Customer income-oriented cloud manufacturing machine tool service resource scheduling method and system

Abstract

The invention belongs to the technical field of resource scheduling, and provides a cloud manufacturing machine tool service resource scheduling method and system oriented to customer income, and the method comprises the steps: distinguishing customer types based on customer personalized preference demands, carrying out the weighting of machine tool service key indexes according to the customer types, and constructing a cloud manufacturing customer task income function; solving the cloud manufacturing customer task revenue function by adopting an improved genetic algorithm to obtain an optimal machine tool service resource scheduling scheme; according to the improved genetic algorithm, an individual fitness function is reconstructed based on a non-cooperative game, a machine tool service resource scheduling scheme is optimized, and multi-task income Nash equilibrium is achieved. The technical problem that dynamic balance of client task income under the influence of multiple factors is usually difficult to consider by an existing scheduling method is solved.

Description

Technical Field

[0001] This invention belongs to the field of resource scheduling technology, and in particular relates to a cloud manufacturing machine tool service resource scheduling method and system oriented towards customer benefits. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] In cloud manufacturing environments, manufacturers' optimal allocation of machine tool resources is often influenced by the diverse and customized task requirements of customers. Existing research on cloud manufacturing scheduling problems mainly focuses on optimizing overall objectives, such as minimizing total completion time. The basic idea is to treat the objectives of multiple manufacturing tasks as a unified global optimization function and solve it centrally.

[0004] However, this approach, to some extent, ignores the personalized preferences and needs of individual customer tasks, making it difficult to reflect the customer-centric nature of cloud manufacturing services. Due to the diversity and customization of customer task requirements, traditional resource scheduling methods are ineffective, resulting in underutilization of machine tool resources and difficulty in maximizing customer benefits.

[0005] Meanwhile, as a service-oriented manufacturing model, cloud manufacturing commonly faces competition for manufacturing resources among customers due to differing needs, such as limitations in the number of machine tools, leading to potential conflicts of interest. As the manufacturing industry increasingly demands energy conservation, emission reduction, and green development, existing scheduling methods often struggle to effectively allocate limited resources while accommodating the personalized manufacturing preferences of different customers, thus failing to achieve a balance of revenue and energy efficiency optimization among them.

[0006] The aforementioned issues, to some extent, limit the fairness and sustainability of cloud manufacturing services, and urgently require new scheduling methods. Summary of the Invention

[0007] The purpose of this invention is to provide a cloud manufacturing machine tool service resource scheduling method and system oriented towards customer benefits. By improving the genetic algorithm, it can effectively schedule limited resources and solve the technical problem that existing scheduling methods often cannot take into account the dynamic balance of customer task benefits under the influence of multiple factors.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: The first aspect of this invention provides a cloud manufacturing machine tool service resource scheduling method oriented towards customer benefits, comprising: Based on the personalized preferences and needs of customers, customer types are distinguished, and key indicators of machine tool services are weighted according to customer type to construct a cloud manufacturing customer task revenue function. An improved genetic algorithm is used to solve the task revenue function of cloud manufacturing customers to obtain the optimal machine tool service resource scheduling scheme; The improved genetic algorithm reconstructs the individual fitness function based on non-cooperative game theory, optimizes the machine tool service resource scheduling scheme, and achieves a multi-task revenue Nash equilibrium.

[0009] A second aspect of the present invention provides a cloud-based machine tool service resource scheduling system for customer benefit, comprising: The function building module is configured to: differentiate customer types based on customer personalized preferences, assign weights to key machine tool service indicators according to customer type, and build a cloud manufacturing customer task revenue function. The function solving module is configured to use an improved genetic algorithm to solve the cloud manufacturing customer task revenue function and obtain the optimal machine tool service resource scheduling scheme. The improved genetic algorithm reconstructs the individual fitness function based on non-cooperative game theory, optimizes the machine tool service resource scheduling scheme, and achieves a multi-task revenue Nash equilibrium.

[0010] A third aspect of the present invention provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps of a cloud manufacturing machine tool service resource scheduling method for customer benefit as described in the first aspect of the present invention.

[0011] A fourth aspect of the present invention provides an electronic device including a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of a cloud manufacturing machine tool service resource scheduling method for customer benefit as described in the first aspect of the present invention.

[0012] The technical solution of the present invention has the following beneficial effects: This invention introduces an improved genetic algorithm, reconstructing the individual fitness function based on non-cooperative game theory. Simultaneously, it adaptively designs crossover and mutation operations, enabling it to comprehensively consider multiple dimensions such as completion time, energy consumption, cost, and overall quality in resource-competitive scenarios involving multiple tasks and process routes. This overcomes the limitation of existing scheduling methods that only focus on the overall optimal goal, achieving optimal scheduling of machine tool service resources and reaching a Nash equilibrium of multi-task benefits. On one hand, by incorporating customer preference requirements, the scheduling process fully reflects the personalized goals of different customers, effectively solving the problem that traditional methods struggle to consider the benefits of individual customer tasks. On the other hand, by using a non-cooperative game theory framework to model resource competition among multiple customer tasks, it avoids uneven customer benefits caused by resource contention, achieving a dynamic equilibrium of customer benefits.

[0013] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0014] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0015] Figure 1 This is a flowchart of the improved genetic algorithm in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the energy consumption calculation upon completion in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the double-layered chain structure chromosome encoding in Embodiment 1 of the present invention; Figure 4 This is a schematic diagram of the cross operation in Embodiment 1 of the present invention; Figure 5 This is a schematic diagram of the mutation operation in Embodiment 1 of the present invention; Figure 6 This is the fitness iteration convergence graph in Embodiment 1 of the present invention; Figure 7 This is the second-generation Gantt chart in Embodiment 1 of the present invention; Figure 8 This is the 555th generation Gantt chart in Embodiment 1 of the present invention; Figure 9 This is a bar chart comparing customer task revenue in Embodiment 1 of the present invention; Figure 10 This is a trend chart of completion time variation in Embodiment 1 of the present invention; Figure 11 This is a graph showing the trend of energy consumption upon completion in Embodiment 1 of the present invention; Figure 12 This is a graph showing the trend of the completion cost in Embodiment 1 of the present invention; Figure 13 This is a graph showing the overall quality change trend in Embodiment 1 of the present invention. Detailed Implementation

[0016] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0017] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0018] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0019] Example 1 like Figure 1 As shown, this embodiment discloses a cloud manufacturing machine tool service resource scheduling method oriented towards customer benefits. It constructs a task benefit function based on the preferences and needs of different customers and designs an improved genetic algorithm to solve the Nash equilibrium of the model. Specifically, it includes the following steps: Step 1: Based on customers' personalized preferences and needs, differentiate customer types, assign weights to key machine tool service indicators according to customer type, and construct a cloud manufacturing customer task revenue function.

[0020] To reflect the diversity of customer needs under the cloud manufacturing model, customers are categorized into five types based on their different preferences for key indicators: "Time-sensitive," "Energy-efficient," "Economical," "Quality-sensitive," and "Comprehensive." For example, "Time-sensitive" customers have high requirements for the completion time of processing tasks. In resource scheduling, these customers usually prioritize machine tool service resources that can respond quickly and have short processing times to improve production efficiency.

[0021] Specifically, four key indicators are constructed with the goals of manufacturing task completion time, energy consumption, cost, and overall quality, and weighted according to customer preferences as customer task benefit functions.

[0022]

[0023]

[0024] in, Indicates the customer's task revenue. This represents a manufacturing task, where i is the manufacturing task number. This indicates the completion time of the last subtask in the manufacturing task. This indicates the energy consumption required to complete a manufacturing task. This represents the cost of completing a manufacturing task. It indicates the overall quality of the manufacturing task. , , , These are the customer preference weights for the four key indicators.

[0025] (1) Completion time

[0026] The calculation process includes subtask processing time, inter-machine transfer time, and machine tool adjustment time. The specific calculation method is as follows:

[0027] in, This represents a subtask in manufacturing task i. Start processing time; Subtasks optional machine tool equipment Processing time; Subtasks The processing end time; Indicates two adjacent subtask machine tools With machine tools Transit time between; Indicates optional machine tool equipment The processing time; Subtasks The processing end time; machine tool equipment Pair Task Adjustment time before processing; As a two-dimensional variable, when the subtask Select machine tool equipment If processing is performed, the value is 1; otherwise, it is 0.

[0028] (2) Energy consumption upon completion

[0029] Processing energy consumption Idle energy consumption Transportation energy consumption It consists of three parts, and the specific calculation method is as follows:

[0030] in, machine tool equipment Processing power. machine tool equipment The no-load power; machine tool equipment The idle time; This indicates the transport power of the transfer equipment. Here, the transfer equipment refers to electrically powered Automated Guided Vehicles (AGVs). The energy consumption for completing a manufacturing task is quantified by combining the energy consumption of machine tools under different conditions and the transport energy consumption of the transfer equipment.

[0031] like Figure 2 As shown, in this embodiment, taking manufacturing task 1 as an example, machine tool equipment M1 is in In this state, the first process 1-1 of manufacturing task 1 is processed, and after completion, it is transferred by a transfer device (such as an automated guided vehicle AGV) at high power. The power is transferred to machine tool M2, at which point the power of M1 is converted to... Upon reaching M2, its power is reduced from... Convert to The second process, steps 1-2, begins. After completion, the equipment transfers the equipment again at high power. The power is transferred to machine tool M3, at which point the power of M2 is converted to... M3 power is from Convert to The third and final process, steps 1-3, begins. Therefore, the energy consumption for completing manufacturing task 1 consists of the processing energy consumption of the three machine tools in their processing state, the idle energy consumption of the three machine tools in their idle state, and the transportation energy consumption of the two transfer devices. Regarding the allocation of the machine tool idle energy consumption, referencing existing allocation models and the principles of activity-based costing, the allocation is based on the time the machine tool adds value to the task, distributing it among the tasks.

[0032] (3) Cost of completion

[0033] Processing costs Energy consumption cost It consists of two parts, and the specific calculation process is as follows:

[0034] in, Subtasks In the equipment The unit time processing cost comprehensively considers the material costs, labor costs, and equipment cloud deployment costs required for the processing; the task energy consumption cost consists of the electricity consumed by machine tools and transportation equipment. This indicates the average price of industrial electricity in the manufacturer's location.

[0035] (4) Overall quality

[0036] Main reliability pass rate With delivery rate It is quantified from three aspects, and its value is calculated based on historical order data. The specific calculation method is as follows:

[0037]

[0038] in, Indicates reliability. Indicates the pass rate. Indicates delivery rate. These represent the weighting coefficients for reliability, pass rate, and delivery rate, respectively.

[0039] (a) Reliability

[0040] reliability This refers to the stability of machine tool equipment during the processing of manufacturing tasks, reflecting the smoothness of machine tool processing and the proficiency of its application. The calculation formula is as follows:

[0041] in, machine tool equipment The degree of reliability. Machine tool equipment The processing time during which the cloud platform operates stably without any failures. express The total processing time within the same time period.

[0042] (b) Pass rate

[0043] pass rate This refers to the degree to which machine tool equipment meets customer quality standards, reflecting the machine tool's ability to satisfy requirements such as product precision and process parameters during processing. The calculation formula is as follows:

[0044] in, machine tool equipment The processing qualification rate. machine tool equipment The number of qualified products produced according to the task process requirements analyzed by the cloud platform. express The total number of products processed within the same period.

[0045] (c) Delivery rate

[0046] Delivery rate This refers to the completion status of machine tool manufacturing tasks within the customer-specified timeframe, reflecting the timeliness of the machine tool's processing. The calculation formula is as follows:

[0047] in, machine tool equipment On-time delivery rate. machine tool equipment According to the number of times the customer delivers on time as specified in the cloud platform, express The total number of deliveries within the same period.

[0048] In this embodiment, the weights of key indicators are allocated evenly based on customer preferences. If a customer pays more attention to one indicator, the weight of that indicator is appropriately increased, while the weights of other indicators are averaged out. Step 2: An improved genetic algorithm is used to solve the task reward function for cloud manufacturing customers to obtain the optimal machine tool service resource scheduling scheme. Specifically, the improved genetic algorithm reconstructs the individual fitness function based on non-cooperative game theory to optimize the machine tool service resource scheduling scheme and achieve a multi-task reward Nash equilibrium.

[0049] Genetic Algorithms (GAs), by simulating the biological evolution mechanism in nature, exhibit strong global search capabilities when solving scheduling optimization problems. However, this algorithm suffers from drawbacks such as being prone to getting trapped in local optima and premature convergence, necessitating improvements for non-cooperative game-theoretic cloud manufacturing resource scheduling optimization problems. The specific steps of the algorithm are as follows: Step 2-1: Set parameters to determine the population size n, as well as the maximum and minimum crossover rates and the maximum and minimum mutation rates.

[0050] Step 2-2: Encoding.

[0051] A two-layer chain structure is used to encode chromosomes. Each chromosome consists of two parts: a subtask layer (ST) and a device layer (MS). The number of genes in each layer is equal to the total number of subtasks to be processed. ST represents the processing execution order of the subtasks corresponding to each task, and its value is represented by the task number and its frequency. MS represents one available machine tool for the corresponding subtask, and its value is represented by the machine tool number. For example... Figure 3 As shown, in the task layer, the first occurrence of 2 indicates the first subtask of task 2, the second occurrence of 2 indicates the second subtask of task 2, and so on. In the equipment layer, the first position 5 indicates that the first subtask of task 1 is processed on machine tool 5, the second position 3 indicates that the first subtask of task 3 is processed on machine tool 3, and so on.

[0052] This type of coding method not only ensures the rationality of the processing order of subtasks and the available machine tools, but also effectively avoids problems such as task order conflicts and machine tool allocation infeasibility caused by unreasonable coding methods. This provides a reliable feasible solution space for subsequent resource scheduling optimization and improves the stability and optimization efficiency of scheduling results.

[0053] Steps 2-3: Use a dual strategy for population initialization.

[0054] (1) Heuristic strategy: Normalize the processing parameters (such as equipment power, unit processing time, and unit processing cost) of the optional machine tool equipment set for sub-tasks, calculate the comprehensive score of each equipment in combination with customer preference weights, and allocate the optimal machine tool equipment combination to each sub-task accordingly.

[0055] For the processing parameters involved in the three negative indicators of time, energy consumption, and cost, the following formula is used for normalization:

[0056] For processing parameters involved in positive indicators such as overall quality, the following formula is used for normalization:

[0057] in, This represents the normalized attribute values ​​of each processing parameter. This indicates the attribute value of the parameter corresponding to the selected machine tool equipment. and These represent the maximum and minimum values ​​of the corresponding attributes for each machine tool.

[0058] (2) Random strategy Under the condition of satisfying resource constraints, a machine tool is randomly assigned to the subtask from the set of available equipment.

[0059] By employing a dual-strategy population initialization mechanism, this approach leverages heuristic strategies to improve local search accuracy and stochastic strategies to ensure solution set diversity. Simultaneously, it introduces customer preference information as a crucial basis for machine tool allocation, thereby guaranteeing that the initial population includes individuals that meet the personalized needs of customers. This dual-strategy population initialization mechanism not only overcomes the lack of individual specificity in existing methods but also significantly improves the overall quality and adaptability of the initial population, providing a more reasonable solution space for subsequent evolutionary search. This, in turn, enhances the algorithm's convergence efficiency and global optimization performance.

[0060] Steps 2-4: Reconstruct the individual fitness function based on non-cooperative game theory, and calculate the optimal reward of the previous generation based on the client's task reward function. , , , Compared with current income , , , The individual fitness value is obtained, where n refers to the nth individual (manufacturing task) and k refers to the kth generation.

[0061] The individual fitness function is designed as follows:

[0062] In the formula, Indicates the first Fitness values ​​of individuals in a generation; Indicates manufacturing task In the Income on behalf of others; Indicates manufacturing task In the The best return in the generation.

[0063] Due to the differences in dimensions among the indicators, the above-mentioned normalization processing is required before calculation. The individual fitness function design solves the problem in existing technologies where fitness evaluation only focuses on the overall goal optimization and ignores the balance of benefits among tasks. By introducing constraints on the differences in benefits among tasks and a threshold determination mechanism, this method can effectively avoid the situation where individual tasks suffer losses due to unfair resource allocation, and achieve a relative balance of customer benefits among multiple tasks, thereby improving the fairness of scheduling results and the stability of system operation.

[0064] Steps 2-5: Implement parent selection and elite retention operations based on individual selection using Stochastic Universal Sampling (SUS).

[0065] Traditional genetic algorithms often employ a roulette wheel selection strategy. While this method prioritizes the retention of superior individuals, the high variance resulting from multiple independent random sampling leads to a rapid decline in population diversity, causing premature convergence. Therefore, SUS (Single-Side Randomization) is introduced to overcome this drawback. SUS generates random pointers and scans the cumulative probability interval at equal intervals, completing the selection of multiple candidate individuals in a single sampling. Its advantage lies in effectively reducing sampling variance and maintaining population diversity. Furthermore, by incorporating an elitist strategy during the selection process, the best individuals are retained for the next generation, enhancing the algorithm's global search capability and convergence stability.

[0066] Steps 2-6: Dynamically adjust crossover or mutation probabilities to update the population, including: Define adaptive crossover probability and mutation probability:

[0067]

[0068] In the formula, Indicates the probability of crossover or mutation; Indicates the maximum crossover or mutation probability; Indicates the minimum crossover or mutation probability; This represents the crossover or mutation adjustment factor, which dynamically adjusts the crossover or mutation probability by combining information from the iteration stage and information about algorithm stalls. Indicates the crossover adjustment factor; Indicates the adjustment factor for the mutation operation; Indicates the current iteration number; Indicates the maximum number of iterations; This indicates the number of consecutive times the fitness value has not improved. Indicates the crossover and mutation probabilities of the baseline; , and These represent the individual fitness value, the average fitness value, and the maximum fitness value of the current population, respectively. All are constants. Traditional genetic algorithms use fixed crossover and mutation probabilities, without dynamically updating them based on individual fitness information and population iteration status. In this embodiment, to improve the search capability for Nash equilibrium solutions, adaptive crossover and mutation probabilities are defined. In the design of this adaptive crossover and mutation probability mechanism, the algorithm enhances the balance between solution space exploration and development by dynamically adjusting the occurrence probabilities of the two types of operators. Simultaneously, it preserves the superior genetic structure of elite individuals based on their fitness values ​​while maintaining population diversity through an adaptive strategy.

[0069] By designing adaptive crossover and mutation probabilities, the probability of occurrence of the two types of operators is dynamically adjusted, enhancing the algorithm's balance between solution space exploration and development. Simultaneously, it preserves the superior genetic structure of elite individuals based on their fitness values ​​while maintaining population diversity through adaptive strategies. (1) Cross operation In genetic algorithms, crossover occurs by exchanging gene information between two parents, thereby achieving population evolution. In this embodiment, the crossover operation occurs at the sub-task level, that is, the sub-task information of two parents is randomly exchanged. Figure 4 As shown, the specific steps are as follows: a) Randomly divide the task set into two non-empty sets. and .

[0070] b) The parent generation Regarding the task set All subtask information is copied to the child in their corresponding order. Middle; Father generation Regarding the task set All subtask information is copied to the child in their corresponding order. middle.

[0071] c) The parent generation About the task set The supplementary set includes all sub-tasks and parent generation The corresponding subtasks perform gene alignment one by one in positional order. When the alignment rate exceeds a set threshold, the next step is performed in the parent generation. Two gene positions are randomly generated, and the gene segments from one position are reversed and then inserted into the offspring. Middle. Offspring Similarly, this crossover method avoids the generation of invalid individuals caused by multiple site repetitions of the gene sequence in the child task layers of the two parents due to relative positions, while preserving the effective gene sequence.

[0072] (2) Mutation operation The mutation operation in a genetic algorithm introduces new genes to improve population diversity by randomly changing some gene values ​​in the individual's encoding. In this embodiment, multi-point mutation is used for mutation operations at both the subtask layer and the device layer. Before designing specific mutation operations, a Lévy fight strategy is introduced to dynamically perturb the mutation process, effectively reducing the probability of the algorithm getting trapped in local optima. The formula for generating the Lévy fight step size is shown below:

[0073] In the formula, and These are two random numbers that follow a normal distribution; It is a constant, usually taking the value 1.5; is the scale parameter. Meanwhile, to ensure that Lévy Fight satisfies the discretization characteristics of the cloud manufacturing scheduling problem, the following method is used to obtain the number of mutated genes in the subtask layer and the device layer.

[0074]

[0075] in, Indicates the number of mutated genes in the subtask layer or device layer; Indicates the maximum number of mutations in the subtask layer gene or the device layer gene; This represents the minimum number of mutations in a subtask layer gene or a device layer gene. This represents the gene variation adjustment factor of the subtask layer or device layer.

[0076] In this embodiment, separate mutation operation procedures were designed for the sub-task layer and the device layer. For example... Figure 5 As shown. Subtask Layer: First, based on the number of mutated genes calculated by Lévy Fight, multiple gene positions are randomly selected, and corresponding subtask sequence combinations are obtained; second, the extracted subtask sequence combinations are randomly combined; finally, the shuffled subtask sequences are inserted into the chromosome to obtain new subtask layer codes. Equipment Layer: Based on the number of mutated genes calculated by Lévy Fight, multiple gene positions are randomly selected, and the available equipment sets for each subtask corresponding to different gene positions are identified, and new machine tools are randomly selected from them.

[0077] Steps 2-7: After completing the population update, determine whether the fitness value of an individual is less than or equal to a preset threshold. If yes, output the optimal individual, i.e., the Nash equilibrium solution; otherwise, continue iterating, while also determining whether the maximum number of iterations has been reached. Specifically, the judgment formula is as follows:

[0078] in, This is considered the threshold for reaching Nash equilibrium, i.e., when the sum of the revenue deviations of all manufacturing tasks steadily converges to... If so, it is considered that an approximate Nash equilibrium solution to the non-cooperative game scheduling problem has been obtained.

[0079] Step 2-8: Determine if the maximum number of iterations has been reached. If yes, output the best individual in the current population as the approximate Nash equilibrium solution; otherwise, let k = k + 1 and return to step 2-4.

[0080] In this embodiment, the method is verified using virtual experimental simulation data of a typical heavy machinery cylinder piston rod cloud manufacturing task. The processing tasks submitted by multiple clients on the cloud platform are set as follows: crane luffing cylinder piston rod (Task 1), petroleum machinery lifting cylinder piston rod (Task 2), press cylinder piston rod (Task 3), tunnel boring machine propulsion cylinder piston rod (Task 4), and rotary drilling rig tilting cylinder piston rod (Task 5). The specific process routes for different types of piston rods are shown in Table 1. The client preference for Task 1 is "time-sensitive"; for Task 2, it is "energy-efficient"; for Task 3, it is "economical"; for Task 4, it is "quality-oriented"; and for Task 5, it is "comprehensive". The weights of different client preferences are shown in Table 2.

[0081] Table 1. Process routes for piston rod machining tasks submitted by different customers

[0082] Table 2 Customer Preference Weight Reference Table

[0083] The machine tools are divided into the following groups according to the process type: turning, welding, electroplating, grinding, drilling, and straightening. Table 3 shows some of the machine tool parameters. The turning group has a total of 3 machines (…). The machining process includes {blank, rough turning, finishing, finish turning of the outer diameter, finish turning of the small end, rough machining, turning, finish turning, end face turning, boring and milling}; the welding group has a total of 1 piece of equipment ( The processing technology includes {welding rod plug, welding earring, welding, welding rod head}; the electroplating group is equipped with 1 piece of equipment ( The processing technology is {electroplating}; the grinding group has a total of 2 machines ( The processing technology includes {outer diameter grinding, polishing, and grinding}; the drilling group has a total of 1 machine ( The processing technology is categorized as {boring}; the straightening group is equipped with 1 piece of equipment ( The processing technology set is {straightening}. Based on the above information, the processing parameters provided by the enterprise are randomly assigned using simulation, resulting in the datasets shown in Tables 4 and 5 (Task 1 is used as an example for easy display). The average price of industrial electricity is taken as 0.61 (unit: yuan / (kW·h)).

[0084] Table 3. Partial Machine Tool Machining Parameter Information

[0085] Table 4 Task Processing Data Information

[0086] Table 5 Equipment Transportation Time (minutes)

[0087] An improved genetic algorithm is used to solve the above non-cooperative game scheduling problem: Parameter settings: Population size 300; minimum and maximum crossover probabilities 0.65 and 0.95; minimum and maximum mutation probabilities 0.07 and 0.35; minimum and maximum mutation counts at the subtask level 3 and 15; minimum and maximum mutation counts at the device level 1 and 20; number of iterations 800. Figure 6 As shown, the convergence trend of the fitness value is illustrated. The results indicate that in the early stages of iteration, adaptive crossover mutation and the Lévy flight strategy enhance population diversity, while the heuristic initialization method accelerates the decrease in fitness value. After approximately 400 generations, the algorithm enters the convergence phase and stabilizes at 0.026 around the 555th generation.

[0088] like Figure 7 , Figure 8As shown, comparing the second and fifth generations of the algorithm reveals that in the second generation, machine tool utilization is insufficient. Some subtasks can be reassigned to other machine tools to shorten their completion time. For example, transferring the first operation of task 2 from M3 to M1 can significantly advance its start time, thereby reducing its completion time. In the fifth generation, however, the completion times of all tasks are improved. Furthermore, under this scheduling scheme, it becomes difficult to reduce the completion time of a single task while keeping the completion times of other tasks unchanged.

[0089] like Figure 9 The table shows the client task rewards in generations 2 and 555 of the improved genetic algorithm. To verify the effectiveness of the improved genetic algorithm, a traditional genetic algorithm is introduced as a control. The results show that in generations 2 and the traditional genetic algorithm, the rewards among tasks did not reach equilibrium. However, in generation 555 of the improved genetic algorithm, the rewards for tasks 1-5 increased by approximately 10%, 12%, 2%, 8%, and 56% respectively compared to generation 2, and the rewards for each task tended to stabilize. At this point, the scheduling scheme reached a stable state, meaning that it is difficult for any task to further increase its own reward without reducing the rewards of other tasks. This aligns with the Nash equilibrium concept in non-cooperative games.

[0090] Table 6 shows the specific performance of each task in terms of completion time, energy consumption, cost, and quality of service (CQS) in the 2nd and 555th generations of the improved genetic algorithm and in the traditional GA (where completion time is rounded to the nearest integer). It can be seen that the improved genetic algorithm significantly optimizes all CQS metrics with increasing iteration count. From generation 2 to generation 555, the completion time, energy consumption, and cost of most tasks show a decreasing trend, while the overall quality improves or remains at a high level. Compared with the traditional genetic algorithm, the improved genetic algorithm has significant advantages in meeting the manufacturing preferences of different customers. For example, for the time-sensitive task 1, the improved genetic algorithm completes the task in 15 hours, less than the 19 hours of the traditional genetic algorithm; for the energy-efficient task 2, the improved genetic algorithm completes the task in 402 kWh, less than the 410 kWh of the traditional genetic algorithm; for the economical task 3, the improved genetic algorithm outperforms the traditional genetic algorithm in all four metrics: completion time, energy consumption, cost, and overall quality; for tasks 4 and 5, the benefits obtained using the improved genetic algorithm are more balanced than those of the traditional genetic algorithm. Experimental results show that the proposed improved genetic algorithm can better optimize the scheduling scheme based on customer preference characteristics, thereby significantly improving the overall scheduling performance while ensuring the balance of benefits for each task.

[0091] Table 6. Comparison of performance metrics between the improved GA generation 2, generation 555, and the traditional GA.

[0092] like Figures 10-13 The chart shows the changing trends of four key indicators for each task: completion time, completion energy consumption, completion cost, and service quality. Taking energy-efficient task 2 as an example, its completion time decreased from 27 hours to 20 hours; completion energy consumption decreased from 464 kWh to 402 kWh; completion cost decreased from 5741 yuan to 5736 yuan; and the overall quality remained around 0.88. Although the total completion energy consumption of task 2 was higher than that of other tasks, this was mainly due to its specific processing technology path: on the one hand, the process route involved in task 2 was the longest among the five tasks, which increased its processing and transportation energy consumption; on the other hand, task 2 involved boring operations performed on high-power machine tools (boring machines), thus increasing the task's processing energy consumption. Overall, the energy consumption reduction of task 2 was 13%, reaching the average level of energy consumption reduction for all tasks.

[0093] Based on the experimental results, the improved genetic algorithm redesigned the fitness function and adaptively designed the crossover and mutation operations. This allows it to overcome the limitations of existing scheduling methods that only focus on the overall optimal goal in resource competition scenarios involving multiple tasks and process routes. On one hand, by introducing customer preference requirements, the scheduling process can fully reflect the personalized goals of different customers, effectively solving the problem that traditional methods struggle to consider the benefits of individual customer tasks. On the other hand, by using a non-cooperative game framework to model resource competition among multiple customer tasks, it avoids uneven customer benefits caused by resource contention, achieving a dynamic balance of customer benefits. Simultaneously, this method explicitly introduces energy consumption indicators during the scheduling optimization process, overcoming the limitations of existing methods that fail to fully consider energy conservation and consumption reduction, thereby improving the green performance and sustainability of the cloud manufacturing system. Furthermore, by balancing differentiated customer needs with energy consumption optimization goals, the scheduling results outperform traditional methods in terms of fairness, stability, and energy efficiency. In summary, this approach not only ensures fairness and personalized response in task scheduling in a multi-client environment, but also improves resource allocation efficiency, reduces energy consumption, and enhances the overall convergence performance and robustness of the system, demonstrating significant application value and promising prospects for wider adoption.

[0094] Example 2 This embodiment discloses a cloud-based machine tool service resource scheduling system oriented towards customer benefits, including: The function building module is configured to: differentiate customer types based on customer personalized preferences, assign weights to key machine tool service indicators according to customer type, and build a cloud manufacturing customer task revenue function. The function solving module is configured to use an improved genetic algorithm to solve the cloud manufacturing customer task revenue function and obtain the optimal machine tool service resource scheduling scheme. The improved genetic algorithm reconstructs the individual fitness function based on non-cooperative game theory, optimizes the machine tool service resource scheduling scheme, and achieves a multi-task revenue Nash equilibrium.

[0095] Example 3 The purpose of this embodiment is to provide a computer-readable storage medium. A computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps of a cloud manufacturing machine tool service resource scheduling method for customer benefit as described in Embodiment 1 of this disclosure.

[0096] Example 4 The purpose of this embodiment is to provide an electronic device. An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of a cloud manufacturing machine tool service resource scheduling method for customer benefit as described in Embodiment 1 of this disclosure.

[0097] The steps and methods involved in the apparatuses of Embodiments 2, 3, and 4 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0098] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0099] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A cloud manufacturing machine tool service resource scheduling method oriented towards customer benefits, characterized in that, include: Based on the personalized preferences and needs of customers, customer types are distinguished, and key indicators of machine tool services are weighted according to customer type to construct a cloud manufacturing customer task revenue function. An improved genetic algorithm is used to solve the task revenue function of cloud manufacturing customers to obtain the optimal machine tool service resource scheduling scheme; The improved genetic algorithm reconstructs the individual fitness function based on non-cooperative game theory, optimizes the machine tool service resource scheduling scheme, and achieves a multi-task revenue Nash equilibrium.

2. The cloud manufacturing machine tool service resource scheduling method oriented towards customer benefits as described in claim 1, characterized in that, The individual fitness function is specifically as follows: Where n is the nth individual, Indicates the first Fitness values ​​of individuals in a generation; Indicates manufacturing task In the Income on behalf of others; Indicates manufacturing task In the The best returns for the generation; This is considered the threshold for reaching Nash equilibrium, i.e., when the sum of the revenue deviations of all manufacturing tasks steadily converges to... If so, it is considered that an approximate Nash equilibrium solution to the non-cooperative game scheduling problem has been obtained.

3. The cloud manufacturing machine tool service resource scheduling method oriented towards customer benefits as described in claim 1, characterized in that, The improved genetic algorithm is used to solve the cloud manufacturing customer task reward function, specifically as follows: Parameter settings are performed to determine the population size, as well as the maximum and minimum crossover rates and maximum and minimum mutation rates; a dual strategy is used to initialize the population, incorporating personalized customer preferences. Calculate the previous generation's best reward and the current reward; based on the previous generation's best reward and the current reward, calculate the individual fitness value; perform individual selection based on random traversal sampling; dynamically adjust the crossover probability or mutation probability to update the population; after completing the population update, determine whether the individual fitness value is less than or equal to a preset threshold. If yes, output the optimal individual, i.e., the Nash equilibrium solution; if not, continue iterating, while also determining whether the maximum number of iterations has been reached; determine whether the maximum number of iterations has been reached. If yes, output the optimal individual in the current population as the approximate Nash equilibrium solution; if not, increment the iteration count by one and continue calculating the previous generation's best reward and the current reward.

4. The cloud manufacturing machine tool service resource scheduling method for customer benefit as described in claim 3, characterized in that, The dynamic adjustment of crossover or mutation probabilities specifically refers to: In the formula, Indicates the probability of crossover or mutation; Indicates the maximum crossover or mutation probability; Indicates the minimum crossover or mutation probability; This represents the crossover or mutation adjustment factor, which dynamically adjusts the crossover or mutation probability by combining information from the iteration stage and information about algorithm stall. Indicates the crossover adjustment factor; Indicates the adjustment factor for the mutation operation; Indicates the current iteration number; Indicates the maximum number of iterations; This indicates the number of consecutive times the fitness value has not improved. Indicates the crossover and mutation probabilities of the baseline; , and These represent the individual fitness value, the average fitness value, and the maximum fitness value in the current population, respectively. All are constants.

5. The cloud manufacturing machine tool service resource scheduling method for customer benefit as described in claim 3, characterized in that, Before initializing the population using a dual strategy, the chromosomes are first encoded using a two-layer chain structure; each chromosome consists of two parts: a subtask layer and a device layer, and the number of genes in each layer is the total number of subtasks to be processed.

6. The cloud manufacturing machine tool service resource scheduling method for customer benefit as described in claim 5, characterized in that, The number of mutated genes in the subtask layer and device layer is calculated as follows: in, Indicates the number of mutated genes in the subtask layer or device layer; Indicates the maximum number of mutations in the subtask layer gene or the device layer gene; This represents the minimum number of mutations in a subtask layer gene or a device layer gene. Indicates the gene variation adjustment factor of the subtask layer or device layer; Indicates the step size.

7. A cloud-based machine tool service resource scheduling system for customer benefit as described in claim 1, characterized in that, In the process of initializing the population using a dual strategy, the processing parameters are normalized, and the comprehensive score of each device is calculated by combining the customer preference weights. Based on this, the optimal combination of machine tools is assigned to each subtask in turn. For the processing parameters involved in negative indicators, the following formula is used for normalization: For the processing parameters involved in the positive indicators, the following formula is used for normalization: in, This represents the normalized attribute values ​​of each processing parameter. This indicates the attribute value of the parameter corresponding to the selected machine tool equipment. and These represent the maximum and minimum values ​​of the corresponding attributes for each machine tool.

8. A cloud-based machine tool service resource scheduling system oriented towards customer benefits, characterized in that, include: The function building module is configured to: differentiate customer types based on customer personalized preferences, assign weights to key machine tool service indicators according to customer type, and build a cloud manufacturing customer task revenue function. The function solving module is configured to use an improved genetic algorithm to solve the cloud manufacturing customer task revenue function and obtain the optimal machine tool service resource scheduling scheme. The improved genetic algorithm reconstructs the individual fitness function based on non-cooperative game theory, optimizes the machine tool service resource scheduling scheme, and achieves a multi-task revenue Nash equilibrium.

9. A computer-readable storage medium having a program stored thereon, characterized in that, When executed by the processor, the program implements the steps of a cloud manufacturing machine tool service resource scheduling method for customer benefit as described in any one of claims 1-7.

10. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the cloud manufacturing machine tool service resource scheduling method for customer benefit as described in any one of claims 1-7.

Images