A blockchain-based river basin environment water quantity allocation system

Abstract

The application discloses a kind of based on watershed environment water quantity allocation system of block chain, it is related to water resource management technical field, including: data acquisition module, obtains real-time data;Risk assessment module, calculate decoupling risk index;Self-adapting control module, combine multiple factors to generate initial control vector;Security verification layer, determine optimal control vector;Intelligent contract execution module, according to risk level, execute corresponding control, form the automation closed loop from perception to execution.The application integrates multi-source heterogeneous data through blockchain technology, builds a trusted shared watershed database;Dynamically calculate decoupling risk index, accurately quantify the conflict between market transaction and ecological safety;With the help of deep reinforcement learning combined with weather forecast, realize forward-looking control;Rely on hierarchical intelligent contract to execute differentiated strategies, form a closed loop from perception, decision-making to execution, realize dynamic and accurate management of watershed water resources, balance market and ecological demand.

Description

Technical Field

[0001] The present invention relates to the technical field of water resource management, and in particular to a watershed environmental water allocation system based on blockchain. Background Art

[0002] At present, river basin data are scattered among different institutions and lack a unified and reliable sharing mechanism, resulting in a lack of data support for decisions such as water rights allocation, trading and ecological compensation; at the same time, water resource scheduling mostly relies on manual experience and is difficult to cope with complex and dynamically changing water conditions. There is an urgent need to achieve precise regulation through intelligent and digital technologies.

[0003] Existing water resource allocation technologies suffer from significant flaws. First, data silos are a serious problem, making it difficult to efficiently integrate multi-source data such as hydrological, meteorological, and agricultural water use, leading to delayed decision-making. Second, risk assessments lack dynamism, failing to deeply integrate market transactions with ecological security, making it difficult to quantify systemic risks. Third, regulatory measures are rigid, often adopting a one-size-fits-all approach, which can easily lead to conflicts of interest and fail to adapt to extreme disasters. Fourth, the decision-making logic and computational processes within AI models are opaque, making it difficult for outsiders to understand how they derive specific regulatory instructions based on input data. This leads to a lack of credibility and a lack of security verification mechanisms, potentially leading to uncontrollable consequences. Summary of the Invention

[0004] The purpose of the present invention is to provide a watershed environmental water allocation system based on blockchain, which solves the problems existing in the background technology.

[0005] To solve the above technical problems, the present invention provides a watershed environmental water allocation system based on blockchain, comprising: a data acquisition module for real-time collection of average market transaction prices and real-time monitoring of water levels;

[0006] The risk assessment module is used to calculate the decoupling risk index based on the average market transaction price, real-time monitored water levels, and preset ecological safety water level thresholds;

[0007] The adaptive control module is used to obtain the decoupling risk index calculated by the risk assessment module, the current water rights distribution vector and the meteorological forecast data to generate an environmental state vector;

[0008] The adaptive control module is also used to calculate the initial control vector based on the environment state vector through the deep reinforcement learning model;

[0009] The security verification layer is used to perform deterministic rule verification on the initial control vector;

[0010] When the initial control vector passes the deterministic rule verification, the initial control vector is determined as the optimal control vector;

[0011] When the initial control vector fails to pass the deterministic rule verification, the preset universal ecological water replenishment instruction is determined as the optimal control vector;

[0012] The smart contract execution module is used to determine the system risk level as safe, warning or crisis based on the decoupling risk index and the preset risk level threshold;

[0013] The smart contract execution module is also used to respond to early warning states and perform market-guided corrections based on the optimal control vector;

[0014] The smart contract execution module is also used to respond to crisis states and perform mandatory redistribution based on the optimal control vector;

[0015] The smart contract execution module is also used to respond to a safe state and not execute water rights regulation.

[0016] Preferably, the calculation of the decoupling risk index by the risk assessment module includes:

[0017] Calculate the price deviation based on the market average transaction price and the preset historical benchmark water price;

[0018] Calculate the ecological water level pressure based on the preset benchmark ecological water level, real-time monitoring water level and ecological safety water level threshold;

[0019] A decoupling risk index is generated by combining price deviation, ecological water level pressure and preset scenario correction coefficient.

[0020] Preferably, the ecological safety water level threshold is determined by taking the larger of the following two values: the minimum allowable water level required to maintain key ecological functions determined based on ecological research, and the water level value corresponding to the 95% guarantee rate based on the statistical distribution of historical minimum water level data.

[0021] Preferably, the historical benchmark water price is calculated by processing no less than 15 years of historical water rights transaction data;

[0022] The processing includes: adjusting the transaction prices of previous years for inflation based on the Consumer Price Index to obtain adjusted prices;

[0023] The adjusted price is weighted averaged using the transaction water volume as the weight.

[0024] Preferably, the reward function that the deep reinforcement learning model relies on for training is generated in the following way:

[0025] Calculate the ecological risk penalty term, which is the result of multiplying the decoupling risk index at the next moment by the first weight coefficient;

[0026] Calculate the control cost penalty term, which is the result of multiplying the sum of the squares of the control amplitudes of each component in the control vector by the second weight coefficient;

[0027] Calculate the agricultural economic impact penalty term, which is the result of multiplying the calculation result of the function for evaluating the potential impact of water rights distribution on agricultural output after regulation by the third weight coefficient;

[0028] The ecological risk penalty, the regulation cost penalty, and the agricultural economic impact penalty are summed to generate a reward function.

[0029] The first, second and third weight coefficients are pre-set by the basin management agency based on different scenarios such as flood season or dry season.

[0030] Preferably, the deterministic rule verification performed by the security verification layer includes:

[0031] Determine whether the downstream ecological water level is lower than the ecological safety water level threshold after the initial control vector is executed;

[0032] Determine whether the single control amount of any subject in the initial control vector exceeds the preset upper limit of the proportion of water rights held by the subject;

[0033] The initial control vector must pass all the above judgments before it is considered to have passed the verification.

[0034] Preferably, a market-guided correction, including:

[0035] After the smart contract is triggered, an offer with a premium is made to the agricultural entity specified in the optimal regulation vector;

[0036] Repurchase water rights from agricultural entities through tender offers and inject them into ecological reserve accounts.

[0037] Preferred, mandatory redistribution includes:

[0038] Smart contracts trigger emergency protocols;

[0039] The emergency agreement, based on the instructions of the optimal control vector, forcibly transfers the specified precise amount of water rights from all designated agricultural water users' accounts to the ecological reserve account;

[0040] Temporarily suspend free market trading of all water rights until the decoupling risk index falls back to the preset safety range. Beneficial effects

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1. Through blockchain, multi-source data can be trusted and uploaded to the chain, and an unalterable watershed database can be built, so that decentralized data can form a unified and reliable decision-making basis, ensuring the transparency and traceability of water rights allocation, transactions and other links, thereby improving credibility.

[0043] 2. By introducing a decoupled risk index to dynamically assess market-ecological conflicts, deeply coupling real-time market prices, ecological water levels, and preset safety thresholds, and combining deep reinforcement learning to achieve forward-looking regulation, the system can respond promptly at the incipient stage of risks, avoid the accumulation of crises, accurately respond to water situation changes, and balance the interests of multiple parties.

[0044] 3. By setting up a security verification layer and hierarchical smart contracts, we ensure that regulatory instructions are safe and controllable. At the same time, we implement differentiated strategies based on risk levels to avoid the drawbacks of one-size-fits-all management. In extreme cases, we can protect the ecological bottom line and reduce conflicts of interest, thus achieving closed-loop trusted control from decision-making to execution, and improving the reliability and adaptability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in 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 only some embodiments of the present invention, and those skilled in the art can derive other drawings based on these drawings without inventive effort.

[0046] Figure 1 It is a logic block diagram of the system of the present invention. DETAILED DESCRIPTION

[0047] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0048] Embodiment 1:

[0049] See also Figure 1 The present invention provides a watershed environmental water allocation system based on blockchain, comprising:

[0050] Data acquisition module, used to collect the average market transaction price and monitor the water level in real time;

[0051] The risk assessment module is used to calculate the decoupling risk index based on the average market transaction price, real-time monitored water levels, and preset ecological safety water level thresholds;

[0052] The adaptive control module is used to obtain the decoupling risk index calculated by the risk assessment module, the current water rights distribution vector and the meteorological forecast data to generate an environmental state vector;

[0053] The adaptive control module is also used to calculate the initial control vector based on the environment state vector through the deep reinforcement learning model;

[0054] The security verification layer is used to perform deterministic rule verification on the initial control vector;

[0055] When the initial control vector passes the deterministic rule verification, the initial control vector is determined as the optimal control vector;

[0056] When the initial control vector fails to pass the deterministic rule verification, the preset universal ecological water replenishment instruction is determined as the optimal control vector;

[0057] The smart contract execution module is used to determine the system risk level as safe, warning or crisis based on the decoupling risk index and the preset risk level threshold;

[0058] The smart contract execution module is also used to respond to early warning states and perform market-guided corrections based on the optimal control vector;

[0059] The smart contract execution module is also used to respond to crisis states and perform mandatory redistribution based on the optimal control vector;

[0060] The smart contract execution module is also used to respond to the safety state and not execute water rights regulation;

[0061] In this embodiment, a blockchain-based watershed environmental water allocation system achieves dynamic, precise, and adaptive management of water resources in the face of extreme disasters through the collaborative work of various modules. The data acquisition module continuously provides the system with real-time market and ecological data required for decision-making. The risk assessment module quantifies the risks faced by the system based on this data, and its output, a decoupled risk index, becomes the core basis for subsequent decision-making. The adaptive control module receives risk and water rights distribution status and, combined with forward-looking meteorological forecast data, constructs a complete environmental state vector, enabling the deep reinforcement learning model to make more predictive and proactive decisions. The security verification layer acts as a "firewall" for decision-making, ensuring that every instruction generated by the artificial intelligence is secure and controllable, addressing concerns about the opacity and uncontrollability of the AI ​​model's decision-making process. Finally, the smart contract execution module converts the verified optimal control vector into an action that can be accurately executed on the chain, adopting differentiated control measures based on different risk levels. This establishes a trustworthy, automated closed-loop control system from perception, decision-making, to execution, successfully preventing the systemic collapse of the "market-ecology" system during extreme drought events. The "market-ecology" system refers to the interconnected and mutually influential organic whole formed by the water rights trading market and the ecosystem within the basin.

[0062] In this embodiment, the deep reinforcement learning model can use the deep deterministic policy gradient (DDPG) algorithm, which is suitable for processing continuous control action space; wherein the environment state vector is the decoupling risk index , the current water rights of each entity, and the key meteorological forecast data for the next week (such as expected rainfall and average temperature) are normalized and concatenated into a one-dimensional vector;

[0063] In order to ensure that data of different dimensions can be effectively processed by the model, the normalization process can adopt the minimum-maximum normalization method; for any data feature , its normalized value The calculation formula is:

[0064] ;

[0065] is any original data feature to be normalized;

[0066] The original data features The value after min-max normalization;

[0067] and The maximum and minimum values ​​of the feature within the preset historical period;

[0068] When splicing, the normalized decoupling risk index (scalar), the current water weight vector of each entity (multidimensional vector), and the daily weather forecast data for the next week (multidimensional vector) are flattened and connected in series in a predetermined order to form a one-dimensional state vector of fixed length , as the input of the deep reinforcement learning model; for example, if there are N water rights entities and the meteorological data contains the rainfall and temperature for the next 7 days, then the state vector The dimension is ;

[0069] Initial control vector output by the model is a vector with the same dimension as the number of water rights subjects, and each component in the vector represents the proposed adjustment amount to the water rights stock of the corresponding subject (which can be positive or negative);

[0070] The universal ecological water replenishment directive is a deterministic rule. For example, the directive system will transfer water from all designated agricultural water users in proportion to their current water rights holdings. The total amount of water transferred is sufficient to ensure that the downstream ecological water level is restored to the ecological safety water level threshold. A preset buffer level above +0.1 m) for the target.

[0071] Example 2:

[0072] The risk assessment module calculates the decoupling risk index, including:

[0073] Calculate the price deviation based on the market average transaction price and the preset historical benchmark water price;

[0074] Calculate the ecological water level pressure based on the preset benchmark ecological water level, real-time monitoring water level and ecological safety water level threshold;

[0075] Combining price deviation, ecological water level pressure and preset scenario correction coefficient, a decoupling risk index is generated;

[0076] The historical benchmark water price is calculated by processing no less than 15 years of historical water rights transaction data;

[0077] The processing includes: adjusting the transaction prices of previous years for inflation based on the Consumer Price Index to obtain adjusted prices;

[0078] The adjusted price is weighted averaged using the transaction volume as weight;

[0079] In this embodiment, the risk assessment module accurately quantifies the degree of conflict between the water rights trading market and ecosystem protection through a comprehensive decoupled risk index formula. By incorporating the ecological security baseline as a key denominator in the calculation of this index, the risk index increases sharply when the real-time water level approaches this threshold, thereby more accurately quantifying the true risk of systemic collapse and providing a more reliable and sensitive decision-making basis for subsequent adaptive regulation.

[0080] Price deviation Defined as the current average market trading price Compared with historical benchmark water prices The ratio of , that is:

[0081] ;

[0082] is the price deviation, a dimensionless indicator used to measure the degree of deviation between the current water price and the historical benchmark water price;

[0083] is the average market transaction price of agricultural water rights in the current period;

[0084] The historical benchmark water price is calculated by adjusting the historical water rights transaction data for inflation and weighted averaging the transaction volume for at least 15 years, and serves as a reference point for the dimensionless price.

[0085] Ecological water level pressure It reflects the real-time water level Towards ecological safety water level threshold The degree of risk approaching, the denominator of which It reflects the near The nonlinear characteristic of the risk will increase sharply when , and its calculation formula is: ;

[0086] is the preset ecological safety water level threshold;

[0087] Real-time monitoring of water levels for downstream wetland ecosystems;

[0088] The baseline ecological water level is a reference point determined based on statistical analysis of historical wetland water level data;

[0089] Specifically, the statistical analysis refers to calculating the arithmetic mean or median of the daily average water level in the same historical period (for example, July of each year) over the past 20 years or more, and using this stable value as the normal reference water level of the ecosystem;

[0090] Finally, decoupling risk index Based on price deviation, ecological water level pressure and scenario correction coefficient Multiplying them together, we can fully quantify the coupling risks between the market and the ecosystem;

[0091] The decoupling risk index is calculated by the following formula:

[0092] ;

[0093] for Decoupling risk index at the moment;

[0094] This is the scenario correction coefficient. Under extreme combined disasters such as drought and heat waves, this coefficient is adjusted higher by preset rules to amplify the sensitivity of risk assessment;

[0095] The preset rule can be defined as follows: When the weather forecast data shows that the cumulative rainfall for the next 7 consecutive days is less than 5mm and the average daily maximum temperature is continuously higher than 38℃, the system will automatically The value of was adjusted from the baseline value of 1.0 to 1.5, thereby immediately increasing the response level of the risk index at the early stage of a disaster.

[0096] Example 3:

[0097] The ecological safety water level threshold is determined by taking the larger of the following two values: the minimum allowable water level required to maintain key ecological functions determined based on ecological research, and the water level corresponding to a 95% assurance rate based on the statistical distribution of historical minimum water level data;

[0098] In this embodiment, the setting of the ecological safety water level threshold adopts a double insurance mechanism, which ensures that the threshold has both solid ecological theoretical support and full consideration of historical extreme situations, making it an insurmountable hard constraint parameter encoded in the smart contract as system protection; based on ecological research, by analyzing the minimum water depth required for the survival of key species (such as fish, waterfowl) and aquatic plants in the basin, the minimum allowable water level required to maintain biodiversity is determined; at the same time, by statistically analyzing the historical lowest water level data for more than 20 years, the water level value corresponding to the 95% guarantee rate is taken as the historical extreme value reference; this method of combining theory and historical data makes the ecological safety water level threshold The setting is more scientific and robust, providing security for the entire risk assessment and control system.

[0099] Embodiment 4:

[0100] The reward function that deep reinforcement learning model training relies on is generated in the following way:

[0101] Calculate the ecological risk penalty term, which is the result of multiplying the decoupling risk index at the next moment by the first weight coefficient;

[0102] Calculate the control cost penalty term, which is the result of multiplying the sum of the squares of the control amplitudes of each component in the control vector by the second weight coefficient;

[0103] Calculate the agricultural economic impact penalty term, which is the result of multiplying the calculation result of the function for evaluating the potential impact of water rights distribution on agricultural output after regulation by the third weight coefficient;

[0104] The ecological risk penalty, the regulation cost penalty, and the agricultural economic impact penalty are summed to generate a reward function.

[0105] The first, second, and third weight coefficients are pre-set by the basin management agency based on different scenarios such as the wet season or the dry season;

[0106] In this example, the reward function of the deep reinforcement learning model has been carefully designed to guide the agent in learning how to balance the three core objectives of "ecological security," "market stability," and "economic impact," so that its decision-making is more in line with the complex real-world needs of watershed management. This reward function clearly defines the model's optimization direction, namely, maximizing long-term cumulative rewards.

[0107] The reward function is generated as follows:

[0108] ;

[0109] The immediate reward returned by the environment after performing the action;

[0110] is the ecological risk penalty item, and the risk index at the next moment The higher the value, the greater the negative reward, which strongly guides the agent to avoid behaviors that threaten the ecological bottom line;

[0111] To regulate cost penalties, it is used to punish excessive or too frequent interventions, encouraging stability at the lowest cost;

[0112] It is an agricultural economic impact penalty item, which is used to punish regulatory actions that may lead to a serious decline in total agricultural output.

[0113] is the evaluation function estimated based on the water resources-crop yield response function;

[0114] The weight coefficients corresponding to the three penalty items are preset by the basin management agency according to different scenarios. is the dimensionless weight coefficient, is a dimensionless coefficient whose dimension is (For example ), The dimension is ;

[0115] In this embodiment, A quadratic function can be used to approximate the penalty for the potential impact of changes in total irrigation water on agricultural output; assuming that the total initial agricultural water rights in the basin are The total agricultural water rights after regulation are , then the evaluation function It can be defined as:

[0116] ;

[0117] is an agricultural economic impact assessment function, which is used to quantify the potential penalty of water rights regulation on agricultural output;

[0118] for At that moment, the total water rights stock of agricultural entities before regulation;

[0119] for At time , the control vector generated by the model represents the water volume adjustment for each water right subject;

[0120] is the total agricultural water right after regulation, and its value is The sum of the water rights of all agricultural entities in the

[0121] is the total amount of agricultural water rights initially allocated;

[0122] is a preset penalty coefficient, dimensionless, used to adjust the weight of economic impact;

[0123] This function indicates that the further the total agricultural water volume after regulation deviates from the initial allocation, whether it increases or decreases, the potential economic losses will be incurred due to deviation from the agricultural production plan, and the greater the penalty will be.

[0124] Example 5:

[0125] The deterministic rule verification performed by the security verification layer includes:

[0126] Determine whether the downstream ecological water level is lower than the ecological safety water level threshold after the initial control vector is executed;

[0127] Determine whether the single control amount of any subject in the initial control vector exceeds the preset upper limit of the proportion of water rights held by the subject;

[0128] The initial control vector must pass all the above judgments before it is considered to have passed the verification;

[0129] In this embodiment, the security verification layer establishes a "firewall" between the output of the deep reinforcement learning model and the execution of the smart contract, adding a hard "firewall" to the opaque decision-making process of AI; the working principle of this verification layer is composed of preset, deterministic rules, ensuring that no matter what the internal logic of AI is, every control instruction it ultimately outputs is tested, meets the bottom line requirements, and will not cause a catastrophic impact on the system; specifically, the verification rules include judging whether the ecological water level after regulation will break through the ecological safety water level threshold , and determine whether the amount of control over any single entity exceeds a preset proportion (for example, 25%) of the total water rights it holds; only instructions that pass all checks will be released, and instructions that fail will be rejected and switched to a preset conservative "safe mode" operation; this design achieves an effective balance between complexity and reliability, providing a solid and reliable core decision-making capability for achieving true risk closed-loop control.

[0130] Example 6:

[0131] Market-guided corrections, including:

[0132] After the smart contract is triggered, an offer with a premium is made to the agricultural entity specified in the optimal regulation vector;

[0133] Repurchase water rights from agricultural entities through tender offers and inject them into ecological reserve accounts;

[0134] A blockchain-based watershed environmental water allocation system with mandatory redistribution, including:

[0135] Smart contracts trigger emergency protocols;

[0136] The emergency agreement, based on the instructions of the optimal control vector, forcibly transfers the specified precise amount of water rights from all designated agricultural water users' accounts to the ecological reserve account;

[0137] Temporarily suspending free market trading of all water rights until the decoupling risk index falls back to a pre-set safety range;

[0138] In this embodiment, the smart contract execution module strongly couples the intelligent decision-making of deep reinforcement learning with the reliable execution of blockchain, building a closed-loop control system in which decision instructions can be executed without deviation and automatically; in response to the early warning state ( ), the smart contract is triggered to execute market-guided corrections based on the optimal control vector Instructions are issued to launch a premium-priced targeted repurchase offer to designated agricultural entities, guide the flow of water rights to the ecological reserve account through market incentives, and fully implement the AI ​​regulation plan in a gentle manner; when the system enters a crisis state ( ), the smart contract triggers the emergency agreement to execute mandatory redistribution, no longer through market negotiation, but based on the optimal control vector The system compulsorily and precisely transfers a specified amount of water rights from designated entities' accounts to ecological reserve accounts, and suspends market transactions to prevent the spread of panic. This hierarchical and precise execution mechanism ensures that the optimal regulation plan generated by AI can be accurately implemented, truly achieving dynamic, precise, and adaptive management of water resources in the basin under extreme disasters.

[0139] The premium can be a function dynamically linked to the degree of system risk, used to dynamically calculate the premium for water rights repurchase when executing market-guided corrections in the early warning state. The premium rate calculation formula is:

[0140] ;

[0141] It is the core indicator for the current decoupling risk index, which quantifies the degree of conflict between the water rights market and the ecosystem;

[0142] For the preset excitation coefficient (e.g. =20%);

[0143] This design ensures that the higher the risk, the greater the incentive to repurchase water rights.

[0144] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A watershed environmental water allocation system based on blockchain, characterized in that: include: Data acquisition module, used to collect the average market transaction price and monitor the water level in real time; The risk assessment module is used to calculate the decoupling risk index based on the average market transaction price, real-time monitored water levels, and preset ecological safety water level thresholds; The adaptive control module is used to obtain the decoupling risk index calculated by the risk assessment module, the current water rights distribution vector and the meteorological forecast data to generate an environmental state vector; The adaptive control module is also used to calculate the initial control vector based on the environment state vector through the deep reinforcement learning model; The security verification layer is used to perform deterministic rule verification on the initial control vector; When the initial control vector passes the deterministic rule verification, the initial control vector is determined as the optimal control vector; When the initial control vector fails to pass the deterministic rule verification, the preset universal ecological water replenishment instruction is determined as the optimal control vector; The smart contract execution module is used to determine the system risk level as safe, warning or crisis based on the decoupling risk index and the preset risk level threshold; The smart contract execution module is also used to respond to early warning states and perform market-guided corrections based on the optimal control vector; The smart contract execution module is also used to respond to crisis states and perform mandatory redistribution based on the optimal control vector; The smart contract execution module is also used to respond to the safety state and not execute water rights regulation; The risk assessment module calculates the decoupling risk index, including: Calculate the price deviation based on the market average transaction price and the preset historical benchmark water price; Calculate the ecological water level pressure based on the preset benchmark ecological water level, real-time monitoring water level and ecological safety water level threshold; Combining price deviation, ecological water level pressure and preset scenario correction coefficient, a decoupling risk index is generated; The deterministic rule verification performed by the security verification layer includes: Determine whether the downstream ecological water level is lower than the ecological safety water level threshold after the initial control vector is executed; Determine whether the single control amount of any subject in the initial control vector exceeds the preset upper limit of the proportion of water rights held by the subject; The initial control vector must pass all the above judgments before it is considered to have passed the verification; Market-guided corrections, including: After the smart contract is triggered, an offer with a premium is made to the agricultural entity specified in the optimal regulation vector; Repurchase water rights from agricultural entities through tender offers and inject them into ecological reserve accounts; Mandatory redistribution, including: Smart contracts trigger emergency protocols; The emergency agreement, based on the instructions of the optimal control vector, forcibly transfers the specified precise amount of water rights from all designated agricultural water users' accounts to the ecological reserve account; Temporarily suspending free market trading of all water rights until the decoupling risk index falls back to a pre-set safety range; The decoupling risk index is calculated by the following formula: ； for Decoupling risk index at the moment; It is the scenario correction coefficient. Under extreme compound disasters, this coefficient is increased by the preset rules to amplify the sensitivity of risk assessment. is the average market transaction price of agricultural water rights in the current period; is the historical benchmark water price; is the preset ecological safety water level threshold; Real-time monitoring of water levels for downstream wetland ecosystems; The baseline ecological water level is a reference point determined based on statistical analysis of historical wetland water level data.

2. A watershed environmental water allocation system based on blockchain according to claim 1, characterized in that: The ecological safety water level threshold is determined by taking the larger of the following two values: the minimum allowable water level required to maintain key ecological functions determined based on ecological research, and the water level value corresponding to the 95% guarantee rate based on the statistical distribution of historical minimum water level data.

3. A watershed environmental water allocation system based on blockchain according to claim 2, characterized in that: The historical benchmark water price is calculated by processing no less than 15 years of historical water rights transaction data; The processing includes: adjusting the transaction prices of previous years for inflation based on the Consumer Price Index to obtain adjusted prices; The adjusted price is weighted averaged using the transaction water volume as the weight.

4. A watershed environmental water allocation system based on blockchain according to claim 1, characterized in that: The reward function that deep reinforcement learning model training relies on is generated in the following way: Calculate the ecological risk penalty term, which is the result of multiplying the decoupling risk index at the next moment by the first weight coefficient; Calculate the control cost penalty term, which is the result of multiplying the sum of the squares of the control amplitudes of each component in the control vector by the second weight coefficient; Calculate the agricultural economic impact penalty term, which is the result of multiplying the calculation result of the function for evaluating the potential impact of water rights distribution on agricultural output after regulation by the third weight coefficient; The ecological risk penalty, the regulation cost penalty, and the agricultural economic impact penalty are summed to generate a reward function. The first, second and third weight coefficients are pre-set by the basin management agency based on different scenarios such as flood season or dry season.

Images