ensure timely delivery of the required number of shipments; This represents the total number of deliveries. Step 4.2: Resource optimization; Step 4.2.1: Construct service provider selection model indicators: In the formula: For service hours; The time the task is served by the service provider; For the transportation time between the various service providers; For service costs; For processing costs; For transportation costs; This refers to the service provider IDs within the existing manufacturing services alliance. , Total number of service providers; For the task The Sub-tasks, , This represents the total number of subtasks; For the task The The service cost of each sub-task in service provider J; For the task The Each task is handled by the service provider. The value is 1 if the service is provided, and 0 otherwise. For the distance between different service providers; Price per unit of transportation; Total available runtime among service providers within the most recent time α; For historical demanders' assessment of the sustainability of the manufacturing services alliance; The total number of times it was evaluated; Step 4.2.2: Grey Relational Analysis First, the service provider selection model indicators are normalized. Then, the relatively optimal grey relational degree matrix is calculated. Next, the selection indicator weights are determined. Finally, the maximum value of the grey weighted relational degree is calculated based on the relatively optimal grey relational degree matrix and the selection indicator weights. Step 4.2.3: The service resource combination corresponding to the maximum value of the gray weighted correlation degree is taken as the optimal service provider combination scheme.
8. A dynamic optimization configuration system for intelligent manufacturing service chains, characterized in that, include: The enterprise internal production anomaly identification and location module is used to acquire real-time production information through intelligent sensing devices in the manufacturing process, construct the judgment criteria for production anomalies in the manufacturing service alliance based on the decision tree method, analyze and evaluate the key event information in the real-time acquired production tasks, diagnose production anomaly events in a timely manner, locate the cause of the anomaly and provide feedback. The abnormal disturbance optimization configuration module within the manufacturing service alliance constructs an optimization configuration mathematical model based on the identification and location results of production abnormality information fed back by the enterprise's internal production abnormality identification and location module, and uses a multi-objective solution algorithm to obtain a better task allocation scheme. The anomaly optimization configuration module within the manufacturing service alliance includes an anomaly optimization module within the enterprise and an anomaly optimization module within the manufacturing service alliance. The inter-manufacturing service alliance anomaly reconfiguration module is used to reconfigure tasks for which the service providers of a single manufacturing service alliance cannot meet the delivery time requirements across multiple manufacturing service alliances. The manufacturing service platform conducts demand analysis and optimizes the allocation of tasks among qualified candidate service providers based on the process of candidate selection, task matching, task implementation tracking and monitoring, and task results and evaluation.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is run, it is used to execute the intelligent manufacturing service chain dynamic optimization configuration method according to any one of claims 1-7.
10. An electronic device, comprising a processor and a memory, wherein a computer program is stored in the memory, characterized in that: The computer program, when run by the processor, is used to execute the intelligent manufacturing service chain dynamic optimization configuration method according to any one of claims 1-7.
Citation Information
Patent Citations
Workshop manufacturing process-oriented active sensing and anomaly analysis method of real-time generating performance
CN104407589A
Intelligent workshop production optimization method based on dynamic bottleneck prediction
CN110163436A
