Combined process and formula optimization method, device and equipment and storage medium

Abstract

The invention discloses a combined process and formula optimization method and device, equipment and a storage medium. The method comprises the steps of obtaining task input information; on the basis of task input information, generating new formula constituent elements and scheme fragments of a process operation module through cooperative work of the agent group; performing dynamic recombination on the scheme fragments through an automatic knowledge reconstruction engine, and adaptively generating a combined process link and formula structure; and verifying the combined process link and the formula structure, and performing iterative optimization on the agent group behavior according to a verification result. According to the method, the scheme fragments are generated through multi-agent cooperation, the knowledge reconstruction engine is used for dynamic recombination, the combined process link and the formula structure are generated in a self-adaptive mode, iteration optimization is conducted on agent group behaviors according to the verification result of the combined process link and the formula structure, combined innovation of the process and the formula is achieved, and the process and the formula are optimized. And the innovation space dimension and the research and development efficiency of the process and formula combination are expanded.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, in particular to a combined process and formula optimization method, device, equipment and storage medium. BACKGROUND

[0002] At present, in the field of material research and development, formula design and new process development, various optimization algorithms and limited experience rules are mainly relied on to carry out related work. However, with the continuous emergence of new materials and new processes, these traditional methods gradually show some shortcomings, which are difficult to meet the increasing innovation needs. For example, 1. The traditional optimization method can only make small adjustments to the existing formula or process parameters, cannot effectively integrate external multi-source knowledge, and is difficult to build a new design space to realize the "leapfrog innovation" of materials or processes, and is difficult to promote major technological breakthroughs in the industry. 2. Most of the existing AI systems use a single agent architecture. The agent under this architecture is passive in parameter search and is difficult to actively carry out innovation activities, making it difficult to realize the cross-combination of different knowledge fields, limiting the breadth and depth of innovation. 3. Process and formula are in a relatively fragmented state. Most existing methods only optimize a single element in the formula or process, without fully considering the coupling effect between the formula and the process, and cannot face the actual complex industrial system to carry out dynamic combination innovation, making it difficult to meet the comprehensive requirements of modern material production.

[0003] Therefore, there is an urgent need for a combined process and formula optimization method that can realize the combined innovation of process and formula and expand the innovation space dimension and research and development efficiency of process and formula combination. SUMMARY

[0004] The main purpose of the present application is to provide a combined process and formula optimization method, device, equipment and storage medium, which aims to solve the technical problems of insufficient innovation and lack of synergy in the combination of process and formula in the process and formula combination optimization of the prior art.

[0005] To achieve the above-mentioned purpose, the present application provides a combined process and formula optimization method, which comprises the following steps:

[0006] Obtain task input information, the task input information including target performance parameters, available raw material constraints and scalable process types;

[0007] Based on the task input information, generate new formula components and process operation module scheme fragments through the collaborative work of the agent group;

[0008] The scheme fragments are dynamically reorganized by an automatic knowledge reconstruction engine to adaptively generate combined process links and formula structures.

[0009] The combined process link and recipe structure is verified, and the agent population behavior is iteratively optimized according to the verification result.

[0010] Optionally, the agent population includes an innovation agent, an evolution agent, an evaluation agent, a memory agent, and a control agent.

[0011] The step of generating a new recipe constituent element and a scheme fragment of a process operation module through the collaborative work of the agent population based on the task input information includes:

[0012] The innovation agent is configured to generate an initial scheme fragment of a new recipe constituent element or a process operation module based on the task input information and a knowledge base.

[0013] The evolution agent is configured to perform a mutation operation on the initial scheme fragment to generate a mutated scheme fragment.

[0014] The evaluation agent is configured to perform multi-objective pre-evaluation on the initial scheme fragment and the mutated scheme fragment to obtain a target evaluation result.

[0015] The memory agent is configured to store historical scheme fragments and the target evaluation result.

[0016] The control agent is configured to coordinate resource allocation and communication among agents in the agent population, and to select a scheme fragment meeting a preset evaluation requirement from the initial scheme fragment and the mutated scheme fragment according to the target evaluation result.

[0017] Optionally, before the step of performing multi-objective pre-evaluation on the initial scheme fragment and the mutated scheme fragment to obtain a target evaluation result, the method further includes:

[0018] Matching the initial scheme fragment and the mutated scheme fragment based on historical scheme fragments stored in the memory agent to obtain a matching result.

[0019] Filtering the initial scheme fragment and the mutated scheme fragment according to the matching result.

[0020] Optionally, the step of dynamically recombining the scheme fragment through an automatic knowledge reconstruction engine to adaptively generate a combined process link and recipe structure includes:

[0021] Extracting knowledge units of recipe constituent elements and process operation modules from the scheme fragment.

[0022] By the automatic knowledge reconfiguration engine, each knowledge unit is associated according to the task input information, and the associated knowledge units are spliced to adaptively generate a combined process link and a recipe structure.

[0023] Optionally, the step of associating each knowledge unit according to the task input information, and splicing the associated knowledge units to adaptively generate a combined process link and a recipe structure, comprises:

[0024] According to the task input information, cross-domain analogy and logical association of each knowledge unit are performed to obtain the associated knowledge units;

[0025] The associated knowledge units are spliced to adaptively generate a combined process link and a recipe structure.

[0026] Optionally, the step of verifying the combined process link and the recipe structure, and iteratively optimizing the agent group behavior according to the verification result, comprises:

[0027] The combined process link and the recipe structure are screened by using virtual simulation technology to obtain a combined process link and a recipe structure screened by virtual simulation technology;

[0028] The combined process link and the recipe structure screened by virtual simulation technology are pushed to an automated experiment platform for verification to obtain a verification result;

[0029] The agent group behavior is iteratively optimized based on the verification result.

[0030] Optionally, the step of iteratively optimizing the agent group behavior based on the verification result comprises:

[0031] The verification result is fed back to an evaluation agent in the agent group;

[0032] After the evaluation agent receives the verification result, the combined process link and the recipe structure are subjected to multi-dimensional quantitative evaluation to obtain a quantitative evaluation result;

[0033] The agent group behavior is iteratively optimized according to the quantitative evaluation result.

[0034] In addition, in order to achieve the above-mentioned purpose, the application further provides a combined process and recipe optimization device, which comprises:

[0035] An information acquisition module is configured to acquire task input information, wherein the task input information comprises target performance parameters, available raw material constraints, and scalable process types;

[0036] A fragment generation module is configured to generate new recipe elements and process operation module scheme fragments based on the task input information through the collaborative work of the agent group.

[0037] A fragment recombination module is configured to dynamically recombine the scheme fragments through an automatic knowledge reconstruction engine to adaptively generate combined process links and recipe structures.

[0038] A result verification module is configured to verify the combined process links and recipe structures and iteratively optimize the agent group behavior according to the verification result.

[0039] In addition, to achieve the above-mentioned purpose, the application further provides a combined process and recipe optimization device, which comprises a memory, a processor and a combined process and recipe optimization program stored in the memory and executable on the processor, and the combined process and recipe optimization program is configured to implement the steps of the combined process and recipe optimization method as described above.

[0040] In addition, to achieve the above-mentioned purpose, the application further provides a storage medium, which stores a combined process and recipe optimization program, and the combined process and recipe optimization program implements the steps of the combined process and recipe optimization method as described above when executed by a processor.

[0041] The application discloses obtaining task input information, which comprises target performance parameters, available raw material constraints and scalable process types; generating new recipe elements and process operation module scheme fragments based on the task input information through the collaborative work of the agent group; dynamically recombining the scheme fragments through an automatic knowledge reconstruction engine to adaptively generate combined process links and recipe structures; verifying the combined process links and recipe structures and iteratively optimizing the agent group behavior according to the verification result. Compared with the prior art, the application realizes the combined innovation of process and recipe, and expands the innovation space dimension and research and development efficiency of the process and recipe combination. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 It is a flowchart of the first embodiment of the combined process and recipe optimization method of the application.

[0043] Figure 2 It is a whole flowchart frame diagram of the combined process and recipe optimization method of the application.

[0044] Figure 3 Fig. 1 is a flowchart of a first embodiment of the combined process and formula optimization method of the present application;

[0045] Figure 4 Fig. 2 is a structural block diagram of a first embodiment of the combined process and formula optimization device of the present application;

[0046] Figure 5 Fig. 3 is a structural diagram of the combined process and formula optimization equipment of the present application.

[0047] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0048] It should be understood that the specific embodiments described herein are merely intended to explain the present application and not to limit the present application.

[0049] The embodiments of the present application provide a combined process and formula optimization method, which will be described below with reference to Figure 1 , Figure 1 Fig. 1 is a flowchart of a first embodiment of the combined process and formula optimization method of the present application.

[0050] In the present embodiment, the combined process and formula optimization method comprises steps S10-S40:

[0051] Step S10: obtaining task input information, wherein the task input information comprises target performance parameters, available raw material constraints and scalable process types.

[0052] It should be noted that the execution subject of the present embodiment can be a computer server device with data processing, network communication and program running functions applied in material research and development, formula design and new process development scenarios, such as servers, tablet computers, personal computers, etc., or an electronic device (such as a combined process and formula optimization equipment) capable of realizing the above functions. The following will take a system (hereinafter referred to as system) containing a combined process and formula optimization equipment as an example to illustrate the present embodiment and the following embodiments.

[0053] Step S20: based on the task input information, generating new formula components and process operation module scheme fragments through the collaborative work of the agent group.

[0054] It should be noted that the agent group comprises a plurality of agents with different functions. Specifically, the agent group can comprise an innovative agent, an evolutionary agent, an evaluation agent, a memory agent and a control agent.

[0055] It needs to be explained that the innovative agent is used to generate an initial scheme fragment of a new recipe constituent element or process operation module based on the task input information and the knowledge base.

[0056] The evolutionary agent is used to perform mutation operation on the initial scheme fragment to generate a mutated scheme fragment.

[0057] The evaluation agent is used to perform multi-objective pre-evaluation on the initial scheme fragment and the mutated scheme fragment to obtain a target evaluation result.

[0058] The memory agent is used to store historical scheme fragments and the target evaluation result.

[0059] The control agent is used to coordinate resource allocation and communication between agents in the agent group, and select a scheme fragment meeting preset evaluation requirements from the initial scheme fragment and the mutated scheme fragment according to the target evaluation result.

[0060] It needs to be noted that the knowledge base including recipe ingredients, process parameters, quality control indicators and rule constraints exists in each agent in the agent group, which is used to ensure that the node semantics queried by all agents are consistent.

[0061] Further, in order to improve the generation efficiency of the scheme fragment and ensure that the new experience generated by each agent in the agent group immediately benefits subsequent planning, an event stream or a message queue can be used to synchronize trigger signals such as “experiment completed” and “evaluation completed” to each agent, so that the results generated by each agent can be almost real-time visible.

[0062] Further, in order to keep the context simple and improve the communication efficiency between agents, only identifiers corresponding to the results generated by each agent (for example, the identifier can be a graph node ID + abstract) can be exchanged between each agent in the agent group.

[0063] It should be understood that the scheme fragment can refer to a partial innovation unit generated by the multi-agent group, which has not formed a complete solution scheme. Each fragment only solves a certain link of the overall task (for example, only optimizes the coating curing process, or only replaces the toughening agent in the recipe), and needs to be dynamically recombined by the knowledge reconstruction engine to form a complete process-recipe chain (i.e., a combined process chain link and recipe structure). The scheme fragment can specifically represent two types of core elements: recipe constituent elements (i.e., recipe structure, which can include raw material selection, ingredient proportion relationship and recipe hierarchical structure, etc.) and process operation modules (which can include process type, process parameter and process sequence fragment, etc.).

[0064] It can be understood that the knowledge base can be a structured database system supporting dynamic extraction, cross-domain association and reorganization of multi-source knowledge units. The knowledge base can provide the intelligent agent group with evolvable and detachable knowledge units, so that the knowledge reconstruction engine can break through the historical data boundary to generate innovative combinations.

[0065] In a specific implementation, in order to identify and avoid repeated or inefficient scheme fragment generation, before the step of performing multi-objective pre-evaluation on the initial scheme fragment and the mutated scheme fragment to obtain a target evaluation result, it further includes: matching the initial scheme fragment and the mutated scheme fragment based on the historical scheme fragments stored in the memory intelligent agent to obtain a matching result; filtering the initial scheme fragment and the mutated scheme fragment according to the matching result.

[0066] Step S30: dynamically reorganizing the scheme fragment through the automatic knowledge reconstruction engine to adaptively generate a combined process link and a recipe structure.

[0067] It needs to be explained that the automatic knowledge reconstruction engine is a dynamic computing system with cross-domain knowledge association and innovative extrapolation capability. By extracting knowledge units in the scheme fragment, establishing non-explicit logical association, and reorganizing the modular structure according to the task target, it can realize breakthrough combination innovation of process-recipe.

[0068] In a specific implementation, it can be that knowledge units of recipe constituent elements and process operation modules are extracted from the scheme fragment; then through the automatic knowledge reconstruction engine, each of the knowledge units is associated according to the task input information, and the associated knowledge units are spliced to adaptively generate a combined process link and a recipe structure.

[0069] Specifically, the above step of associating each of the knowledge units according to the task input information, and splicing the associated knowledge units to adaptively generate a combined process link and a recipe structure can include: cross-domain analogy and logical association of each of the knowledge units according to the task input information to obtain associated knowledge units; splicing the associated knowledge units to adaptively generate a combined process link and a recipe structure.

[0070] It needs to be noted that splicing the associated knowledge units, i.e., modular splicing the associated knowledge units, can include realizing sequence combination, parallel combination or parameter coupling configuration of multiple different process operation modules, and deducing the response relationship of the adaptive recipe structure with the process link change.

[0071] Step S40: verifying the combined process link and the recipe structure, and iteratively optimizing the behavior of the intelligent agent group according to the verification result.

[0072] In a specific implementation, the combined process link and the formula structure can be verified through virtual simulation technology, the agent group behavior is iteratively optimized according to the verification result, and the knowledge base is updated.

[0073] For example, referring to Figure 2 , Figure 2 is the overall flow framework diagram of the combined process and formula optimization method of the application. Based on the knowledge base or experience data, the task input information (such as designing a new low-cost high-durability coating system) is received, the agent group (that is, the agent group) independently or cooperatively generates innovative fragments (such as new resin introduction + temperature gradient curing + nano additive), and then the combined process route assembly and new formula structure splicing are performed through the knowledge reconstruction engine, and then a plurality of new process-formula chains (that is, the combined process link and the formula structure) are output. Real simulation and automatic test verification are carried out, the corresponding verification results are obtained, the whole process performance is evaluated, if the evaluation is unqualified, the agent group is independently corrected and recombined, if the evaluation is qualified, the knowledge base is output and automatically expanded.

[0074] The embodiment discloses obtaining task input information, the task input information including target performance parameters, available raw material constraints and scalable process types; based on the task input information, a new formula constituent element and a process operation module scheme fragment are generated through collaborative work of an agent group; the scheme fragment is dynamically reorganized through an automatic knowledge reconstruction engine to adaptively generate a combined process link and a formula structure; the combined process link and the formula structure are verified, and the agent group behavior is iteratively optimized according to the verification result. Compared with the prior art, the embodiment realizes the combined innovation of process and formula, and expands the innovation space dimension and research and development efficiency of the process and formula combination.

[0075] For example, referring to Figure 3 , Figure 3 is the flow schematic diagram of the second embodiment of the combined process and formula optimization method of the application.

[0076] Based on the above first embodiment, in the embodiment, the step S40 includes steps S401-S403:

[0077] Step S401: The combined process link and the formula structure are screened through virtual simulation technology to obtain the combined process link and the formula structure screened through virtual simulation technology.

[0078] Step S402: push the combined process link and the recipe structure screened by the virtual simulation technology to an automated experiment platform for verification, and obtain a verification result.

[0079] Step S403: iteratively optimize the agent group behavior based on the verification result.

[0080] In a specific implementation, the verification result can also be fed back to an evaluation agent in the agent group; after the evaluation agent receives the verification result, the evaluation agent performs multi-dimensional quantitative evaluation on the combined process link and the recipe structure, and obtains a quantitative evaluation result; and iteratively optimizes the agent group behavior according to the quantitative evaluation result and updates the knowledge base.

[0081] It should be noted that iteratively optimizing the agent group behavior can include adjusting the generation strategy of the innovation agent, the mutation rule of the evolution agent, or the experience weight of the memory agent, and the like.

[0082] It should be explained that the above multi-dimensional quantitative evaluation on the combined process link and the recipe structure can be multi-dimensional quantitative evaluation on performance, process cost, and environmental impact of the combined process link and the recipe structure, wherein the performance, the process cost, and the environmental impact can be respectively assigned different weight coefficients.

[0083] It should be understood that the embodiment shortens the verification period and reduces the research and development cost by mixing the verification mechanism, integrating virtual simulation and automated experiment. In addition, the verification result is fed back to the agent group in real time, driving dynamic adjustment of a subsequent scheme generation strategy, and forming a "innovation-verification-evolution" closed loop.

[0084] The embodiment discloses screening the combined process link and the recipe structure by using a virtual simulation technology, obtaining a combined process link and a recipe structure screened by the virtual simulation technology, pushing the combined process link and the recipe structure screened by the virtual simulation technology to an automated experiment platform for verification, obtaining a verification result, and iteratively optimizing the agent group behavior based on the verification result. Compared with the prior art, the embodiment integrates virtual simulation and automated experiment, shortens the verification period, and reduces the research and development cost.

[0085] In addition, the embodiment of the application also proposes a storage medium, and the storage medium stores a combined process and recipe optimization program. When the combined process and recipe optimization program is executed by a processor, the steps of the combined process and recipe optimization method described above are implemented.

[0086] Reference Figure 4 , Figure 4 is a structural block diagram of a first embodiment of a combined process and recipe optimization device of the application.

[0087] As shown in Figure 4 The combined process and formula optimization device comprises an information acquisition module 501, a segment generation module 502, a segment reorganization module 503, and a result verification module 504.

[0088] The information acquisition module 501 is configured to acquire task input information, wherein the task input information comprises target performance parameters, available raw material constraints, and expandable process types.

[0089] The segment generation module 502 is configured to generate new formula elements and process operation module scheme segments through collaborative work of an agent group based on the task input information.

[0090] The segment reorganization module 503 is configured to dynamically reorganize the scheme segments through an automatic knowledge reconstruction engine to adaptively generate combined process links and formula structures.

[0091] The result verification module 504 is configured to verify the combined process links and formula structures, and iteratively optimize behaviors of the agent group according to a verification result.

[0092] The device embodiment discloses acquiring task input information, wherein the task input information comprises target performance parameters, available raw material constraints, and expandable process types; generating new formula elements and process operation module scheme segments through collaborative work of an agent group based on the task input information; dynamically reorganizing the scheme segments through an automatic knowledge reconstruction engine to adaptively generate combined process links and formula structures; and verifying the combined process links and formula structures, and iteratively optimizing behaviors of the agent group according to a verification result. Compared with the prior art, the device embodiment realizes combined innovation of processes and formulas, and expands innovation space dimensions and research and development efficiency of process and formula combination, because the device embodiment generates scheme segments through multi-agent collaboration, adaptively generates combined process links and formula structures through dynamic reorganization of the knowledge reconstruction engine, and iteratively optimizes behaviors of the agent group according to a verification result of the combined process links and formula structures.

[0093] Based on the first embodiment of the above-mentioned combined process and formula optimization device, the second embodiment of the combined process and formula optimization device is proposed.

[0094] In the embodiment, the result verification module 504 is further configured to screen the combined process link and recipe structure by using a virtual simulation technology, obtain a combined process link and recipe structure that passes the virtual simulation technology screening, push the combined process link and recipe structure that passes the virtual simulation technology screening to an automated experiment platform for verification, and obtain a verification result; and iteratively optimize the agent group behavior based on the verification result.

[0095] Other embodiments or specific implementations of the combined process and recipe optimization apparatus can refer to the above-mentioned method embodiments, and will not be described here.

[0096] The present application provides a combined process and recipe optimization device, which comprises at least one processor and a memory in communication with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the combined process and recipe optimization method in the above-mentioned embodiment one.

[0097] Reference will be made to the following Figure 5 , which shows a structural diagram of a combined process and recipe optimization device suitable for implementing the embodiments of the present application. The combined process and recipe optimization device in the embodiments of the present application can include, but is not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), and the like, and fixed terminals such as digital TVs, desktop computers, and the like. Figure 5 The combined process and recipe optimization device shown is only an example, and should not impose any limitation on the functions and use range of the embodiments of the present application.

[0098] As Figure 5As shown, the combined process and recipe optimization device can include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.) that can perform various appropriate actions and processes according to programs stored in a read-only memory 1002 or loaded from a storage device 1003 into a random access memory 1004. Various programs and data required for the operation of the combined process and recipe optimization device are also stored in the random access memory 1004. The processing device 1001, the read-only memory 1002, and the random access memory 1004 are connected to each other by a bus 1005. An input / output interface 1006 is also connected to the bus. Generally, the following systems can be connected to the input / output interface 1006: input devices 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; the storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the combined process and recipe optimization device to communicate with other devices wirelessly or by wire to exchange data. Although the combined process and recipe optimization device with various systems is shown in the figure, it should be understood that all the shown systems are not required to be implemented or possessed. More or fewer systems can be alternatively implemented or possessed.

[0099] In particular, according to embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as a computer software program. For example, embodiments of the present disclosure include a computer program product comprising a computer program carried on a computer readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network by a communication device, or installed from the storage device 1003, or installed from the read-only memory 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the methods of the embodiments of the present disclosure are performed.

[0100] The combined process and recipe optimization device provided by the present disclosure adopts the combined process and recipe optimization method in the above embodiments, and can solve the technical problems of insufficient innovation and lack of synergy of process and recipe combination in the process and recipe combination optimization in the prior art. Compared with the prior art, the combined process and recipe optimization device provided by the present disclosure has the same beneficial effects as the combined process and recipe optimization method provided by the above embodiments, and other technical features in the combined process and recipe optimization device are the same as the features disclosed in the previous embodiment method, which will not be repeated here.

[0101] It should be understood that various parts of the present application can be realized in hardware, software, firmware, or a combination thereof. In the above description of embodiments, specific functional, structural, material or characteristic features are combined in a manner that is appropriate for the particular embodiment.

[0102] The above description is merely illustrative of the application and is not intended to limit the scope of the application. Any variations and modifications that are obvious to those skilled in the art are intended to be within the scope of the application. The scope of the application is defined by the claims.

[0103] It should be noted that the terms "comprising", "including", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that processes, methods, articles, or systems that comprise a list of elements do not include only those elements recited, but can also include other elements not expressly listed or inherent to such processes, methods, articles, or systems. Without further limitation, an element defined by the phrase "comprising a" does not exclude the presence of additional identical elements in the process, method, article, or system that includes the element.

[0104] The above-mentioned embodiment numbers of the present application are only for description and do not represent the advantages or disadvantages of the embodiments.

[0105] From the above description of the embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be realized by means of software and necessary general hardware platforms, of course, they can also be realized by hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solutions of the present application or the parts that contribute to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access memory, a magnetic disk, an optical disk) and includes a number of instructions to make a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) execute the methods described in various embodiments of the present application.

[0106] The above is only the preferred embodiment of the present application, and does not limit the scope of the present application, and any equivalent structure or equivalent process transformation, or direct or indirect application in other related technical fields, is also included in the protection scope of the present application.

Claims

1. A combined process and recipe optimization method, characterized by, The method comprises: acquiring task input information, the task input information comprising target performance parameters, available raw material constraints, and scalable process types; generating new recipe components and process operation module scheme fragments through collaborative work of an agent group based on the task input information; dynamically recombining the scheme fragments through an automatic knowledge reconfiguration engine to adaptively generate combined process links and recipe structures; verifying the combined process links and recipe structures and iteratively optimizing the agent group behavior according to the verification results.

2. The combined process and recipe optimization method of claim 1, wherein, The agent group comprises innovative agents, evolutionary agents, evaluation agents, memory agents, and control agents; The step of generating new recipe components and process operation module scheme fragments through collaborative work of an agent group based on the task input information comprises: The innovative agent is configured to generate initial scheme fragments of new recipe components or process operation modules based on the task input information and a knowledge base; The evolutionary agent is configured to perform mutation operations on the initial scheme fragments to generate mutated scheme fragments; The evaluation agent is configured to perform multi-objective pre-evaluation on the initial scheme fragments and the mutated scheme fragments to obtain target evaluation results; The memory agent is configured to store historical scheme fragments and the target evaluation results; The control agent is configured to coordinate resource allocation and communication among agents in the agent group and select scheme fragments meeting preset evaluation requirements from the initial scheme fragments and the mutated scheme fragments according to the target evaluation results.

3. The modular process and recipe optimization method of claim 2, wherein, Before the step of performing multi-objective pre-evaluation on the initial scheme fragments and the mutated scheme fragments to obtain target evaluation results, the method further comprises: matching the initial scheme fragments and the mutated scheme fragments based on historical scheme fragments stored in the memory agent to obtain matching results; filtering the initial scheme fragments and the mutated scheme fragments according to the matching results.

4. The modular process and recipe optimization method of claim 1, wherein, The step of dynamically recombining the scheme fragments through an automatic knowledge reconfiguration engine to adaptively generate combined process links and recipe structures comprises: extracting knowledge units of recipe components and process operation modules from the scheme fragments; associating the knowledge units according to task input information through an automatic knowledge reconfiguration engine, splicing the associated knowledge units, and adaptively generating combined process links and recipe structures.

5. The modular process and recipe optimization method of claim 4, wherein, The step of associating the knowledge units according to task input information and splicing the associated knowledge units to adaptively generate combined process links and recipe structures comprises: cross-domain analogy and logical association of the knowledge units according to task input information to obtain associated knowledge units; splicing the associated knowledge units to adaptively generate combined process links and recipe structures.

6. The modular process and recipe optimization method of claim 1, wherein, The step of verifying the combined process links and recipe structures and iteratively optimizing the agent group behavior according to the verification results comprises: Screening the combined process link and recipe structure by using virtual simulation technology to obtain the combined process link and recipe structure screened by the virtual simulation technology; Pushing the combined process link and recipe structure screened by the virtual simulation technology to an automatic experiment platform for verification to obtain a verification result; Iteratively optimizing the agent group behavior based on the verification result.

7. The modular process and recipe optimization method of claim 6, wherein, The step of iteratively optimizing the agent group behavior based on the verification result includes: feeding back the verification result to an evaluation agent in the agent group; after the evaluation agent receives the verification result, performing multi-dimensional quantitative evaluation on the combined process link and recipe structure to obtain a quantitative evaluation result; iteratively optimizing the agent group behavior according to the quantitative evaluation result.

8. A combined process and recipe optimization apparatus, characterized by, The device includes: an information acquisition module configured to acquire task input information, the task input information including target performance parameters, available raw material constraints, and scalable process types; a segment generation module configured to generate new recipe constituent elements and process operation module scheme segments by collaborative work of an agent group based on the task input information; a segment recombination module configured to dynamically recombine the scheme segments by an automatic knowledge reconstruction engine to adaptively generate a combined process link and recipe structure; a result verification module configured to verify the combined process link and recipe structure and iteratively optimize the agent group behavior according to a verification result.

9. A combined process and recipe optimization apparatus, characterized by, The device includes a memory, a processor, and a combined process and recipe optimization program stored on the memory and executable on the processor, the combined process and recipe optimization program being configured to implement the steps of the combined process and recipe optimization method according to any one of claims 1 to 7.

10. A storage medium, characterized by The storage medium stores a combined process and recipe optimization program, and the combined process and recipe optimization program, when executed by a processor, implements the steps of the combined process and recipe optimization method according to any one of claims 1 to 7.

Images