Crop planting management method and device, electronic equipment and storage medium

Abstract

The invention provides a crop planting management method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring current environment state information; inputting the current environment state information into a pre-trained reinforcement learning model to obtain a global management decision output by the reinforcement learning model; wherein the reinforcement learning model is determined by completing various crop planting management decisions through cooperation of multiple agents according to a multi-objective optimization reward function. According to the crop planting management method provided by the invention, climate change information, soil change information and crop growth change information are taken as real-time environment change, and further through a complex reinforcement learning model, a multi-objective optimization reward function and a collaborative optimization mechanism, the crop planting management information is obtained. Various crop planting management decisions in agricultural production are continuously optimized according to real-time environment changes, the efficiency and accuracy of agricultural intelligent decision making are greatly improved, and the method is suitable for large-scale crop planting, especially intelligent management in a dynamic environment.

Description

Technical Field

[0001] The present invention relates to the field of agricultural information technology, and in particular to a crop planting management method, device, electronic equipment and storage medium. Background Art

[0002] Crops are an important part of daily life and play a vital role in the sustainable development of agriculture. Therefore, the planting of crops requires scientific and sustainable planting management methods to ensure the rational use of resources and the long-term stability of agricultural production. The planting of crops is affected by many factors, especially climate change, soil conditions, pests and diseases.

[0003] The existing crop planting management methods are based on fixed rules or historical data decisions and lack dynamic adjustment capabilities, which makes it difficult to adapt to changing environments and leads to poor decision-making quality. Especially when performing multi-task optimization at the same time, there are problems of uneven resource utilization and delayed decision-making. Summary of the invention

[0004] The present invention provides a crop planting management method, device, electronic equipment and storage medium, which are used to solve the problem of low accuracy of the crop planting management method in the prior art.

[0005] In a first aspect, the present invention provides a method for crop planting management, comprising: Acquiring current environmental status information; the current environmental status information includes at least one of climate change information, soil change information, and crop growth change information; Inputting the current environmental state information into a pre-trained reinforcement learning model to obtain a global management decision output by the reinforcement learning model; the global management decision includes a multi-dimensional crop planting management decision; Among them, the reinforcement learning model is determined by multiple agents collaborating to complete various crop planting management decisions based on a multi-objective optimization reward function.

[0006] In one embodiment, the reinforcement learning model is pre-trained in the following manner: A plurality of intelligent agents are defined; any of the intelligent agents is responsible for completing a specific crop planting management decision; Under the current environmental state sample, each of the intelligent agents determines the crop planting management decision according to the corresponding strategy function; By each of the intelligent agents executing corresponding crop planting management decisions, the current reward value of each of the intelligent agents and the next environmental state sample are observed; Performing a weighted summation on the current reward values ​​to obtain a global reward value for the multiple agents; Update the value function according to the current environment state sample and the global reward value; According to the updated value function, updating the corresponding strategy function of each agent, and determining the global strategy function of the multiple agents according to the updated strategy function; The next environment state sample is updated to the current environment state sample, and the steps of each agent determining the crop planting management decision according to the corresponding strategy function under the current environment state sample are executed until the preset training termination condition is reached to obtain the reinforcement learning model.

[0007] In one embodiment, the current reward value of each agent is determined by: After the agent executes the corresponding crop planting management decision, obtaining crop yield, crop quality, crop production cost and environmental impact information; The crop yield, the crop quality, the crop production cost and the environmental impact information are weighted and summed through a multi-objective optimization reward function to obtain a current reward value of the intelligent agent.

[0008] In one embodiment, the obtaining of current environment status information includes: Inputting the meteorological data monitored in the current environment into the climate change prediction model to obtain the climate change information for a period of time in the future output by the climate change prediction model; Inputting the soil data monitored under the current environment into a soil change prediction model to obtain soil change information for a period of time in the future output by the soil change prediction model; Inputting the crop growth data monitored under the current environment into the crop growth prediction model to obtain crop growth change information for a period of time in the future output by the crop growth change prediction model; At least one of the climate change information, the soil change information and the crop growth change information is combined into current environmental status information.

[0009] In one embodiment, the climate change prediction model makes predictions based on the following formula: in, indicating said climate change information; represents the temperature in the meteorological data; Indicates the humidity in the meteorological data; represents the precipitation in the meteorological data; Represents the interaction between the temperature, the humidity and the precipitation.

[0010] In one embodiment, the soil change prediction model performs prediction based on the following formula: in, Indicates the soil change information; represents the soil irrigation amount in the soil data; represents the soil moisture content in the soil data; represents the water evaporation coefficient in the soil data; Indicates the amount of change in soil moisture in the soil data caused by fertilization.

[0011] In one embodiment, the crop growth change prediction model performs prediction based on the following formula: in, Indicates the crop growth change information; Indicates the increase in photosynthesis yield in the crop growth data; Indicates the loss caused by environmental stress in the crop growth data; represents the impact coefficient of the said production increase; Indicates the factors affecting the loss.

[0012] In a second aspect, the present invention further provides a crop planting management device, comprising: A state acquisition module, used to acquire current environmental state information; the current environmental state information includes at least one of climate change information, soil change information and crop growth change information; A global management decision prediction module, used to input the current environmental state information into a pre-trained reinforcement learning model to obtain a global management decision output by the reinforcement learning model; the global management decision includes a multi-dimensional crop planting management decision; Among them, the reinforcement learning model is determined by multiple agents collaborating to complete various crop planting management decisions based on a multi-objective optimization reward function.

[0013] In a third aspect, the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of any of the above-mentioned crop planting management methods are implemented.

[0014] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of any of the above-mentioned crop planting management methods are implemented.

[0015] The crop planting management method, device, electronic device and storage medium provided by the present invention take into account the influence of factors such as climate conditions, soil conditions, and growth conditions during the crop planting process, and use climate change information, soil change information and crop growth change information as real-time environmental changes. Further, through complex reinforcement learning models, multi-objective optimization reward functions and collaborative optimization mechanisms, various crop planting management decisions in agricultural production are continuously optimized according to real-time environmental changes, which greatly improves the efficiency and accuracy of agricultural intelligent decision-making. It is suitable for large-scale crop planting, especially intelligent management in dynamic environments, and solves the problem that traditional agricultural decisions have poor adaptability and cannot cope with dynamic environmental changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 It is a schematic diagram of the process of the crop planting management method provided by the present invention.

[0018] Figure 2 It is a schematic diagram of the multi-agent collaborative optimization mechanism provided by the present invention.

[0019] Figure 3 This is a system architecture diagram of the intelligent optimization decision-making for crop planting provided by the present invention.

[0020] Figure 4 It is a structural schematic diagram of the crop planting management device provided by the present invention.

[0021] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] The terms "first", "second", etc. in the present invention are used to distinguish similar objects, rather than to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention can be implemented in an order other than those illustrated or described herein.

[0024] Combine the following Figure 1-Figure 5 The present invention describes the crop planting management method, device, electronic device and storage medium provided by the present invention.

[0025] It should be noted that the crop planting management method provided in the embodiment of the present invention is implemented based on the crop planting management device. The crop planting management method provided in the embodiment of the present invention continuously optimizes various decisions in agricultural production through dynamic learning of environmental data and multi-agent collaboration, thereby realizing dynamic adaptive resource management and decision optimization, and improving crop yield and resource utilization efficiency.

[0026] The embodiment of the present invention describes the crop planting management method by taking the crop planting management device as an execution subject as an example.

[0027] Combination Figure 1 , Figure 1 It is a schematic diagram of the process of the crop planting management method provided by the present invention.

[0028] like Figure 1 As shown, the method includes the following: Step 101: Obtain current environment status information; Step 102: input the current environment state information into a pre-trained reinforcement learning model to obtain a global management decision output by the reinforcement learning model.

[0029] Specifically, in the field environment where crops are planted, a variety of monitoring equipment is deployed to monitor the environmental data of the planting fields, including at least one of meteorological data, soil data, and crop growth data. Of course, it does not only include these environmental data, and other environmental data can be taken into account according to actual conditions. The embodiment of the present invention uses meteorological data, soil data and crop growth data as environmental data for description.

[0030] Through a variety of monitoring equipment, meteorological data, soil data and crop growth data in the current environment are collected and used as input data into the environmental change prediction model. The input data is analyzed and predicted through the environmental change prediction model, and climate change information, soil change information and crop growth change information in the future are output.

[0031] At least one of the predicted climate change information, soil change information and crop growth change information is determined as current environmental status information.

[0032] Furthermore, the current environmental state information is input into the pre-trained reinforcement learning model to obtain a global management decision output by the reinforcement learning model, wherein the global management decision includes multi-dimensional crop planting management decisions, which may specifically include crop planting management decisions in terms of irrigation, fertilization, pest and disease control, etc., but of course are not limited to crop planting management decisions in these dimensions.

[0033] Crop management decisions in irrigation, fertilization, pest control, etc. are closely related to climate change information, soil change information, and crop growth change information. Climate change information affects temperature, precipitation patterns, and extreme weather, directly affecting the frequency and timing of irrigation, fertilization, and pests and diseases. Soil change information affects the supply of soil moisture and nutrients and the pest and disease environment of the soil, determining the adjustment of irrigation and fertilization decisions and pest and disease control measures. Crop growth change information directly affects the demand for irrigation, fertilization, and pest and disease control decisions. The growth stage and health status of crops determine their needs for water, nutrients, and protective measures.

[0034] Among them, the reinforcement learning model is determined by the collaborative completion of various crop planting management decisions by multiple agents based on a multi-objective optimization reward function. That is to say, through real-time environmental data such as climate data, soil data and crop growth data, climate change information, soil change information and crop growth change information in the future are predicted. The reinforcement learning model can adopt different agents to complete corresponding specific tasks according to different agricultural scenarios while achieving multi-objective optimization (such as yield, quality, cost, and environmental impact), optimize crop planting management decisions in a single dimension, and finally complete various tasks through the collaboration of multiple agents, output global management decisions for crops planted in the field, realize real-time optimization of crop planting management decisions in irrigation, fertilization, pest and disease control, and complete dynamic global management.

[0035] The crop planting management method provided by the present invention takes into account the influence of factors such as climate conditions, soil conditions, and growth conditions during the crop planting process, and uses climate change information, soil change information, and crop growth change information as real-time environmental changes. It further uses complex reinforcement learning models, multi-objective optimization reward functions, and collaborative optimization mechanisms to continuously optimize various crop planting management decisions in agricultural production according to real-time environmental changes, greatly improving the efficiency and accuracy of agricultural intelligent decision-making. It is suitable for large-scale crop planting, especially intelligent management in dynamic environments, and solves the problem that traditional agricultural decisions have poor adaptability and cannot cope with dynamic environmental changes.

[0036] In some embodiments, based on step 101, obtaining current environment status information includes: Inputting the meteorological data monitored in the current environment into the climate change prediction model to obtain the climate change information for a period of time in the future output by the climate change prediction model; Inputting the soil data monitored under the current environment into a soil change prediction model to obtain soil change information for a period of time in the future output by the soil change prediction model; Inputting the crop growth data monitored under the current environment into the crop growth prediction model to obtain crop growth change information for a period of time in the future output by the crop growth change prediction model; At least one of the climate change information, the soil change information and the crop growth change information is combined into current environmental status information.

[0037] Specifically, meteorological data, such as temperature, humidity, precipitation, etc., can be monitored through weather stations or remote sensors. Soil data, including but not limited to soil moisture, soil irrigation, etc., can be monitored through soil sensors. Crop growth data, such as crop yield, crop loss, etc., can be monitored through remote sensing images taken by drones, ground sampling, or sensors. Therefore, according to the monitoring platform or monitoring equipment, meteorological data, soil data, and crop growth data in the current environment can be collected, and these data can be input into the environmental change prediction model for prediction, including climate change prediction models, soil change prediction models, and crop growth change prediction models.

[0038] The meteorological data monitored in the current environment is input into the climate change prediction model to obtain the climate change information for a period of time in the future output by the climate change prediction model.

[0039] When the climate change prediction model makes predictions, the crop growth change prediction model makes predictions based on the following formula: in, indicating said climate change information; represents the temperature in the meteorological data; Indicates the humidity in the meteorological data; represents the precipitation in the meteorological data; Represents the interaction between the temperature, the humidity and the precipitation.

[0040] The soil data monitored under the current environment is input into the soil change prediction model to obtain the soil change information for a period of time in the future output by the soil change prediction model.

[0041] When the soil change prediction model makes predictions, the soil change prediction model makes predictions based on the following formula: in, Indicates the soil change information; represents the soil irrigation amount in the soil data; represents the soil moisture content in the soil data; represents the water evaporation coefficient in the soil data; Indicates the amount of change in soil moisture in the soil data caused by fertilization.

[0042] The crop growth data monitored under the current environment is input into the crop growth prediction model to obtain the crop growth change information for a period of time in the future output by the crop growth change prediction model.

[0043] When the crop growth change prediction model makes a prediction, the crop growth change prediction model makes a prediction based on the following formula: in, Indicates the crop growth change information; Indicates the increase in photosynthesis yield in the crop growth data; Indicates the loss caused by environmental stress in the crop growth data; represents the impact coefficient of the said production increase; Indicates the factors affecting the loss.

[0044] The embodiment of the present invention inputs the monitored meteorological data, soil data and crop growth data into the climate change prediction model, the soil change prediction model and the crop growth change prediction model respectively for prediction, and outputs climate change information, soil change information and crop growth change information. These models provide the background knowledge required for system decision-making, and output environmental change information in future time periods, ensuring that the optimal global agricultural management decisions can be made based on environmental change information, crop needs and other external conditions.

[0045] In some embodiments, the reinforcement learning model is pre-trained by: A plurality of intelligent agents are defined; any of the intelligent agents is responsible for completing a specific crop planting management decision; Under the current environmental state sample, each of the intelligent agents determines the crop planting management decision according to the corresponding strategy function; By each of the intelligent agents executing corresponding crop planting management decisions, the current reward value of each of the intelligent agents and the next environmental state sample are observed; Performing a weighted summation on the current reward values ​​to obtain a global reward value for the multiple agents; Update the value function according to the current environment state sample and the global reward value; According to the updated value function, updating the corresponding strategy function of each agent, and determining the global strategy function of the multiple agents according to the updated strategy function; The next environment state sample is updated to the current environment state sample, and the steps of each agent determining the crop planting management decision according to the corresponding strategy function under the current environment state sample are executed until the preset training termination condition is reached to obtain the reinforcement learning model.

[0046] Specifically, multiple intelligent agents are defined, and each intelligent agent completes a specific crop planting management decision.

[0047] The current environment status samples are input through the input layer, including at least one of climate change information samples, soil change information samples and crop growth change information samples.

[0048] The input state is mapped to the decision space through the neural network, and the initialization strategy function is: in, Represents the current environment status sample, and denote the weight matrices of the hidden layer and the output layer respectively, represents the activation function, and They represent the bias terms of the hidden layer and the output layer respectively.

[0049] Through the value function To evaluate the benefits in the current state, the initialization value function is: in, Represents the current environment status sample, represents the discount factor, Represents the reward value.

[0050] Furthermore, under the current environmental state sample, each agent determines the crop planting management decision according to the corresponding strategy function.

[0051] Furthermore, each agent executes the corresponding crop planting management decision, and observes the current reward value of each agent and the next environmental state sample. Each agent makes autonomous decisions based on the reinforcement learning model. Status and actions Represents the current state and action selection of the agent, so the global management decision of the multi-agent is expressed as: ,in, For the The action of an agent at time t, is the joint action of all agents.

[0052] Furthermore, when multiple agents complete tasks through writing, each agent can obtain a local current reward value. By weighted summing up the current reward values, the global reward value of multiple agents can be obtained, which is expressed as: ,in, For the The weight coefficient of each agent is dynamically adjusted according to the importance of the task. For the The reward value of an agent at time t.

[0053] Furthermore, the value function is updated according to the current environment state sample and the global reward value, and further according to the updated value function, each agent is updated according to the corresponding policy function, and the global policy function of multiple agents is determined according to the updated policy function.

[0054] That is, through joint optimization, multiple agents share the current environment state samples. and the local policy function Collaborate to eventually form a global optimal strategy. Collaborative learning is described by the following formula: in, represents the discount factor, represents the expected value of future rewards, represents the joint action of all agents, is the global policy function (i.e., the global optimal policy).

[0055] Furthermore, the next environment state sample is updated to the current environment state sample, and under the current environment state sample, each agent determines the steps of crop planting management decision according to the corresponding strategy function until the preset training termination condition is reached, which can be a preset number of iterations or the value function and the strategy function reaching a preset performance, and finally a pre-trained reinforcement learning model is obtained.

[0056] The embodiment of the present invention adopts multi-agent collaboration to complete a multi-objective optimization mechanism. The intelligent agent is trained by sharing environmental states and collaborative strategies, so that the trained reinforcement learning model can generate optimized decisions for multi-dimensional crop planting management such as irrigation, fertilization, and pest and disease control, ensuring efficient use of resources for crop planting while improving crop yield and quality, etc., thereby solving the problem that traditional agricultural decisions have poor adaptability and cannot cope with dynamic environmental changes.

[0057] Based on the above, the current reward value of each agent is determined by: After the agent executes the corresponding crop planting management decision, obtaining crop yield, crop quality, crop production cost and environmental impact information; The crop yield, the crop quality, the crop production cost and the environmental impact information are weighted and summed through a multi-objective optimization reward function to obtain a current reward value of the intelligent agent.

[0058] Specifically, after the intelligent agent executes the corresponding crop planting management decision, it obtains crop yield, crop quality, crop production cost and environmental impact information.

[0059] Furthermore, the crop yield, crop quality, crop production cost and environmental impact information are weighted and summed through a multi-objective optimization reward function to obtain the current reward value of the agent.

[0060] The current reward value of the agent can be expressed by the following formula: Among them, P is crop yield, Q is crop quality, C is crop production cost, and E is environmental impact information. is the weight coefficient of each target.

[0061] In the iterative process, the weights of the reward function can be optimized using stochastic gradient descent: in, is the learning rate, is the partial derivative of the reward function with respect to the weights.

[0062] Based on the above reinforcement learning model training process, you can refer to Figure 2 , Figure 2 It is a schematic diagram of the multi-agent collaborative optimization mechanism provided by the present invention. Figure 2 The center displays the multi-agent collaboration mechanism, including irrigation agent, fertilization agent and pest control agent, which respectively execute irrigation decision-making tasks, fertilization decision-making tasks and pest control decision-making tasks. Figure 2The area around the center shows the optimization process of the multi-objective reward function, including the balance of crop yield, crop quality, crop production cost, and environmental impact information.

[0063] The embodiment of the present invention adopts a multi-objective optimization reward function during the reinforcement learning model training process, taking into account multiple optimization objectives such as yield, crop quality, crop production cost and environmental impact information, and balancing the trade-offs between these objectives during the training process to obtain a solution that can meet the requirements of each objective and optimize the overall performance as much as possible, so as to realize the global management decision of agricultural production.

[0064] In order to adapt to the dynamic changes in the agricultural production environment, climate data, soil data and crop growth data are monitored in real time through a variety of monitoring equipment, thereby predicting climate change information, soil change information and crop growth change information, and dynamically updating model parameters based on this real-time environmental change information to achieve continuous optimization.

[0065] The online learning process is as follows: 1. Data collection and status update: Obtain environmental change information such as climate change information, soil change information, and crop growth change information, and update it to the reinforcement learning model in real time.

[0066] 2. Strategy optimization and update: The model optimizes the strategy according to the current environment state and uses the gradient descent algorithm to update the model parameters. The strategy update equation is: in, represents the learning rate, represents the objective function of the strategy, Represents the gradient.

[0067] 3. Regular model updates and continuous optimization: Regularly update the global model based on the accumulated data, and adjust the priority and resource allocation ratio of each task (irrigation, fertilization, pest and disease control) to adapt to new environmental conditions.

[0068] Combining all the above embodiments, Figure 3 , Figure 3 This is a system architecture diagram of the intelligent optimization decision-making for crop planting provided by the present invention. Figure 3 The overall structure of crop planting environment modeling, reinforcement learning strategy generation, and online learning and model optimization is demonstrated. Figure 3 The upper part shows environmental modeling, which refers to the modeling of climate change prediction model, soil change prediction model and crop growth change prediction model. Figure 3 The middle part shows the input layer, hidden layer, and output layer of the reinforcement learning policy network, as well as the generated optimization decisions, including irrigation decisions, fertilization decisions, and pest and disease control decisions. Figure 3The bottom of the figure shows the online learning and model optimization process, demonstrating how to update model parameters through real-time monitoring data.

[0069] The crop planting management device provided by the present invention is described below. The crop planting management device described below and the crop planting management method described above can be referenced to each other.

[0070] Reference Figure 4 , Figure 4 It is a structural schematic diagram of the crop planting management device provided by the present invention.

[0071] The crop planting management device comprises: The state acquisition module 410 is used to acquire current environmental state information; the current environmental state information includes at least one of climate change information, soil change information and crop growth change information.

[0072] A global management decision prediction module 420 is used to input the current environmental state information into a pre-trained reinforcement learning model to obtain a global management decision output by the reinforcement learning model; the global management decision includes a multi-dimensional crop planting management decision; Among them, the reinforcement learning model is determined by multiple agents collaborating to complete various crop planting management decisions based on a multi-objective optimization reward function.

[0073] The crop planting management device provided by the present invention takes into account the influence of factors such as climate conditions, soil conditions, and growth conditions during the crop planting process, and uses climate change information, soil change information, and crop growth change information as real-time environmental changes. It further uses complex reinforcement learning models, multi-objective optimization reward functions, and collaborative optimization mechanisms to continuously optimize various crop planting management decisions in agricultural production according to real-time environmental changes, greatly improving the efficiency and accuracy of agricultural intelligent decision-making. It is suitable for large-scale crop planting, especially intelligent management in dynamic environments, and solves the problem that traditional agricultural decisions have poor adaptability and cannot cope with dynamic environmental changes.

[0074] Furthermore, the state acquisition module 410 is also used for: Inputting the meteorological data monitored in the current environment into the climate change prediction model to obtain the climate change information for a period of time in the future output by the climate change prediction model; Inputting the soil data monitored under the current environment into a soil change prediction model to obtain soil change information for a period of time in the future output by the soil change prediction model; Inputting the crop growth data monitored under the current environment into the crop growth prediction model to obtain crop growth change information for a period of time in the future output by the crop growth change prediction model; At least one of the climate change information, the soil change information and the crop growth change information is combined into current environmental status information.

[0075] Furthermore, the crop planting management device is also used for: A plurality of intelligent agents are defined; any of the intelligent agents is responsible for completing a specific crop planting management decision; Under the current environmental state sample, each of the intelligent agents determines the crop planting management decision according to the corresponding strategy function; By each of the intelligent agents executing corresponding crop planting management decisions, the current reward value of each of the intelligent agents and the next environmental state sample are observed; Performing a weighted summation on the current reward values ​​to obtain a global reward value for the multiple agents; Update the value function according to the current environment state sample and the global reward value; According to the updated value function, updating the corresponding strategy function of each agent, and determining the global strategy function of the multiple agents according to the updated strategy function; The next environment state sample is updated to the current environment state sample, and the steps of each agent determining the crop planting management decision according to the corresponding strategy function under the current environment state sample are executed until the preset training termination condition is reached to obtain the reinforcement learning model.

[0076] Furthermore, the crop planting management device is also used for: After the agent executes the corresponding crop planting management decision, obtaining crop yield, crop quality, crop production cost and environmental impact information; The crop yield, the crop quality, the crop production cost and the environmental impact information are weighted and summed through a multi-objective optimization reward function to obtain a current reward value of the intelligent agent.

[0077] It should be noted that the crop planting management device provided by the present invention can execute the crop planting management method described in any of the above embodiments during specific operation, which will not be elaborated in this embodiment.

[0078] Figure 5 is a schematic diagram of the structure of the electronic device provided by the present invention, such as Figure 5As shown, the electronic device may include: a processor 510, a communication interface 520, a memory 530 and a communication bus 540, wherein the processor 510, the communication interface 520 and the memory 530 communicate with each other through the communication bus 540. The processor 510 may call the logic instructions in the memory 530 to execute the crop planting management method, which includes: obtaining current environmental state information; the current environmental state information includes at least one of climate change information, soil change information and crop growth change information; inputting the current environmental state information into a pre-trained reinforcement learning model to obtain a global management decision output by the reinforcement learning model; the global management decision includes a multi-dimensional crop planting management decision; wherein the reinforcement learning model is determined by multiple agents collaborating to complete various crop planting management decisions based on a multi-objective optimization reward function.

[0079] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0080] On the other hand, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the crop planting management methods provided by the above-mentioned embodiments, the method including: obtaining current environmental status information; the current environmental status information includes at least one of climate change information, soil change information and crop growth change information; inputting the current environmental status information into a pre-trained reinforcement learning model to obtain a global management decision output by the reinforcement learning model; the global management decision includes a multi-dimensional crop planting management decision; wherein the reinforcement learning model is determined by multiple intelligent agents collaborating to complete various crop planting management decisions based on a multi-objective optimization reward function.

[0081] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the crop planting management method provided in the above-mentioned embodiments, the method comprising: obtaining current environmental state information; the current environmental state information comprises at least one of climate change information, soil change information and crop growth change information; inputting the current environmental state information into a pre-trained reinforcement learning model to obtain a global management decision output by the reinforcement learning model; the global management decision comprises a multi-dimensional crop planting management decision; wherein the reinforcement learning model is determined by multiple intelligent agents collaborating to complete various crop planting management decisions based on a multi-objective optimization reward function.

[0082] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. Those of ordinary skill in the art may understand and implement the present invention without creative effort.

[0083] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0084] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for crop planting management, characterized in that: include: Acquiring current environmental status information; the current environmental status information includes at least one of climate change information, soil change information, and crop growth change information; Inputting the current environment state information into a pre-trained reinforcement learning model to obtain a global management decision output by the reinforcement learning model; The global management decision includes a multi-dimensional crop planting management decision; Among them, the reinforcement learning model is determined by multiple agents collaborating to complete various crop planting management decisions based on a multi-objective optimization reward function.

2. The crop planting management method according to claim 1, characterized in that: The reinforcement learning model is pre-trained in the following way: A plurality of intelligent agents are defined; any of the intelligent agents is responsible for completing a specific crop planting management decision; Under the current environmental state sample, each of the intelligent agents determines the crop planting management decision according to the corresponding strategy function; By each of the intelligent agents executing corresponding crop planting management decisions, the current reward value of each of the intelligent agents and the next environmental state sample are observed; Performing a weighted summation on the current reward values ​​to obtain a global reward value for the multiple agents; Update the value function according to the current environment state sample and the global reward value; According to the updated value function, updating the corresponding strategy function of each agent, and determining the global strategy function of the multiple agents according to the updated strategy function; The next environment state sample is updated to the current environment state sample, and the steps of each agent determining the crop planting management decision according to the corresponding strategy function under the current environment state sample are executed until the preset training termination condition is reached to obtain the reinforcement learning model.

3. The crop planting management method according to claim 2, characterized in that: The current reward value of each agent is determined by: After the agent executes the corresponding crop planting management decision, obtaining crop yield, crop quality, crop production cost and environmental impact information; The crop yield, the crop quality, the crop production cost and the environmental impact information are weighted and summed through a multi-objective optimization reward function to obtain a current reward value of the intelligent agent.

4. The crop planting management method according to any one of claims 1 to 3, characterized in that: The obtaining of current environment status information includes: Inputting the meteorological data monitored in the current environment into the climate change prediction model to obtain the climate change information for a period of time in the future output by the climate change prediction model; Inputting the soil data monitored under the current environment into a soil change prediction model to obtain soil change information for a period of time in the future output by the soil change prediction model; Inputting the crop growth data monitored under the current environment into the crop growth prediction model to obtain crop growth change information for a period of time in the future output by the crop growth change prediction model; At least one of the climate change information, the soil change information and the crop growth change information is combined into current environmental status information.

5. The crop planting management method according to claim 4, characterized in that: The climate change prediction model is based on the following formula: in, indicating said climate change information; represents the temperature in the meteorological data; Indicates the humidity in the meteorological data; represents the precipitation in the meteorological data; Represents the interaction between the temperature, the humidity and the precipitation.

6. The crop planting management method according to claim 4, characterized in that: The soil change prediction model is based on the following formula: in, Indicates the soil change information; represents the soil irrigation amount in the soil data; represents the soil moisture content in the soil data; represents the water evaporation coefficient in the soil data; Indicates the amount of change in soil moisture in the soil data caused by fertilization.

7. The crop planting management method according to claim 4, characterized in that: The crop growth change prediction model is based on the following formula for prediction: in, Indicates the crop growth change information; Indicates the increase in photosynthesis yield in the crop growth data; Indicates the loss caused by environmental stress in the crop growth data; represents the impact coefficient of the said production increase; Indicates the factors affecting the loss.

8. A crop planting management device, characterized in that: include: A state acquisition module, used to acquire current environmental state information; the current environmental state information includes at least one of climate change information, soil change information and crop growth change information; A global management decision prediction module, used to input the current environmental state information into a pre-trained reinforcement learning model to obtain a global management decision output by the reinforcement learning model; the global management decision includes a multi-dimensional crop planting management decision; Among them, the reinforcement learning model is determined by multiple agents collaborating to complete various crop planting management decisions based on a multi-objective optimization reward function.

9. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the crop planting management method according to any one of claims 1 to 7 are implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the crop planting management method according to any one of claims 1 to 7 are implemented.

Images