Decision making method and device

Abstract

The present disclosure provides a decision generation method and device, which relates to the field of computer technology, and specifically to the field of big data and artificial intelligence technology. The specific implementation scheme is: first, multiple pre-trained decision models corresponding to multiple decision strategies are obtained, and the pre-trained decision models correspond to the decision strategies one by one, and then the problem to be processed and the decision scenario corresponding to the problem to be processed are obtained, and based on the decision scenario corresponding to the problem to be processed, a target decision model is selected from multiple pre-trained decision models, and finally, based on the problem to be processed and the target decision model, a target decision corresponding to the problem to be processed is generated, and multiple pre-trained decision models corresponding to the decision strategies are obtained in advance, and in the process of using the pre-trained decision models, the same pre-trained decision model can be used to make decisions on different problems to be processed, thereby realizing the decoupling between the decision model and the problem to be processed.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, specifically to the field of big data and artificial intelligence technology, and especially to a decision-making method and device. Background Art

[0002] With the continuous advancement of science and technology, more and more problems are emerging, and decision analysis is needed for different problems, thus proposing a decision model for operations optimization. The decision model is a model established for management decision-making, that is, a mathematical model developed to assist decision-making. With the development of operations research, effective decision analysis methods such as linear rules, dynamic rules, game theory, queuing theory, deposit and loan models, and scheduling models have emerged.

[0003] With the rise of machine learning in recent years, some solutions have used machine learning to accelerate solutions, applying precise methods, heuristic methods, or direct end-to-end solutions. Corresponding decision models are trained for different problems, so that the decision models can make decisions on the problem. Summary of the invention

[0004] The present disclosure provides a decision making method, device, electronic device, storage medium and computer program product.

[0005] According to one aspect of the present disclosure, a decision generation method is provided, the method comprising: obtaining a plurality of pre-trained decision models corresponding to a plurality of decision strategies, wherein the pre-trained decision models correspond one-to-one to the decision strategies; obtaining a problem to be processed and a decision scenario corresponding to the problem to be processed; based on the decision scenario corresponding to the problem to be processed, selecting a target decision model from the plurality of pre-trained decision models; based on the problem to be processed and the target decision model, generating a target decision corresponding to the problem to be processed.

[0006] According to another aspect of the present disclosure, a decision generation device is provided, which includes: an acquisition module, configured to acquire multiple pre-trained decision models corresponding to multiple decision strategies, wherein the pre-trained decision models correspond to the decision strategies one-to-one; acquire a problem to be processed and a decision scenario corresponding to the problem to be processed; a selection module, configured to select a target decision model from the multiple pre-trained decision models based on the decision scenario corresponding to the problem to be processed; and a generation module, configured to generate a target decision corresponding to the problem to be processed based on the problem to be processed and the target decision model.

[0007] According to another aspect of the present disclosure, an electronic device is provided, which includes at least one processor; and a memory communicatively connected to 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 so that the at least one processor can execute the above-mentioned decision-making method.

[0008] According to another aspect of the present disclosure, a computer-readable medium is provided, on which computer instructions are stored, and the computer instructions are used to enable a computer to execute the above-mentioned decision-making method.

[0009] According to another aspect of the present disclosure, an embodiment of the present application provides a computer program product, which includes a computer program / instructions, and the computer program / instructions implement the above-mentioned decision making method when executed by a processor.

[0010] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0012] Figure 1 is a flow chart of an embodiment of a decision making method according to the present disclosure;

[0013] Figure 2 is a flowchart of an embodiment of obtaining multiple pre-trained decision models corresponding to multiple decision strategies according to the present disclosure;

[0014] Figure 3 is a flow chart of another embodiment of a decision making method according to the present disclosure;

[0015] Figure 4 is a flow chart of yet another embodiment of a decision making method according to the present disclosure;

[0016] Figure 5 is a flow chart of an embodiment of a monitoring target decision model according to the present disclosure;

[0017] Figure 6 is a schematic structural diagram of an embodiment of a decision making device according to the present disclosure;

[0018] Figure 7 It is a block diagram of an electronic device used to implement the decision making method of the embodiment of the present disclosure. DETAILED DESCRIPTION

[0019] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0020] refer to Figure 1 , Figure 1 A flowchart 100 of an embodiment of a decision making method that can be applied to the present disclosure is shown. The decision making method comprises the following steps:

[0021] Step 110, obtaining multiple pre-trained decision models corresponding to multiple decision strategies.

[0022] In this embodiment, the executor of the decision generation method (such as a terminal device or a server) can obtain multiple pre-trained decision models through network reading and other methods. The pre-trained decision models correspond to decision strategies one by one. Each decision strategy can correspond to a pre-trained decision model. The decision strategy can characterize the application strategies of different decision models, and can include precise solution methods, heuristic solution methods, and methods for scenario problems and solutions. Among them, the precise solution method can be to imitate and learn the pre-trained decision model through the algorithm in the solver; the heuristic solution method can be to predict the processing problem based on the pre-trained decision model, obtain the evaluation index of the problem result, and judge which group of decisions is the best to select through the evaluation index; the method for scenario problems and solutions can be the artificial experience of a certain scenario or the result of solving it with a heuristic algorithm / precise algorithm, and restore the result to a sequence decision. Furthermore, each pre-trained decision model is obtained by training an initial decision model based on sample data of multiple processing problems for the current decision strategy, so that each pre-trained decision model can make decisions on multiple different processing problems. The multiple pre-trained decision models may include solver-type pre-trained decision models, heuristic algorithm-type pre-trained decision models, and end-to-end model-type pre-trained decision models.

[0023] Step 120, obtaining the problem to be processed and the decision scenario corresponding to the problem to be processed.

[0024] In this embodiment, the above-mentioned execution entity can receive user input or obtain pending problems that need to be decided through network reading, etc. The pending problems can be problems that need to be decided based on decision models, such as how to maximize the number of parking spaces in a parking management system.

[0025] After the execution subject obtains the problem to be processed, it can also obtain the decision scenario corresponding to the problem to be processed. The decision scenario can be an application scenario of the pre-trained decision model, and can be an offline decision scenario or an online decision scenario.

[0026] Step 130 , based on the decision scenario corresponding to the problem to be processed, select a target decision model from a plurality of pre-trained decision models.

[0027] In this embodiment, after the execution subject obtains the decision scenario corresponding to the problem to be processed, the corresponding target decision model type can be determined in the corresponding relationship table between decision scenarios and decision model types according to the decision scenario, and different decision scenarios can correspond to different decision model types. Then the execution subject determines the corresponding target decision model type, and selects a pre-trained decision model corresponding to the target decision model type from multiple pre-trained decision models as the target decision model.

[0028] As an example, if the decision scenario is an online decision scenario, the above-mentioned execution entity can determine that the target decision model type corresponding to the online decision scenario is an end-to-end model type, and then a pre-trained decision model corresponding to the end-to-end model type can be selected from multiple pre-trained decision models and used as the target decision model for the problem to be processed in the online decision scenario.

[0029] Step 140: Generate a target decision corresponding to the problem to be processed based on the problem to be processed and the target decision model.

[0030] In this embodiment, after the above-mentioned execution subject obtains the problem to be processed and the target decision model, it can input the problem to be processed into the target decision model, make a decision on the problem to be processed through the target decision model, and generate a target decision corresponding to the problem to be processed.

[0031] Alternatively, the above-mentioned execution entity can collect sample data corresponding to the problem to be processed, train and adjust the target decision model based on the sample data to obtain an adjusted decision model, and input the problem to be processed into the adjusted decision model, and make a decision on the problem to be processed through the adjusted decision model to generate a target decision corresponding to the problem to be processed.

[0032] The decision generation method provided by the embodiments of the present disclosure obtains multiple pre-trained decision models corresponding to multiple decision strategies, and the pre-trained decision models correspond to the decision strategies one by one. Then, the problem to be processed and the decision scenarios corresponding to the problem to be processed are obtained, and based on the decision scenarios corresponding to the problem to be processed, a target decision model is selected from the multiple pre-trained decision models. Finally, based on the problem to be processed and the target decision model, a target decision corresponding to the problem to be processed is generated. Multiple pre-trained decision models corresponding to the decision strategies are obtained in advance, and in the process of using the pre-trained decision models, the same pre-trained decision model can be used to make decisions on different problems to be processed, thereby realizing the decoupling between the decision model and the problem to be processed. The decision model no longer exists only for one problem to be processed, but can solve multiple different problems to be processed, thereby improving the flexibility and versatility of the application of the decision model.

[0033] See also Figure 2 , Figure 2 A flowchart of an embodiment of obtaining multiple pre-trained decision models corresponding to multiple decision strategies is shown, that is, the above step 110, obtaining multiple pre-trained decision models corresponding to multiple decision strategies, may include the following steps:

[0034] Step 210, obtaining training conditions corresponding to the decision model.

[0035] In this step, the above-mentioned execution entity may obtain the training conditions corresponding to the decision model before pre-training the decision model. The training conditions may characterize the model accuracy requirements of the decision model obtained by training, etc. The training conditions may be pre-set by the operator according to the model requirements of the decision model, and the present disclosure does not make specific requirements for this.

[0036] Step 220, based on the training conditions, determine the corresponding model training steps.

[0037] In this step, after the above-mentioned execution subject obtains the training conditions of the decision model, it can determine the corresponding model training steps in the corresponding relationship table between the training conditions and the model training steps according to the training conditions. The model training steps may include the training method of the decision model, for example, including different model training methods such as supervised learning and reinforcement learning.

[0038] As an optional implementation method, the above step 220, based on the training conditions, determines the corresponding model training steps, and may include the following steps: constructing multiple training sub-steps for training the decision model; based on the training conditions, selecting corresponding target training sub-steps from the multiple training sub-steps to form a model training step.

[0039] Specifically, after the above-mentioned execution subject obtains the training conditions of the decision model, it can construct multiple training sub-steps for training the decision model. The training sub-steps can be method steps for training the model, which can include decision strategies, sample collection, supervised learning and reinforcement learning, etc., among which the decision strategies can include precise solution methods, heuristic solution methods, and scenario problem and solution methods; sample collection can include strategy selection methods, data aggregation methods and parallel sampling methods; supervised learning can include rule model methods, regression model methods, classification model methods and generation model methods; reinforcement learning can include solving the optimal reward sum method (Value based), solving the optimal policy method (Policybased) and hybrid solution method (hybrid).

[0040] The above-mentioned execution subject can select a target training sub-step that can meet the training conditions from multiple training sub-steps according to the training conditions. The selected target training sub-step can include reinforcement learning alone, or three target training sub-steps of decision strategy, sample collection, and supervised learning, or four target training sub-steps of decision strategy, sample collection, supervised learning, and reinforcement learning, wherein each training sub-step can adopt any one of the training methods included, without specific limitation. Thus, the above-mentioned execution subject can combine the selected target training sub-steps into a model training step, and obtain the model training step corresponding to the training condition.

[0041] In this implementation, different model training steps are obtained by combining different training sub-steps, thereby improving the diversity of the training steps. The training steps are further used to train the decision model, so that pre-trained decision models of different accuracies can be obtained, thereby improving the diversity of the decision model.

[0042] Step 230, for multiple decision strategies, the initial decision model is trained using a model training step to obtain multiple pre-trained decision models corresponding to the multiple decision strategies.

[0043] In this step, after the above-mentioned execution subject obtains the model training steps corresponding to the training conditions, for each of the multiple decision strategies, the model training steps can be used to train the initial decision model to obtain a pre-trained decision model corresponding to the decision strategy, so that the above-mentioned execution subject can obtain the pre-trained decision model corresponding to each decision strategy, and obtain multiple pre-trained decision models corresponding to multiple decision strategies.

[0044] The present disclosure can use relevant technical means to implement the model training step to train the initial decision model to obtain a pre-trained decision model corresponding to the decision strategy, and the present disclosure does not make specific limitations on this.

[0045] In this embodiment, the decision model is trained through the model training steps corresponding to the training conditions, so that a pre-trained decision model of corresponding accuracy can be obtained, thereby improving the pertinence and accuracy of the decision model.

[0046] refer to Figure 3 , Figure 3 A flow chart showing another embodiment of a decision making method is shown, and the decision making method may include the following steps:

[0047] Step 310, obtaining multiple pre-trained decision models corresponding to multiple decision strategies.

[0048] Step 310 of this embodiment can be performed according to Figure 1 Step 110 in the illustrated embodiment is performed in a similar manner and will not be described in detail here.

[0049] Step 320, obtaining the problem to be processed and the decision scenario corresponding to the problem to be processed.

[0050] Step 320 of this embodiment can be performed according to Figure 1 Step 120 in the illustrated embodiment is performed in a similar manner and will not be described in detail here.

[0051] Step 330, obtaining the decision requirement corresponding to the problem to be processed.

[0052] The above decision-making scenarios include offline decisions, so the pre-trained decision model is applied in offline decision-making scenarios.

[0053] In this step, the above-mentioned execution entity can obtain the decision requirements corresponding to the problem to be processed by receiving user input or network reading, and the decision requirements can characterize the effect requirements of the decision results of the problem to be processed, and may include high decision timeliness, high accuracy of the processing results corresponding to the decision results, high decision timeliness and high accuracy of the processing results corresponding to the decision results, etc.

[0054] Step 340: Select a target decision model from a plurality of pre-trained decision models based on the decision requirements corresponding to the problem to be processed.

[0055] In this step, after the execution subject obtains the decision requirement corresponding to the problem to be processed, it can determine the corresponding target decision model type in the corresponding relationship table between decision requirements and decision model types according to the decision requirement, and different decision requirements can correspond to different decision model types. Then the execution subject determines the corresponding target decision model type, and selects a pre-trained decision model corresponding to the target decision model type from multiple pre-trained decision models as the target decision model.

[0056] The above-mentioned decision-making requirements with high decision timeliness can correspond to the pre-trained decision model of the end-to-end model type, the decision-making requirements with high accuracy of the processing results corresponding to the decision results can correspond to the pre-trained decision model of the solver type, and the decision-making requirements with high decision timeliness and high accuracy of the processing results corresponding to the decision results can correspond to the pre-trained decision model of the heuristic algorithm type.

[0057] Step 350: Generate a target decision corresponding to the problem to be processed based on the problem to be processed and the target decision model.

[0058] Step 350 of this embodiment can be performed according to Figure 1 Step 140 in the illustrated embodiment is performed in a similar manner and will not be described in detail here.

[0059] In this embodiment, the corresponding target decision model is determined by different decision requirements, which improves the pertinence of the decision model under different requirements, thereby improving the pertinence of the decision results of the problem to be processed and meeting different requirements.

[0060] refer to Figure 4 , Figure 4 A flowchart of another embodiment of a decision making method is shown, and the decision making method may include the following steps:

[0061] Step 410, obtaining multiple pre-trained decision models corresponding to multiple decision strategies.

[0062] Step 410 of this embodiment can be performed according to Figure 1 Step 110 in the illustrated embodiment is performed in a similar manner and will not be described in detail here.

[0063] Step 420, obtaining the problem to be processed and the decision scenario corresponding to the problem to be processed.

[0064] Step 420 of this embodiment can be performed according to Figure 1 Step 120 in the illustrated embodiment is performed in a similar manner and will not be described in detail here.

[0065] Step 430 , based on the decision scenario corresponding to the problem to be processed, select a target decision model from a plurality of pre-trained decision models.

[0066] Step 430 of this embodiment can be performed according to Figure 1 Step 130 in the illustrated embodiment is performed in a similar manner and will not be described in detail here.

[0067] Step 440, obtaining the current status corresponding to the problem to be processed.

[0068] The above decision-making scenarios include online decision-making, so the pre-trained decision model is applied in the online decision-making scenario. Since online decision-making requires high timeliness, the above execution subject can determine that the target decision model is a pre-trained decision model of the end-to-end model type according to the online decision.

[0069] In this step, the above-mentioned execution entity can apply the pre-trained decision model of the end-to-end model type to the online decision-making scenario of the problem to be processed. By analyzing the information of the problem to be processed, the current state corresponding to the problem to be processed can be obtained. The current state can represent the event state corresponding to the problem to be processed, and different decisions can correspond to different states.

[0070] Step 450: Based on the current state and the issues to be processed, a corresponding decision list is output through the target decision model.

[0071] In this step, the execution subject may use the target decision model to make decisions on the current state and pending issues, and output a decision list corresponding to the pending issues in the current state, which may include multiple different processing decisions.

[0072] Step 460, based on the decision conditions corresponding to the problem to be processed, determine the target decision corresponding to the problem to be processed from the decision list.

[0073] In this step, after the execution subject obtains the decision list corresponding to the problem to be processed, it can further obtain the decision conditions corresponding to the problem to be processed. The decision conditions can be the selection indicators for the problem to be processed pre-set by the operator, which are used to select multiple decisions. Then, the execution subject determines the target decision that meets the decision conditions from the decision list according to the decision conditions as the target decision corresponding to the problem to be processed.

[0074] Furthermore, the execution entity may execute the target decision and update the current state to obtain an updated state.

[0075] In this embodiment, the target decision is determined through the target decision model and decision conditions, which improves the pertinence and accuracy of the target decision.

[0076] refer to Figure 5 , Figure 5 A flowchart showing an embodiment of a monitoring target decision model may include the following steps:

[0077] Step 510, obtaining decision data corresponding to the target decision model within a preset time period.

[0078] In this step, the above-mentioned execution entity can use the target decision model to make decisions on the problem to be processed, and can obtain decision data corresponding to the target decision model within a preset time period. The decision data may include the problem to be processed and the target decision generated by the target decision model. The present disclosure does not specifically limit the preset time period.

[0079] Step 520: Analyze the decision data to obtain analysis results.

[0080] In this step, after the above-mentioned execution entity obtains the decision data within the preset time period, it can perform data analysis on the decision data according to the best decision corresponding to the problem to be processed, such as the tendency of the data, the gap value of the data, etc., to obtain the corresponding analysis results. The analysis results may include the decision analysis results of the target decision model for multiple problems to be processed, the decision analysis results for a certain problem to be processed, etc.

[0081] Step 530: Generate performance results of the target decision model based on the analysis results.

[0082] In this step, after the above-mentioned execution entity obtains the analysis results of the decision data, it evaluates the effectiveness of the decision data of the target decision model. Based on the analysis results of the decision data, the performance results of the target decision model can be generated, which may include the decision performance for multiple problems to be processed, the decision performance for a certain problem to be processed, etc.

[0083] In this embodiment, by performing data analysis on the target decision model within a preset time period, regular evaluation of the decision results is achieved, thereby enabling regular monitoring of the target decision model to ensure the performance of the decision model.

[0084] Continue to refer Figure 5 , and may also include the following steps:

[0085] Step 540, determining whether the performance result of the target decision model meets the preset conditions.

[0086] In this step, after the above-mentioned execution entity obtains the performance result of the target decision model, it can compare the performance result with the preset conditions to determine whether the performance result of the target decision model meets the preset conditions. The preset conditions can indicate that the performance of the decision model cannot be lower than the performance threshold, that is, the effect after the decision of the target decision model cannot be lower than the preset threshold. The preset conditions disclosed in the present invention can also include other forms of expression, which are not specifically limited.

[0087] Step 550 , in response to determining that the performance result of the target decision model does not meet the preset condition, obtain sample data of the target decision model within a preset time period.

[0088] In this step, the above-mentioned execution entity determines that the performance results of the target decision model do not meet the preset conditions and determines that data drift has occurred, and then obtains sample data of the target decision model within a preset time period through collection. The sample data may include the problem to be processed and the corresponding optimal decision.

[0089] Step 560, performing model iteration on the target decision model based on the sample data.

[0090] In this step, after the execution subject obtains the sample data, it can perform model iteration on the target decision model based on the sample data to update and adjust the target decision model. The present disclosure can implement model iteration on the target decision model using sample data through relevant technical means.

[0091] In this embodiment, by regularly evaluating the effectiveness of the decision results, if the effectiveness of the target decision model is low and data drift occurs, the model is reviewed or iterated to achieve regular monitoring of the target decision model and ensure the performance of the decision model.

[0092] refer to Figure 6 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a decision making device. Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0093] like Figure 6 As shown, the decision making device 600 of this embodiment includes: an acquisition module 610 , a selection module 620 and a generation module 630 .

[0094] The acquisition module 610 is configured to acquire multiple pre-trained decision models corresponding to multiple decision strategies, wherein the pre-trained decision models correspond to the decision strategies one by one; acquire the problem to be processed and the decision scenario corresponding to the problem to be processed;

[0095] A selection module 620 is configured to select a target decision model from a plurality of pre-trained decision models based on a decision scenario corresponding to the problem to be processed;

[0096] The generation module 630 is configured to generate a target decision corresponding to the problem to be processed based on the problem to be processed and the target decision model.

[0097] In some optional embodiments of this embodiment, the acquisition module 610 includes: an acquisition unit, configured to acquire training conditions corresponding to the decision model; a determination unit, configured to determine the corresponding model training steps based on the training conditions; and a training unit, configured to train the initial decision model using the model training steps for multiple decision strategies to obtain multiple pre-trained decision models corresponding to the multiple decision strategies.

[0098] In some optional embodiments of this embodiment, the determination unit is further configured to: construct multiple training sub-steps for training the decision model; based on the training conditions, select corresponding target training sub-steps from the multiple training sub-steps to form a model training step.

[0099] In some optional embodiments of this embodiment, the decision scenario corresponding to the problem to be processed includes offline decision-making; and the selection module 620 is further configured to: obtain the decision requirements corresponding to the problem to be processed; and select a target decision model from multiple pre-trained decision models based on the decision requirements corresponding to the problem to be processed.

[0100] In some optional embodiments of this embodiment, the decision scenario corresponding to the problem to be processed includes online decision-making; and the generation module 630 is further configured to: obtain the current state corresponding to the problem to be processed; based on the current state and the problem to be processed, output the corresponding decision list through the target decision model; based on the decision conditions corresponding to the problem to be processed, determine the target decision corresponding to the problem to be processed from the decision list.

[0101] In some optional embodiments of this embodiment, the device also includes an analysis module; an acquisition module, further configured to: acquire decision data corresponding to the target decision model within a preset time period; the analysis module, configured to: perform data analysis on the decision data to obtain analysis results; and a generation module, further configured to: generate performance results of the target decision model based on the analysis results.

[0102] In some optional embodiments of this embodiment, the device also includes a judgment module and a model iteration module; the judgment module is configured to: judge whether the performance results of the target decision model meet the preset conditions; the acquisition module is further configured to: in response to determining that the performance results of the target decision model do not meet the preset conditions, acquire sample data of the target decision model within a preset time period; the model iteration module is configured to: perform model iteration on the target decision model based on the sample data.

[0103] The decision generation device provided by the embodiment of the present disclosure obtains multiple pre-trained decision models corresponding to multiple decision strategies, and the pre-trained decision models correspond to the decision strategies one by one. Then, the problem to be processed and the decision scenarios corresponding to the problem to be processed are obtained, and based on the decision scenarios corresponding to the problem to be processed, a target decision model is selected from the multiple pre-trained decision models. Finally, based on the problem to be processed and the target decision model, a target decision corresponding to the problem to be processed is generated. Multiple pre-trained decision models corresponding to the decision strategies are obtained in advance, and in the process of using the pre-trained decision models, the same pre-trained decision model can be used to make decisions on different problems to be processed, thereby realizing the decoupling between the decision model and the problem to be processed. The decision model no longer exists only for one problem to be processed, but can solve multiple different problems to be processed, thereby improving the flexibility and versatility of the application of the decision model.

[0104] Those skilled in the art will appreciate that the above device also includes some other well-known structures, such as a processor, a memory, etc. In order to unnecessarily obscure the embodiments of the present disclosure, these well-known structures are described in detail. Figure 6 Not shown in FIG.

[0105] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0106] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0107] Figure 7 A schematic block diagram of an example electronic device 700 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0108] like Figure 7As shown, the electronic device 700 includes a computing unit 701, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 702 or a computer program loaded from a storage unit 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the device 700 can also be stored. The computing unit 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 805 is also connected to the bus 704.

[0109] Multiple components in the electronic device 700 are connected to the I / O interface 705, including: an input unit 706, such as a keyboard, a mouse, etc.; an output unit 707, such as various types of displays, speakers, etc.; a storage unit 708, such as a disk, an optical disk, etc.; and a communication unit 709, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 709 allows the device 700 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0110] The computing unit 701 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 701 performs the various methods and processes described above, such as a decision-making method. For example, in some embodiments, the decision-making method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 808. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 700 via ROM 702 and / or a communication unit 709. When the computer program is loaded into RAM 703 and executed by the computing unit 701, one or more steps of the decision-making method described above may be performed. Alternatively, in other embodiments, the computing unit 701 may be configured to perform the decision-making method in any other appropriate manner (e.g., by means of firmware).

[0111] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0112] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0113] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0114] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0115] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0116] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0117] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0118] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A decision making method, comprising: Acquire a plurality of pre-trained decision models corresponding to a plurality of decision strategies, wherein the pre-trained decision models correspond one-to-one to the decision strategies; Obtaining a problem to be processed and a decision scenario corresponding to the problem to be processed, wherein the decision scenario corresponding to the problem to be processed includes an offline decision and an online decision; Based on the decision scenario corresponding to the problem to be processed, selecting a target decision model from the multiple pre-trained decision models, including: obtaining a decision requirement corresponding to the problem to be processed; based on the decision requirement corresponding to the problem to be processed, selecting a target decision model from the multiple pre-trained decision models; Based on the problem to be processed and the target decision model, generating a target decision corresponding to the problem to be processed; Wherein, generating a target decision corresponding to the problem to be processed based on the problem to be processed and the target decision model includes: Obtaining a current state corresponding to the pending issue, where the current state is used to represent an event state corresponding to the pending issue; Based on the current state and the pending issues, output a corresponding decision list through the target decision model; Based on the decision conditions corresponding to the problem to be processed, determining the target decision corresponding to the problem to be processed from the decision list; The step of obtaining a plurality of pre-trained decision models corresponding to a plurality of decision strategies includes: Determine the corresponding model training steps based on the training conditions corresponding to the decision model; For multiple decision strategies, the model training step is used to train the initial decision model to obtain multiple pre-trained decision models corresponding to the multiple decision strategies.

2. The method according to claim 1, wherein: The step of determining the corresponding model training step based on the training conditions corresponding to the decision model includes: Constructing multiple training sub-steps for training the decision model; Based on the training conditions, a corresponding target training sub-step is selected from the multiple training sub-steps to form the model training step.

3. The method according to any one of claims 1 to 2, further comprising: Obtaining decision data corresponding to the target decision model within a preset time period; Performing data analysis on the decision data to obtain analysis results; Based on the analysis results, a performance result of the target decision model is generated.

4. The method according to claim 3, further comprising: Determining whether the performance result of the target decision model meets the preset conditions; In response to determining that the performance result of the target decision model does not meet the preset condition, obtaining sample data of the target decision model within the preset time period; The target decision model is iterated based on the sample data.

5. A decision making device, comprising: An acquisition module is configured to acquire a plurality of pre-trained decision models corresponding to a plurality of decision strategies, wherein the pre-trained decision models correspond to the decision strategies one by one; acquire a problem to be processed and a decision scenario corresponding to the problem to be processed, wherein the decision scenario corresponding to the problem to be processed includes an offline decision and an online decision; The selection module is configured to select a target decision model from the plurality of pre-trained decision models based on a decision scenario corresponding to the problem to be processed, including: obtaining a decision requirement corresponding to the problem to be processed; selecting a target decision model from the plurality of pre-trained decision models based on the decision requirement corresponding to the problem to be processed; A generating module is configured to generate a target decision corresponding to the problem to be processed based on the problem to be processed and the target decision model; Wherein, the acquisition module includes: A determination unit, configured to determine a corresponding model training step based on a training condition corresponding to the decision model; A training unit is configured to train the initial decision model using the model training step for multiple decision strategies to obtain multiple pre-trained decision models corresponding to the multiple decision strategies; The generation module is further configured to: Obtaining a current state corresponding to the pending issue, where the current state is used to represent an event state corresponding to the pending issue; Based on the current state and the pending issues, output a corresponding decision list through the target decision model; Based on the decision conditions corresponding to the problem to be processed, a target decision corresponding to the problem to be processed is determined from the decision list.

6. The device according to claim 5, wherein: The determining unit is further configured to: Constructing multiple training sub-steps for training the decision model; Based on the training conditions, a corresponding target training sub-step is selected from the multiple training sub-steps to form the model training step.

7. The device according to any one of claims 5-6, further comprising an analysis module; The acquisition module is further configured to: acquire decision data corresponding to the target decision model within a preset time period; The analysis module is configured to: perform data analysis on the decision data to obtain analysis results; The generation module is further configured to generate a performance result of the target decision model based on the analysis result.

8. The device according to claim 7, further comprising a judgment module and a model iteration module; The judgment module is configured to: judge whether the performance result of the target decision model meets the preset conditions; The acquisition module is further configured to: in response to determining that the performance result of the target decision model does not meet the preset condition, acquire sample data of the target decision model within the preset time period; The model iteration module is configured to perform model iteration on the target decision model based on the sample data.

9. An electronic device, comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed 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 method according to any one of claims 1 to 4.

10. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-4.

11. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images