Artificial intelligence quantification evaluation and autonomous evolution method and device based on parallel system

Abstract

The application provides an artificial intelligence quantitative evaluation and autonomous evolution method and device based on a parallel system. The method comprises steps 1 to 7. The application can objectively quantitatively evaluate the intelligent level of a power grid regulation intelligent system, facilitate power practitioners to compare and optimize the intelligent levels of different intelligent systems, and promote the autonomous evolution of the intelligent system in the direction of intelligent level improvement, and autonomously find an intelligent system with a better intelligent level and a better regulation effect on a power grid.

Description

Technical Field

[0001] The embodiments of the present invention relate to the fields of artificial intelligence and power system technology, and in particular to a method and device for quantitative evaluation and autonomous evolution of artificial intelligence based on parallel systems. Background Technology

[0002] Using intelligent agents to solve power system regulation problems is an important research topic in the practical application of AI technology, with the application of power grid correction and control agents in the power grid being a significant branch of this research. To optimize agent design and generate more human-like agents, it is necessary to compare the strengths and weaknesses of different agents and to compare the similarity between different agents and between agents and humans. However, to date, it is difficult to evaluate and compare intelligent agents using a unified standard.

[0003] Intelligent agent evaluation refers to methods for describing the intelligence level of an agent through subjective or objective measurements; intelligent quantitative evaluation, on the other hand, is an objective quantitative measurement of the intelligence level of an agent, describing the measurement results in a quantitative form. As one of the classic algorithms for building intelligent agents, reinforcement learning (RL) faces the following problems: reinforcement learning struggles to quickly explore intelligent agent models that perform well in real-world environments; related algorithms are fragile; and manually finding moderately strong hyperparameters is already very expensive for problems in complex domains such as power grids; in domains requiring large amounts of computation, the flexibility of reinforcement learning algorithms needs to be improved. Currently, these problems can be solved using automated reinforcement learning (AutoRL) technology. AutoRL technology is an automated machine learning (AutoML) technology specifically designed for the evolution of RL agents, aiming to achieve autonomous evolution of the agent by autonomously adjusting its network architecture, hyperparameters, and algorithms. AutoML is a data-driven intelligent agent generation and optimization system that automates the entire machine learning process, significantly reducing the development cost of agents while obtaining high-performing agents. Common optimization search methods used in AutoRL technology include grid search, random search, evolutionary algorithms, and Bayesian optimization. Bayesian optimization, as a widely used method for autonomous system evolution, is applied in industry and various scientific experiments due to its sequential decision-making concept. However, it faces the same difficulty in understanding and interpreting the evolutionary results using AutoRL technology as intelligent agents. Therefore, developing a method and device for quantitative evaluation and autonomous evolution of artificial intelligence based on parallel systems, which can effectively overcome the shortcomings of the aforementioned related technologies, has become a pressing technical problem for the industry. Summary of the Invention

[0004] To address the aforementioned problems in existing technologies, embodiments of the present invention provide a method and apparatus for quantitative evaluation and autonomous evolution of artificial intelligence based on parallel systems.

[0005] In a first aspect, embodiments of the present invention provide a method for quantitative evaluation and autonomous evolution of artificial intelligence based on parallel systems, comprising: Step 1: constructing an intelligent quantitative evaluation index system and selecting an intelligent index m from it; Step 2: setting and generating a test scenario n based on the selected intelligent index m. m Step 3: Based on the indicator scoring formula, calculate the test scenario n of the tested intelligent agent in terms of intelligent indicator m. m The intelligence level in the middle is quantitatively scored. m,n Step 4: Construct an intelligent quantitative evaluation function and calculate the comprehensive intelligence level of the tested intelligent agent; Step 5: Construct an objective function for autonomous optimization evolution based on intelligent evaluation; Step 6: Based on the parallel system as the basic platform, and according to the objective function, introduce intelligent quantitative evaluation to construct a basic framework for autonomous optimization evolution of the intelligent agent based on intelligent evaluation; Step 7: Based on the basic framework for autonomous evolution, construct an autonomous optimization evolution system based on intelligent evaluation.

[0006] Based on the above method embodiments, the artificial intelligence quantitative evaluation and autonomous evolution method based on parallel systems provided in this embodiment of the invention specifically includes the following steps in step 1: Step 1.1: Dividing machine intelligence into 35 indicators; Step 1.2: According to the principles of representativeness and comprehensiveness, classifying and summarizing the 35 indicators based on their similarity and superiority to humans, task completion, and the intelligence level that the intelligent agent should possess; Step 1.3: Constructing an intelligence quantitative evaluation indicator system based on the classification and summary of the 35 intelligence indicators; Step 1.4: Based on the tested intelligent agent T... (k) The specific tasks to be performed and the capabilities required are determined by selecting appropriate intelligent indicators m from the intelligent quantitative evaluation indicator system constructed in step 1.3 for subsequent intelligent evaluation.

[0007] Based on the above method embodiments, the AI ​​quantitative evaluation and autonomous evolution method based on parallel systems provided in this embodiment of the invention specifically includes the following steps in step 2: Step 2.1: Based on the selected intelligence index m, clarify the characteristics of each test scenario and the requirements for each test scenario; Step 2.2: Compare the training scenarios of the tested intelligent agent, and based on the characteristics and requirements of the test scenarios, generate test scenarios n of intelligence index m in the parallel system. m .

[0008] Based on the above method embodiments, the artificial intelligence quantitative evaluation and autonomous evolution method based on parallel systems provided in this embodiment of the invention specifically includes steps 3 and 4 as follows: Step 3.1: In the test scenario generated by the parallel system, the tested intelligent agent is tested and run to obtain the observed Y of its running state in the test scenario. t A,(k)Step 3.2: Based on the tasks performed and problems solved by the tested agent, select or construct appropriate quantitative scoring formulas, and use these scoring formulas with the help of Y... t A,(k) The computational agent scores E in the test scenario. m,n Step 3.3: Using the entropy weight method, through E m,n Calculate or select the weight w of each scoring function in the smart metric. n Step 3.4: Using the weighted summation method shown in equation (1), and employing E m,n and w n Calculate the test score E for each indicator. m (Y t A,(k) ,T (k) ,n m ),

[0009]

[0010] Where n is the indicator scoring formula, N is the total number of indicator scoring formulas, and E' m,n It is the normalized E m,n ;

[0011] Step 4.1: Calculate the weight w of each intelligent indicator using the entropy weight method. m , or take w m =1 / M, where M is the total number of selected intelligence indicators; Step 4.2: Using (2), calculate the comprehensive intelligence level I of the agent. (k) ,

[0012]

[0013] Based on the above method embodiments, the artificial intelligence quantitative evaluation and autonomous evolution method based on parallel systems provided in this embodiment of the invention specifically includes step 5: Step 5.1: Analyze and summarize the objective function of AutoRL, combine it with intelligent quantitative evaluation, and establish the objective function of autonomous optimization evolution based on intelligent evaluation as shown in equations (3) and (4):

[0014] T * =argmax 1≤k≤K I (k) (T (k) (p (k) ),n m (3)

[0015] p * =argmax 1≤k≤K J(T (k) (p (k) (4)

[0016] Among them, T * and p * They represent the optimal comprehensive intelligence level I, respectively. * The corresponding optimal agent and its model parameters and hyperparameters; p (k) The agent T generated in each iteration (k) The model parameters and hyperparameters, and their combined intelligence level is I. (k) ; (k) represents the number of iterations during agent evolution; J is the reward function for reinforcement learning.

[0017] Based on the above method embodiments, the artificial intelligence quantitative evaluation and autonomous evolution method based on parallel systems provided in this embodiment of the invention specifically includes step 6: Step 6.1: Based on the management and control process of the parallel system, obtain the mathematical expression of the autonomous evolution process based on intelligent evaluation, as shown in equation (5):

[0018]

[0019] in, This represents the artificial system control quantity of the virtual and real systems in a parallel system. This represents the actual system control quantity of the virtual and real systems in a parallel system;

[0020] Step 6.2: Since the evaluation and evolution of agents only consider virtual systems within parallel systems, and By T (k) T (k) Control by p (k) The basic framework for autonomous optimization evolution based on intelligent evaluation is obtained from equation (5), as shown in equation (6):

[0021]

[0022] in, The optimized agent model and hyperparameters are given.

[0023] Based on the above method embodiments, the AI ​​quantitative evaluation and autonomous evolution method based on parallel systems provided in this embodiment of the invention includes the following steps in step 7: Step 7.1: Selecting a Bayesian optimization algorithm for an agent and establishing or selecting a suitable AutoRL system; Step 7.2: Based on the basic framework obtained in step 6, embedding the intelligent quantitative evaluation method into the AutoRL system selected in step 7.1 to construct an autonomous optimization evolution system based on intelligent evaluation, wherein the autonomous optimization evolution system based on intelligent evaluation is an intelligent evaluation and autonomous evolution system based on parallel systems.

[0024] Secondly, embodiments of the present invention provide an artificial intelligence quantitative evaluation and autonomous evolution device based on a parallel system, comprising: a first main module, used to implement step 1: constructing an intelligent quantitative evaluation index system and selecting an intelligent index m from it; step 2: setting and generating a test scenario n based on the selected intelligent index m. m The second main module is used to implement step 3: based on the indicator scoring formula, calculate the test scenario n of the tested intelligent agent in terms of intelligent indicator m. m The intelligence level in the middle is quantitatively scored. m,n Step 4: Construct an intelligent quantitative evaluation function and calculate the comprehensive intelligence level of the tested intelligent agent; The third main module is used to implement Step 5: Construct an objective function for autonomous optimization evolution based on intelligent evaluation; Step 6: Based on the parallel system as the basic platform, and according to the objective function, introduce intelligent quantitative evaluation to construct a basic framework for autonomous optimization evolution of the intelligent agent based on intelligent evaluation; The fourth main module is used to implement Step 7: Construct an autonomous optimization evolution system based on intelligent evaluation according to the basic framework of autonomous evolution.

[0025] Thirdly, embodiments of the present invention provide an electronic device, comprising:

[0026] At least one processor; and

[0027] At least one memory communicatively connected to the processor, wherein:

[0028] The memory stores program instructions that can be executed by the processor. The processor can call the program instructions to execute the artificial intelligence quantitative evaluation and autonomous evolution method based on parallel systems provided by any of the various implementations of the first aspect.

[0029] Fourthly, embodiments of the present invention provide a non-transitory computer-readable storage medium storing computer instructions that cause a computer to execute a parallel system-based artificial intelligence quantitative evaluation and autonomous evolution method provided by any of the various implementations of the first aspect.

[0030] The artificial intelligence quantitative evaluation and autonomous evolution method and equipment based on parallel systems provided in this invention can objectively and quantitatively evaluate the intelligence level of power grid control intelligent systems, which facilitates power practitioners to compare and select the intelligence level of different intelligent systems. It can also promote the autonomous evolution of intelligent systems with the goal of improving intelligence level, and autonomously find intelligent systems with better intelligence level and better power grid control effect. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 A flowchart of an artificial intelligence quantitative evaluation and autonomous evolution method based on parallel systems provided in this embodiment of the invention;

[0033] Figure 2 A schematic diagram of the structure of an artificial intelligence quantitative evaluation and autonomous evolution device based on a parallel system provided in an embodiment of the present invention;

[0034] Figure 3 A schematic diagram of the physical structure of an electronic device provided in an embodiment of the present invention;

[0035] Figure 4 This is a schematic diagram of the basic process of intelligent quantitative evaluation provided in the embodiments of the present invention;

[0036] Figure 5 This is a schematic diagram illustrating the effect of constructing an intelligent quantitative evaluation index system provided in an embodiment of the present invention. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. In addition, the technical features of the various embodiments or individual embodiments provided by the present invention can be arbitrarily combined with each other to form feasible technical solutions. Such combinations are not constrained by the order of steps and / or structural composition patterns, but must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0038] This invention provides a method for quantitative evaluation and autonomous evolution of artificial intelligence based on parallel systems. See [link to relevant documentation]. Figure 1 The method includes: Step 1: Constructing an intelligent quantitative evaluation index system and selecting intelligent index m from it; Step 2: Setting and generating test scenarios n based on the selected intelligent index m. m Step 3: Based on the indicator scoring formula, calculate the test scenario n of the tested intelligent agent in terms of intelligent indicator m.m The intelligence level in the middle is quantitatively scored. m,n Step 4: Construct an intelligent quantitative evaluation function and calculate the comprehensive intelligence level of the tested intelligent agent; Step 5: Construct an objective function for autonomous optimization evolution based on intelligent evaluation; Step 6: Based on the parallel system as the basic platform, and according to the objective function, introduce intelligent quantitative evaluation to construct a basic framework for autonomous optimization evolution of the intelligent agent based on intelligent evaluation; Step 7: Based on the basic framework for autonomous evolution, construct an autonomous optimization evolution system based on intelligent evaluation.

[0039] This invention also provides an artificial intelligence quantitative evaluation and autonomous evolution system based on parallel systems, comprising: an intelligent quantitative evaluation module, implemented by steps 1 to 4; and an autonomous evolution module guided by the intelligent evaluation results, implemented by steps 5 to 7.

[0040] See Figure 4 Based on the above method embodiments, as an optional embodiment, the artificial intelligence quantitative evaluation and autonomous evolution method based on parallel systems provided in this embodiment of the invention specifically includes the following steps in step 1: Step 1.1: Dividing machine intelligence into 35 indicators; Step 1.2: According to the principles of representativeness and comprehensiveness, classifying and summarizing the 35 indicators based on their similarity and superiority to humans, task completion, and the intelligence level that the intelligent agent should possess; Step 1.3: Based on the classification and summary of the 35 intelligence indicators, constructing an intelligence quantitative evaluation indicator system. The intelligence quantitative evaluation indicator system can be found in [reference needed]. Figure 5 Step 1.4: Based on the tested intelligent agent T (k) The specific tasks to be performed and the capabilities required are determined by selecting appropriate intelligent indicators m from the intelligent quantitative evaluation indicator system constructed in step 1.3 for subsequent intelligent evaluation.

[0041] Based on the above method embodiments, as an optional embodiment, the AI ​​quantitative evaluation and autonomous evolution method based on parallel systems provided in this embodiment of the invention specifically includes the following steps in step 2: Step 2.1: Based on the selected intelligence index m, clarify the characteristics of each test scenario and the requirements for each test scenario; Step 2.2: Compare the training scenarios of the tested intelligent agent, and based on the characteristics and requirements of the test scenarios, generate test scenarios n of intelligence index m in the parallel system. m .

[0042] Based on the above method embodiments, as an optional embodiment, the artificial intelligence quantitative evaluation and autonomous evolution method based on parallel systems provided in this embodiment of the invention specifically includes steps 3 and 4 as follows: Step 3.1: In the test scenario generated by the parallel system, the tested intelligent agent is tested and run to obtain the observed Y of its running state in the test scenario. tA,(k) Step 3.2: Based on the tasks performed and problems solved by the tested agent, select or construct appropriate quantitative scoring formulas, and use these scoring formulas with the help of Y... t A,(k) The computational agent scores E in the test scenario. m,n Step 3.3: Using the entropy weight method, through E m,n Calculate or select the weight w of each scoring function in the smart metric. n Step 3.4: Using the weighted summation method shown in equation (1), and employing E m,n and w n Calculate the test score E for each indicator. m (Y t A,(k) ,T (k) ,n m ),

[0043]

[0044] Where n is the indicator scoring formula, N is the total number of indicator scoring formulas, and E' m,n It is the normalized E m,n ;

[0045] Step 4.1: Calculate the weight w of each intelligent indicator using the entropy weight method. m , or take w m =1 / M, where M is the total number of selected intelligence indicators; Step 4.2: Using (2), calculate the comprehensive intelligence level I of the agent. (k) ,

[0046]

[0047] Based on the above method embodiments, as an optional embodiment, the artificial intelligence quantitative evaluation and autonomous evolution method based on parallel systems provided in this embodiment of the invention specifically includes the following step 5: Step 5.1: Analyze and summarize the objective function of AutoRL, and combine it with intelligent quantitative evaluation to establish an objective function for autonomous optimization evolution based on intelligent evaluation, as shown in equations (3) and (4):

[0048] T * =argmax 1≤k≤K I (k) (T (k) (p (k) ),n m (3)

[0049] p * =argmax 1≤k≤K J(T (k) (p (k) (4)

[0050] Among them, T * and p * They represent the optimal comprehensive intelligence level I, respectively. * The corresponding optimal agent and its model parameters and hyperparameters; p (k) The agent T generated in each iteration (k) The model parameters and hyperparameters, and their combined intelligence level is I. (k) ; (k) represents the number of iterations during agent evolution; J is the reward function for reinforcement learning.

[0051] Based on the above method embodiments, as an optional embodiment, the artificial intelligence quantitative evaluation and autonomous evolution method based on parallel systems provided in this embodiment of the invention specifically includes step 6 as follows: Step 6.1: Based on the management and control process of the parallel system, obtain the mathematical expression of the autonomous evolution process based on intelligent evaluation, as shown in equation (5):

[0052]

[0053] in, This represents the artificial system control quantity of the virtual and real systems in a parallel system. This represents the actual system control quantity of the virtual and real systems in a parallel system;

[0054] Step 6.2: Since the evaluation and evolution of agents only consider virtual systems within parallel systems, and By T (k) T (k) Control by p (k) The basic framework for autonomous optimization evolution based on intelligent evaluation is obtained from equation (5), as shown in equation (6):

[0055]

[0056] in, The optimized agent model and hyperparameters are given.

[0057] Based on the above method embodiments, as an optional embodiment, the AI ​​quantitative evaluation and autonomous evolution method based on parallel systems provided in this embodiment of the invention specifically includes step 7 as follows: Step 7.1: Selecting a Bayesian optimization algorithm for an agent and establishing or selecting a suitable AutoRL system; Step 7.2: Based on the basic framework obtained in step 6, embedding the intelligent quantitative evaluation method into the AutoRL system selected in step 7.1 to construct an autonomous optimization evolution system based on intelligent evaluation, wherein the autonomous optimization evolution system based on intelligent evaluation is an intelligent evaluation and autonomous evolution system based on parallel systems.

[0058] The artificial intelligence quantitative evaluation and autonomous evolution method based on parallel systems provided in this invention can objectively and quantitatively evaluate the intelligence level of power grid control intelligent systems, which facilitates power practitioners to compare and select the intelligence level of different intelligent systems. It can also promote the autonomous evolution of intelligent systems with the goal of improving intelligence level, and autonomously find intelligent systems with better intelligence level and better power grid control effect.

[0059] The implementation of the various embodiments of the present invention is based on programmed processing through a device with processor functionality. Therefore, in practical engineering, the technical solutions and functions of the various embodiments of the present invention can be encapsulated into various modules. Based on this reality, and building upon the above embodiments, the embodiments of the present invention provide an artificial intelligence quantitative evaluation and autonomous evolution device based on parallel systems. This device is used to execute the artificial intelligence quantitative evaluation and autonomous evolution method based on parallel systems in the above method embodiments. See also... Figure 2 The device includes: a first main module, used to implement step 1: constructing an intelligent quantitative evaluation index system and selecting intelligent index m from it; step 2: setting and generating test scenario n based on the selected intelligent index m. m The second main module is used to implement step 3: based on the indicator scoring formula, calculate the test scenario n of the tested intelligent agent in terms of intelligent indicator m. m The intelligence level in the middle is quantitatively scored. m,n Step 4: Construct an intelligent quantitative evaluation function and calculate the comprehensive intelligence level of the tested intelligent agent; The third main module is used to implement Step 5: Construct an objective function for autonomous optimization evolution based on intelligent evaluation; Step 6: Based on the parallel system as the basic platform, and according to the objective function, introduce intelligent quantitative evaluation to construct a basic framework for autonomous optimization evolution of the intelligent agent based on intelligent evaluation; The fourth main module is used to implement Step 7: Construct an autonomous optimization evolution system based on intelligent evaluation according to the basic framework of autonomous evolution.

[0060] The artificial intelligence quantitative evaluation and autonomous evolution device based on parallel systems provided in this invention adopts... Figure 2 Several modules within it can objectively and quantitatively evaluate the intelligence level of the power grid control intelligent system, facilitating power industry professionals to compare and select the best intelligence level of different intelligent systems. Furthermore, it can promote the autonomous evolution of intelligent systems by focusing on improving intelligence levels, and autonomously find intelligent systems with better intelligence levels and better power grid control effects.

[0061] It should be noted that the apparatus in the device embodiments provided by the present invention can be used not only to implement the methods in the above method embodiments, but also to implement the methods in other method embodiments provided by the present invention. The difference lies only in the setting of corresponding functional modules. Its principle is basically the same as that of the above device embodiments provided by the present invention. As long as those skilled in the art, based on the above device embodiments and referring to the specific technical solutions in other method embodiments, obtain corresponding technical means and technical solutions composed of these technical means by combining technical features, and improve the apparatus in the above device embodiments while ensuring the practicality of the technical solutions, they can obtain corresponding device-type embodiments for implementing the methods in other method-type embodiments. For example:

[0062] Based on the above-described device embodiments, as an optional embodiment, the artificial intelligence quantitative evaluation and autonomous evolution device based on parallel systems provided in this embodiment of the invention further includes: a first sub-module, used to implement step 1, specifically including: step 1.1: dividing machine intelligence into 35 indicators; step 1.2: classifying and summarizing the 35 indicators according to the principles of representativeness and comprehensiveness, based on their similarity and superiority to humans, task completion, and the intelligence level that the intelligent agent should possess; step 1.3: constructing an intelligence quantitative evaluation indicator system based on the classification and summary of the 35 intelligence indicators; step 1.4: based on the tested intelligent agent T... (k) The specific tasks to be performed and the capabilities required are determined by selecting appropriate intelligent indicators m from the intelligent quantitative evaluation indicator system constructed in step 1.3 for subsequent intelligent evaluation.

[0063] Based on the above-described device embodiments, as an optional embodiment, the artificial intelligence quantitative evaluation and autonomous evolution device based on parallel systems provided in this embodiment of the invention further includes: a second sub-module, used to implement step 2, specifically including: step 2.1: based on the selected intelligence index m, clarifying the characteristics of each test scenario and the requirements for each test scenario; step 2.2: comparing with the training scenario of the tested intelligent agent, and based on the characteristics and requirements of the test scenario, generating test scenario n of intelligence index m in the parallel system. m .

[0064] Based on the above device embodiments, as an optional embodiment, the artificial intelligence quantitative evaluation and autonomous evolution device based on parallel systems provided in this embodiment of the invention further includes: a third submodule, used to implement steps 3 and 4, specifically including: step 3.1: in the test scenario generated by the parallel system, the tested intelligent agent is tested and run, and the observed Y of its running state in the test scenario is obtained. t A,(k)Step 3.2: Based on the tasks performed and problems solved by the tested agent, select or construct appropriate quantitative scoring formulas, and use these scoring formulas with the help of Y... t A,(k) The computational agent scores E in the test scenario. m,n Step 3.3: Using the entropy weight method, through E m,n Calculate or select the weight w of each scoring function in the smart metric. n Step 3.4: Using the weighted summation method shown in equation (1), and employing E m,n and w n Calculate the test score E for each indicator. m (Y t A,(k) ,T (k) ,n m ),

[0065]

[0066] Where n is the indicator scoring formula, N is the total number of indicator scoring formulas, and E' m,n It is the normalized E m,n ;

[0067] Step 4.1: Calculate the weight w of each intelligent indicator using the entropy weight method. m , or take w m =1 / M, where M is the total number of selected intelligence indicators; Step 4.2: Using (2), calculate the comprehensive intelligence level I of the agent. (k) ,

[0068]

[0069] Based on the above device embodiments, as an optional embodiment, the artificial intelligence quantitative evaluation and autonomous evolution device based on parallel systems provided in this embodiment of the invention further includes: a fourth sub-module, used to implement step 5, specifically including: step 5.1: analyze and summarize the objective function of AutoRL, combine it with intelligent quantitative evaluation, and establish an objective function for autonomous optimization evolution based on intelligent evaluation, as shown in equations (3) and (4):

[0070] T * =arg max 1≤k≤K I (k) (T (k) (p (k) ),n m (3)

[0071] p * =arg max 1≤k≤K J(T (k) (p (k) (4)

[0072] Among them, T * and p * They represent the optimal comprehensive intelligence level I, respectively. * The corresponding optimal agent and its model parameters and hyperparameters; p (k) The agent T generated in each iteration (k) The model parameters and hyperparameters, and their combined intelligence level is I. (k) ; (k) represents the number of iterations during agent evolution; J is the reward function for reinforcement learning.

[0073] Based on the above device embodiments, as an optional embodiment, the artificial intelligence quantitative evaluation and autonomous evolution device based on parallel systems provided in this embodiment of the invention further includes: a fifth sub-module, used to implement step 6, specifically including: step 6.1: Based on the management and control process of the parallel system, obtain the mathematical expression of the autonomous evolution process based on intelligent evaluation, as shown in equation (5):

[0074]

[0075] in, This represents the artificial system control quantity of the virtual and real systems in a parallel system. This represents the actual system control quantity of the virtual and real systems in a parallel system;

[0076] Step 6.2: Since the evaluation and evolution of agents only consider virtual systems in parallel systems, and U t A,(k) By T (k) T (k) Control by p (k) The basic framework for autonomous optimization evolution based on intelligent evaluation is obtained from equation (5), as shown in equation (6):

[0077]

[0078] in, The optimized agent model and hyperparameters are given.

[0079] Based on the above-described device embodiments, as an optional embodiment, the artificial intelligence quantitative evaluation and autonomous evolution device based on parallel systems provided in this embodiment of the invention further includes: a sixth sub-module, used to implement step 7, specifically including: step 7.1: selecting a Bayesian optimization algorithm for an agent and establishing or selecting a suitable AutoRL system; step 7.2: based on the basic framework obtained in step 6, embedding the intelligent quantitative evaluation method into the AutoRL system selected in step 7.1 to construct an autonomous optimization evolution system based on intelligent evaluation, wherein the autonomous optimization evolution system based on intelligent evaluation is an intelligent evaluation and autonomous evolution system based on parallel systems.

[0080] The method in this embodiment of the invention is implemented using an electronic device; therefore, it is necessary to introduce the relevant electronic device. For this purpose, this embodiment of the invention provides an electronic device, such as... Figure 3 As shown, the electronic device includes at least one processor, a communications interface, at least one memory, and a communications bus, wherein the at least one processor, the communications interface, and the at least one memory communicate with each other via the communications bus. The at least one processor can invoke logical instructions stored in the at least one memory to execute all or part of the steps of the methods provided in the foregoing method embodiments.

[0081] Furthermore, when the logical instructions in at least one of the aforementioned memories can be implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various method embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0082] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0083] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0084] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. Based on this understanding, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, or sometimes in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0085] It should be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0086] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for quantitative evaluation and autonomous evolution of artificial intelligence based on parallel systems, characterized in that, include: Step 1: Construct an intelligent quantitative evaluation index system and select intelligent index m from it; Step 2: Based on the selected intelligent indicator m, set up and generate test scenario n. m Step 3: Based on the indicator scoring formula, calculate the test scenario n of the tested intelligent agent in terms of intelligent indicator m. m The intelligence level in the middle is quantitatively scored. m,n Step 4: Construct an intelligent quantitative evaluation function and calculate the comprehensive intelligence level of the tested intelligent agent; Step 5: Construct an objective function for autonomous optimization evolution based on intelligent evaluation; Step 6: Based on the parallel system as the basic platform, and according to the objective function, introduce intelligent quantitative evaluation to construct a basic framework for autonomous optimization evolution of the intelligent agent based on intelligent evaluation. Step 7: Based on the basic framework of autonomous evolution, construct an autonomous optimization-oriented evolutionary system based on intelligent evaluation; Step 5 specifically includes: Step 5.1: Analyze and summarize the objective function of AutoRL, and combine it with intelligent quantitative evaluation to establish an objective function for autonomous optimization evolution based on intelligent evaluation, as shown in equations (3) and (4): (3) (4) in, and They represent the optimal comprehensive intelligence level, respectively. The corresponding optimal agent and its model parameters and hyperparameters; The intelligent agent generated in each iteration The model parameters and hyperparameters, and their overall intelligence level are: (k) represents the number of iterations during agent evolution; J is the reward function for reinforcement learning; Step 6 specifically includes: Step 6.1: Based on the management and control process of the parallel system, obtain the mathematical expression of the autonomous evolution process based on intelligent evaluation, as shown in equation (5): (5) in, This represents the artificial system control quantity of the virtual and real systems in a parallel system. This represents the actual system control quantity of the virtual and real systems in a parallel system; Step 6.2: Since the evaluation and evolution of agents only consider virtual systems within parallel systems, and Depend on , Control by The basic framework for autonomous optimization evolution based on intelligent evaluation is obtained from equation (5), as shown in equation (6): (6) in, The optimized agent model and hyperparameters are given.

2. The method for quantitative evaluation and autonomous evolution of artificial intelligence based on parallel systems according to claim 1, characterized in that, Step 1 specifically includes: Step 1.1: Dividing machine intelligence into 35 indicators; Step 1.2: According to the principles of representativeness and comprehensiveness, classifying and summarizing the 35 indicators based on their similarity and superiority to humans, task completion, and the intelligence level that the intelligent agent should possess; Step 1.3: Constructing a quantitative evaluation index system for intelligence based on the classification and summary of the 35 intelligence indicators; Step 1.4: Based on the tested intelligent agent... The specific tasks to be performed and the capabilities required are determined by selecting appropriate intelligent indicators m from the intelligent quantitative evaluation indicator system constructed in step 1.3 for subsequent intelligent evaluation.

3. The method for quantitative evaluation and autonomous evolution of artificial intelligence based on parallel systems according to claim 2, characterized in that, Step 2 specifically includes: Step 2.1: Based on the selected intelligence index m, clarify the characteristics of each test scenario and the requirements for each test scenario; Step 2.2: Compare the training scenarios of the tested agent with the characteristics and requirements of the test scenarios, and generate test scenarios for intelligence index m in the parallel system. .

4. The method for quantitative evaluation and autonomous evolution of artificial intelligence based on parallel systems according to claim 3, characterized in that, Steps 3 and 4 specifically include: Step 3.1: In the test scenario generated by the parallel system, the tested agent is run to obtain observations of its running state in the test scenario. Step 3.2: Based on the tasks performed and problems solved by the tested agent, select or construct appropriate quantitative scoring formulas, and apply these scoring formulas with the help of... Scoring of computational agents in test scenarios Step 3.3: Using the entropy weight method, through... Calculate or select the weights of each scoring function in the smart metric. Step 3.4: Using the weighted summation method shown in equation (1), and Calculate the test scores for each indicator. , (1) Where n is the indicator scoring formula, and N is the total number of indicator scoring formulas. It is the normalized version ; Step 4.1: Calculate the weights of each intelligent indicator using the entropy weight method. , or take ,in This is the total number of selected intelligent indicators; Step 4.2: Using (2), calculate the overall intelligence level of the agent. , (2)。 5. The method for quantitative evaluation and autonomous evolution of artificial intelligence based on parallel systems according to claim 4, characterized in that, Step 7 specifically includes: Step 7.1: Selecting a Bayesian optimization algorithm for the agent and establishing or selecting a suitable AutoRL system; Step 7.2: Based on the basic framework obtained in Step 6, embedding the intelligent quantitative evaluation method into the AutoRL system selected in Step 7.1 to construct an autonomous optimization evolution system based on intelligent evaluation, wherein the autonomous optimization evolution system based on intelligent evaluation is an intelligent evaluation and autonomous evolution system based on parallel systems.

6. A device for quantitative evaluation and autonomous evolution of artificial intelligence based on parallel systems using the method of any one of claims 1 to 5, characterized in that, include: The first main module is used to implement step 1: constructing an intelligent quantitative evaluation index system and selecting intelligent index m from it; Step 2: Based on the selected intelligent indicator m, set up and generate test scenario n. m The second main module is used to implement step 3: based on the indicator scoring formula, calculate the test scenario n of the tested intelligent agent in terms of intelligent indicator m. m The intelligence level in the middle is quantitatively scored. m,n Step 4: Construct an intelligent quantitative evaluation function and calculate the comprehensive intelligence level of the tested intelligent agent; The third main module is used to implement Step 5: Construct an objective function for autonomous optimization evolution based on intelligent evaluation; Step 6: Based on the parallel system as the basic platform, and according to the objective function, introduce intelligent quantitative evaluation to construct a basic framework for autonomous optimization evolution of the intelligent agent based on intelligent evaluation; The fourth main module is used to implement Step 7: Construct an autonomous optimization evolution system based on intelligent evaluation according to the basic framework of autonomous evolution.

7. An electronic device, characterized in that, include: At least one processor, at least one memory, and a communication interface; wherein, The processor, memory, and communication interface communicate with each other; The memory stores program instructions that can be executed by the processor, which invokes the program instructions to perform the method described in any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium stores computer instructions that cause the computer to perform the method described in any one of claims 1 to 5.

Images