Internet of Things data value affirmation system based on dynamic game theory

Abstract

The invention relates to the field of Internet of Things data, and discloses an Internet of Things data value determination system based on a dynamic game theory, and the system comprises an intelligent quality evaluation module which is used for carrying out the quality evaluation of Internet of Things data through a general artificial intelligence model, and screening a high-value data subset; the information funnel processing module is connected with the intelligent quality evaluation module and is used for carrying out classification and redundancy screening on the data and generating a utility ordinal number of each piece of data; and the dynamic value attenuation modeling module is connected with the information funnel processing module. According to the invention, by introducing the time-sensitive factor, dynamically adjusting the data value and dynamically pricing, the characteristic that the data value of the Internet of Things is rapidly attenuated along with time can be effectively reflected, the pricing accuracy is improved, redundant data is screened through the information funnel, the data value density is improved, and the data value is estimated through the utility ordinal number obtained through the information funnel.

Description

Technical Field

[0001] The present invention relates to the field of Internet of Things data, and in particular to an Internet of Things data value identification system based on dynamic game theory. Background Art

[0002] With the rapid development of IoT technology, the number of IoT devices is growing exponentially. According to data from the Ministry of Industry and Information Technology, China's IoT connections will exceed 6.6 billion in 2023, with a market size of 3.35 trillion yuan. IoT data has become a key production factor in the digital economy. Determining the value of IoT data faces multiple challenges. Data is highly time-sensitive (for example, the value of industrial warning data decreases by 80% within an hour), highly heterogeneous (structured and unstructured data coexist), and historical data is scarce due to the immaturity of the trading market.

[0003] Currently, the mainstream technologies for IoT data value identification mainly include the following two solutions:

[0004] Static cost pricing method (see "IEEE Internet of Things Journal" Volume 8, 2021): Based on the fixed cost allocation of data collection and storage, it ignores the time sensitivity and dynamic value changes of data.

[0005] Machine learning valuation models (such as patent CN114491944A): Use historical transaction data to train regression models to predict prices. However, in the early stages of the IoT data trading market, insufficient historical data limits the model's generalization capabilities.

[0006] In the existing IoT data value recognition system, the static pricing model cannot reflect the real-time decay characteristics of data value, resulting in the pricing of highly time-sensitive data seriously deviating from actual needs. It does not distinguish between the value differences of structured and unstructured data, making it difficult to achieve refined pricing.

[0007] These problems seriously restrict the healthy development of the IoT data trading market. A dynamic, adaptive and multimodal data value recognition theory is needed to promote the efficient circulation and value release of data elements. Summary of the Invention

[0008] In order to solve the problem in the above-mentioned IoT data value recognition system that the static pricing model cannot reflect the real-time decay characteristics of data value, resulting in the pricing of highly time-sensitive data seriously deviating from actual needs, failing to distinguish the value difference between structured and unstructured data, and difficulty in achieving refined pricing, the present invention is achieved through the following technical solution.

[0009] The IoT data value recognition system based on dynamic game theory includes:

[0010] Intelligent quality assessment module: used to assess the quality of IoT data through a general artificial intelligence model and screen high-value data subsets;

[0011] Information funnel processing module: connected to the intelligent quality assessment module, used to classify and filter redundancy of data, and generate the utility ordinal number of each data;

[0012] Dynamic value decay modeling module: connected to the information funnel processing module, builds a model for data value decay over time based on time-sensitive factors, and dynamically adjusts data value;

[0013] Game pricing execution module: connected to the dynamic value decay modeling module, estimates data prices through the Vickrey auction mechanism and utility function to achieve supply and demand matching.

[0014] Preferably, the intelligent quality assessment module includes:

[0015] Data acquisition unit, used to collect multimodal data from IoT devices in real time;

[0016] The multimodal feature fusion unit processes text, image and sensor data respectively through the Transformer model, convolutional neural network and time series analysis to generate a unified feature vector.

[0017] Preferably, the data acquisition unit includes structured data, unstructured data and sensor stream data.

[0018] Preferably, the information funnel processing module includes:

[0019] Semantic enhancement classification unit, clustering multimodal features based on hierarchical naming rules and semantic similarity algorithm (cosine similarity);

[0020] The dynamic redundancy screening unit calculates data redundancy through information entropy. The formula is:

[0021]

[0022] Among them, P i The entropy value is the proportion of a certain type of data in the total data set. When the entropy value is higher than the threshold, redundant elimination is triggered.

[0023] Preferably, the dynamic value decay modeling module is configured as follows:

[0024] The reinforcement learning algorithm is used to dynamically adjust the time-sensitive factor λ and the storage cost function C(t), and the reward function is designed as:

[0025] Preferably, the dynamic value decay modeling module adopts the following model for industrial real-time warning data scenarios:

[0026] P(t)=A(t)P0e -λt

[0027] Where A(t) is a monotonically increasing function satisfying A(0)=1.

[0028] Preferably, the state value decay modeling module adopts the following model for historical analysis data scenarios:

[0029] P(t)=P0e -λt +C(t)

[0030] Where P0 represents the initial value of the data, λ represents the attenuation coefficient (the degree of data depreciation), and C(t) is a monotonically increasing function.

[0031] Preferably, the game pricing execution module includes:

[0032] A blockchain-powered smart contract unit for executing the Vickrey auction process, including sealed bidding, on-chain bid opening, and transaction record storage;

[0033] The decentralized reputation system calculates the reputation score based on historical transaction data. The formula is:

[0034]

[0035] High-reputation users will be matched first and enjoy fee discounts.

[0036] Preferably, the intelligent quality assessment module further includes:

[0037] The data encryption unit is used to encrypt the transmission and storage of collected IoT data. The encryption algorithm uses a combination of AES-256 and homomorphic encryption to ensure the privacy and security of data during transmission and calculation.

[0038] Preferably, the blockchain-driven smart contract unit includes:

[0039] The dynamic parameter adjustment unit automatically adjusts auction parameters based on real-time market supply and demand data. The dynamic parameter adjustment unit includes:

[0040] Reserve price: dynamically set based on the median of historical transaction prices;

[0041] Bidding time window: shorten or extend according to the urgency of the transaction;

[0042] Transaction fee rate: Inversely correlated with the user's credit score, with lower fees for users with high credit rating.

[0043] This invention provides an IoT data value determination system based on dynamic game theory. Compared with existing technologies, it has the following advantages: by introducing a time-sensitive factor to dynamically adjust data value, dynamic pricing can effectively reflect the rapid decay of IoT data value over time, improving pricing accuracy. It also uses an information funnel to filter redundant data, increasing data value density. The utility ordinal number obtained through the information funnel is used to estimate data value. An AI model is used to assess IoT data quality, automatically locating core high-value data subsets. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 This is a principle block diagram of the present invention.

[0045] Figure 2 This is a flow chart of the system proposed by the present invention. DETAILED DESCRIPTION

[0046] The present invention is further described below with reference to specific examples. It should be understood that these examples are only used to illustrate the present invention and are not used to limit the scope of protection of the present invention.

[0047] The following describes the embodiments of the present invention through specific examples. Those skilled in the art will readily understand the other advantages and benefits of the present invention from the disclosure herein. The present invention may also be implemented or applied through various other specific embodiments, and the details in this specification may be modified or altered based on different viewpoints and applications without departing from the spirit of the present invention.

[0048] Example 1: Reference Figure 1-Figure 2 , the IoT data value recognition system based on dynamic game theory includes:

[0049] Intelligent quality assessment module: used to assess the quality of IoT data through a general artificial intelligence model and screen high-value data subsets;

[0050] Intelligent quality assessment modules include:

[0051] Data collection unit, used to collect multimodal data of IoT devices in real time. The data collection unit includes structured data, unstructured data and sensor stream data;

[0052] The multimodal feature fusion unit processes text, image, and sensor data separately through the Transformer model, convolutional neural network, and time series analysis to generate a unified feature vector;

[0053] The data encryption unit is used to encrypt the transmission and storage of collected IoT data. The encryption algorithm uses a combination of AES-256 and homomorphic encryption to ensure the privacy and security of data during transmission and calculation.

[0054] Information funnel processing module: connected to the intelligent quality assessment module, used to classify and filter redundancy of data, and generate the utility ordinal number of each data;

[0055] The information funnel processing module includes:

[0056] Semantic enhancement classification unit, clustering multimodal features based on hierarchical naming rules and semantic similarity algorithm (cosine similarity);

[0057] The dynamic redundancy screening unit calculates data redundancy through information entropy. The formula is:

[0058]

[0059] Among them, P i The entropy value is the proportion of a certain type of data in the total data set. When the entropy value is higher than the threshold, redundant elimination is triggered.

[0060] Dynamic value decay modeling module: connected to the information funnel processing module, builds a model for data value decay over time based on time-sensitive factors, and dynamically adjusts data value;

[0061] The dynamic value decay modeling module is configured as follows:

[0062] The reinforcement learning algorithm is used to dynamically adjust the time-sensitive factor λ and the storage cost function C(t), and the reward function is designed as:

[0063] in:

[0064] It is a comprehensive reward value used to measure the overall performance of the system in a certain period of time, by maximizing The reinforcement learning algorithm dynamically optimizes the time-sensitive factor λ and the storage cost function C(t);

[0065] α is the weight coefficient of the transaction success rate, with a value range of α≥0, which adjusts the contribution ratio of the transaction success rate to the total reward. For example, if the trading market focuses on fast transaction, a higher α can be set;

[0066] The transaction success rate is the percentage of successfully completed data transactions to the total number of transaction requests within a specific time period;

[0067] β is the weight coefficient of buyer satisfaction, ranging from β≥0, which controls the impact of buyer satisfaction on rewards. A high β value indicates that the system attaches more importance to buyer experience.

[0068] Buyer satisfaction is the buyer's satisfaction with the transaction results and can be quantified by:

[0069] Scoring mechanism: Buyers rate data quality, price, delivery speed, etc. (e.g., 1-5 points);

[0070] Repeat transaction rate: the frequency with which a buyer initiates another transaction within a certain period;

[0071] Buyer satisfaction measures the long-term trust of market participants and system sustainability;

[0072] γ is the weight coefficient of storage cost, ranging from γ ≥ 0, which weighs the negative impact of storage cost on rewards. A high γ value encourages the system to prioritize reducing storage overhead.

[0073] Storage costs refer to the resources required to store IoT data, including:

[0074] Hardware costs: maintenance costs of servers and storage devices;

[0075] Energy costs: electricity consumption in the data center;

[0076] Redundancy management costs: the cost of redundant data backup and cleanup;

[0077] By driving the reinforcement learning algorithm through the reward function, λ and the storage cost function C(t) are optimized in real time to achieve dynamic adjustment of data value and efficient utilization of resources.

[0078] In this embodiment, the dynamic value decay modeling module adopts the following model for industrial real-time warning data scenarios:

[0079] P(t)_=A(t)P0e -λt

[0080] Where A(t) is a monotonically increasing function satisfying A(0)=1.

[0081] Game pricing execution module: connects to the dynamic value decay modeling module, estimates data prices through the Vickrey auction mechanism and utility function, and achieves supply and demand matching;

[0082] The game pricing execution module includes:

[0083] A blockchain-powered smart contract unit for executing the Vickrey auction process, including sealed bidding, on-chain bid opening, and transaction record storage;

[0084] Incorporating Vickrey auction theory, different data demanders compete for data resources, determine a base price for a portion of the data, and estimate the utility function using a regression algorithm. Based on Vickrey auction theory, the winner only needs to pay the second-highest price to trade the product. This auction method can be substituted with other auction methods depending on the situation. The utility function is defined as U(k,n), where k is the utility ordinal of the current data and n is the total amount of data. The utility function can be fitted using monotonic functions such as exponential and logarithmic functions.

[0085] According to the marginal rate of substitution equal to the price ratio, the initial prices of the remaining data are estimated through the utility function, that is,

[0086] P A =P B ×U(k A ,n) / U(k B ,n)

[0087] Among them, P A and P B These are the prices of data A and B respectively. B is data with known price, and A is data with unknown price. After obtaining the data price, the real-time price of the data can be calculated through the price change model.

[0088] The blockchain-driven smart contract unit includes: a dynamic parameter adjustment unit that automatically adjusts auction parameters based on real-time market supply and demand data. The dynamic parameter adjustment unit includes:

[0089] Reserve price: dynamically set based on the median of historical transaction prices;

[0090] Bidding time window: shorten or extend according to the urgency of the transaction;

[0091] Transaction fee rate: Inversely correlated with the user's credit score, with lower fees for users with high credit rating.

[0092] The decentralized reputation system calculates the reputation score based on historical transaction data. The formula is:

[0093]

[0094] in:

[0095] Successful transaction volume i represents the sum of all successfully completed transactions by the user. The "transaction volume" here can be the number of transactions or the transaction amount, which needs to be determined according to the system design. If it is the amount, it can better reflect the user's actual contribution;

[0096] ∑Total transaction volume i represents the sum of all transactions initiated by the user, including successful and failed (e.g., canceled, timed out) transactions;

[0097] δ represents the penalty coefficient, which determines the negative impact of each dispute on the score;

[0098] The number of disputes indicates the number of complaints or arbitrations received during user transactions, reflecting disputes or breaches of contract in transactions.

[0099] Malicious behavior is suppressed through a penalty mechanism. The more disputes there are, the lower the credit score becomes, reducing the market competitiveness of problematic users. High-reputation users are given priority in matching and enjoy fee discounts.

[0100] Example 2: This example differs from Example 1 in that the dynamic value decay modeling module uses the following model for historical analysis data scenarios:

[0101] P(t)=P0e -λt +C(t)

[0102] Where P0 represents the initial value of the data, λ represents the attenuation coefficient (the degree of data depreciation), C(t) is a monotonically increasing function, and C(t) = 0. As time goes by, the storage cost of data will gradually increase, the redundancy of data will also decrease, and the value of the corresponding data will also increase.

[0103] Example 3: This example provides an application of an IoT data value recognition system based on dynamic game theory. A large farm deploys IoT devices to monitor soil moisture, meteorological conditions, and crop growth status. It needs to obtain accurate irrigation recommendation data in real time to optimize water resource utilization. Due to the significant differences in soil properties and crop types across different plots, the timeliness and accuracy of irrigation data directly affect yield and cost. However, it faces the following challenges:

[0104] Time sensitivity: The value of weather forecast data decreases by 50% within 6 hours (e.g., a change in rainfall probability causes a sudden drop in irrigation demand);

[0105] Heterogeneity: Soil sensor data (structured), crop images taken by drones (unstructured), and historical yield data (multimodal) need to be processed uniformly.

[0106] Lack of trust: Data providers exaggerate the accuracy of data, making it difficult for farmers to verify data quality.

[0107] Technical Implementation

[0108] 1. Intelligent quality assessment module:

[0109] Data collection:

[0110] Soil moisture sensor (structured data, updated every minute);

[0111] The probability of rainfall in the next 24 hours provided by the weather station (time series data);

[0112] Multispectral images of crops taken by drones (unstructured data, updated daily).

[0113] Quality Assessment:

[0114] Use lightweight CNN models to analyze crop images and identify water-stressed areas (e.g., leaf color and texture features);

[0115] The LSTM model is used to verify the temporal consistency of meteorological forecast data and filter out outliers.

[0116] 2. Information funnel processing module:

[0117] Classification and redundancy screening:

[0118] The first level of classification: by data type (sensor data, image data, meteorological data);

[0119] Second-level semantic clustering: Based on the correlation between soil moisture and crop images, the data is clustered into "high water demand plots" and "low water demand plots";

[0120] Redundancy elimination: Calculate the entropy value of meteorological data and eliminate prediction results with high repetitiveness (such as the probability of rainfall for three consecutive hours is 80%).

[0121] Utility ordinal generation:

[0122] The data priority of high water demand plots is set to 1, and that of low water demand plots is 2;

[0123] Meteorological data are sorted according to the forecast time window (data utility ordinal number within 6 hours is 1, and data utility ordinal number between 6-12 hours is 2).

[0124] 3. Dynamic Value Decay Modeling Module:

[0125] Model selection:

[0126] Weather forecast data uses industrial real-time early warning models:

[0127] P(t)=A(t)P0e -λt

[0128] A(t) is a step function, and the decay accelerates to A(t) = 0.5 after 6 hours;

[0129] Soil moisture data uses a historical analysis model:

[0130] P(t)=P0e -λt +C(t)

[0131] Parameter optimization:

[0132] The reinforcement learning reward function is set as:

[0133]

[0134] Game pricing execution module:

[0135] Blockchain smart contract process:

[0136] The data provider submits an irrigation recommendation data package (including utility ordinal number, quality score, and encrypted raw data);

[0137] Farmers bid anonymously, and the smart contract executes the Vickrey auction:

[0138] The highest-paying farmer pays the second-highest price and receives the data decryption key;

[0139] Transaction records are uploaded to the chain, including data hash value, transaction time, and price.

[0140] Reputation System:

[0141] Data provider reputation scores are based on historical data accuracy (e.g., the error rate between predicted rainfall and actual rainfall);

[0142] Data from highly reputable providers will be displayed first, and the handling fee will be reduced by 10%.

[0143] Implementation Effect

[0144] 1. Improved efficiency:

[0145] Irrigation decision response time reduced from 2 hours to 15 minutes;

[0146] The accuracy of high-value data recognition has increased to 92% (the original solution was 75%).

[0147] 2. Cost optimization:

[0148] The dynamic pricing model reduces the cost of purchasing weather data for farmers by 25% (avoiding the purchase of expired data);

[0149] Storage costs are reduced by 30% (by eliminating duplicate weather data through redundancy filtering).

[0150] 3. Trust Building:

[0151] Blockchain-enabled transparent transactions have reduced farmer complaints by 60%;

[0152] The proportion of transaction volume from high-reputation data providers increased from 40% to 75%.

[0153] Example 4: Based on Example 1 and Example 2, the specific process steps of the system are as follows:

[0154] 1. Data Collection and Preprocessing

[0155] Multi-source data collection:

[0156] Real-time collection of multimodal data through IoT devices (such as sensors, cameras, and smart terminals), including:

[0157] Structured data: sensor values (temperature, humidity, location, etc.);

[0158] Unstructured data: images, audio, log files;

[0159] Time series data: continuously monitored flow rate, energy consumption, etc.

[0160] Protocol adaptation: Compatible with communication protocols such as MQTT, CoAP, HTTP, etc., ensuring seamless access to heterogeneous devices.

[0161] Data encryption and secure transmission:

[0162] Use AES-256 encryption algorithm to encrypt the original data;

[0163] Homomorphic encryption technology is used to support preliminary calculations in an encrypted state (such as data feature extraction).

[0164] Output: Encrypted multimodal dataset.

[0165] 2. Intelligent quality assessment

[0166] Multimodal feature fusion:

[0167] Text data: Extract semantic features through the Transformer model;

[0168] Image data: Using convolutional neural networks (CNNs) to identify key visual features (e.g., abnormal regions in images of equipment failures);

[0169] Sensor data: Use the LSTM model to analyze time series features and detect abnormal fluctuations.

[0170] Quality Scoring and Filtering:

[0171] Assess data quality based on preset scoring algorithms (such as accuracy, completeness, and timeliness);

[0172] Filter out high-value data subsets (such as data with a score ≥ 80 points).

[0173] Output: A subset of high-value data and their quality scores.

[0174] 3. Information funnel processing

[0175] Hierarchical classification:

[0176] First-level classification: rough classification by data source (industrial equipment, smart city, medical equipment, etc.);

[0177] Second-level semantic clustering: Using the cosine similarity algorithm, data is subdivided into subcategories based on semantic features (such as "equipment failure warning" and "energy consumption optimization suggestions").

[0178] Redundancy screening and utility ordinal generation:

[0179] Calculate data redundancy: through information entropy formula Eliminate redundant data (entropy value > threshold);

[0180] Generate a utility ordinal number (e.g., 1-5, with 1 being the highest priority) based on quality score, semantic relevance, and timeliness.

[0181] Output: classified data set and utility ordinal number.

[0182] 4. Dynamic Value Decay Modeling

[0183] Model selection and parameter initialization:

[0184] Industrial real-time warning data: using P(t)=A(t)P0e -λt , initial value A(0)=1,λ=0;

[0185] Historical analysis data: using P(t)=P0e -λt +C(t).

[0186] Reinforcement Learning Dynamic Optimization:

[0187] Based on real-time transaction data and user feedback, the time-sensitive factor λ and storage cost function C(t) are adjusted through the Q-Learning algorithm;

[0188]

[0189] Output: Dynamically adjusted data value curve.

[0190] 5. Game Pricing and Transaction Execution

[0191] Vickrey auction mechanism:

[0192] Sealed Bid: The buyer submits an anonymous bid to the blockchain smart contract;

[0193] Bid opening and settlement: The highest bidder wins the deal at the second highest bid, and the transaction record is uploaded to the chain (including data hash, timestamp, and price).

[0194] Utility function estimation:

[0195] Fit the utility function U(k,n) through known price data, the formula is: P A =P B ×U(kA ,n) / U(k B ,n)

[0196] Estimate the initial price of unpriced data.

[0197] Reputation System Intervention:

[0198]

[0199] High-reputation users are matched first and have lower handling fees (e.g., fees for users with a rating ≥ 0.8 are reduced by 10%).

[0200] Output: transaction price, on-chain transaction records, and updated credit score.

[0201] 6. Data delivery and post-processing

[0202] Data decryption and delivery:

[0203] After the buyer successfully pays, the smart contract automatically releases the data decryption key;

[0204] The data is transmitted to the buyer via edge computing nodes or the cloud.

[0205] Feedback collection and model iteration:

[0206] The buyer rates the data quality and price rationality (1-5 points);

[0207] The system updates AI model parameters (such as quality assessment weights), dynamic attenuation models, and credit scoring rules based on feedback.

[0208] Output: decrypted data, user feedback, and optimized system parameters.

[0209] In summary, compared with the existing technology, it has the following beneficial effects:

[0210] By introducing time-sensitive factors and dynamically adjusting data value, dynamic pricing can effectively reflect the characteristic that the value of IoT data decays rapidly over time, improve pricing accuracy, filter redundant data through information funnels, increase data value density, estimate data value through the utility ordinal obtained through information funnels, and conduct quality assessment of IoT data through AI models to automatically locate core high-value data subsets.

[0211] Through intelligent quality assessment and information funnel theory, a unified data quality assessment and pricing standard is established to reduce the value perception differences between supply and demand sides. Redundant data is filtered through the information funnel to improve data value density and reduce the value perception differences between supply and demand sides. Through the Vickrey auction mechanism, efficient matching between supply and demand sides is achieved to promote data transactions.

[0212] Thus, although the invention has been described herein with reference to specific embodiments thereof, freedom of modification, various changes and substitutions are contemplated within the foregoing disclosure, and it should be understood that in some cases, some features of the invention will be employed without the corresponding use of other features without departing from the scope and spirit of the claimed invention. Thus, many modifications may be made to adapt a particular environment or material to the true scope and spirit of the invention. The invention is not intended to be limited to the specific terminology used in the claims below and / or to the specific embodiments disclosed as the best mode contemplated for carrying out the invention, but the invention is intended to include any and all embodiments and equivalents falling within the scope of the appended claims. Thus, the scope of the invention will be determined solely by the appended claims.

Claims

1. The IoT data value recognition system based on dynamic game theory is characterized by: The system includes: Intelligent quality assessment module: used to assess the quality of IoT data through a general artificial intelligence model and screen high-value data subsets; Information funnel processing module: connected to the intelligent quality assessment module, used to classify and filter redundancy of data, and generate the utility ordinal number of each data; Dynamic value decay modeling module: connected to the information funnel processing module, builds a model for data value decay over time based on time-sensitive factors, and dynamically adjusts data value; Game pricing execution module: connected to the dynamic value decay modeling module, estimates data prices through the Vickrey auction mechanism and utility function to achieve supply and demand matching.

2. The IoT data value recognition system based on dynamic game theory according to claim 1 is characterized in that: The intelligent quality assessment module includes: Data acquisition unit, used to collect multimodal data from IoT devices in real time; The multimodal feature fusion unit processes text, image and sensor data respectively through the Transformer model, convolutional neural network and time series analysis to generate a unified feature vector.

3. The IoT data value recognition system based on dynamic game theory according to claim 2 is characterized in that: The data collection unit includes structured data, unstructured data and sensor stream data.

4. The IoT data value recognition system based on dynamic game theory according to claim 1 is characterized in that: The information funnel processing module includes: Semantic enhancement classification unit, clustering multimodal features based on hierarchical naming rules and semantic similarity algorithm (cosine similarity); The dynamic redundancy screening unit calculates data redundancy through information entropy. The formula is: Among them, P i The entropy value is the proportion of a certain type of data in the total data set. When the entropy value is higher than the threshold, redundant elimination is triggered.

5. The IoT data value recognition system based on dynamic game theory according to claim 1 is characterized in that: The dynamic value decay modeling module is configured as follows: The reinforcement learning algorithm is used to dynamically adjust the time-sensitive factor λ and the storage cost function C(t), and the reward function is designed as:

6. The IoT data value recognition system based on dynamic game theory according to claim 5 is characterized in that: The dynamic value decay modeling module adopts the following model for industrial real-time warning data scenarios: P(t)=A(t)P0e -λt Where A(t) is a monotonically increasing function satisfying A(0)=1.

7. The IoT data value recognition system based on dynamic game theory according to claim 5 is characterized in that: The dynamic value decay modeling module uses the following model for historical analysis data scenarios: P(t)=P0e -λt +C(t) Where P0 represents the initial value of the data, λ represents the attenuation coefficient (the degree of data depreciation), and C(t) is a monotonically increasing function.

8. The IoT data value recognition system based on dynamic game theory according to claim 1 is characterized in that: The game pricing execution module includes: A blockchain-powered smart contract unit for executing the Vickrey auction process, including sealed bidding, on-chain bid opening, and transaction record storage; The decentralized reputation system calculates the reputation score based on historical transaction data. The formula is: Users with high credibility will be matched first and enjoy fee discounts.

9. The IoT data value recognition system based on dynamic game theory according to claim 2 is characterized in that: The intelligent quality assessment module also includes: The data encryption unit is used to encrypt the transmission and storage of collected IoT data. The encryption algorithm uses a combination of AES-256 and homomorphic encryption to ensure the privacy and security of data during transmission and calculation.

10. The IoT data value recognition system based on dynamic game theory according to claim 8 is characterized in that: The blockchain-driven smart contract unit includes: The dynamic parameter adjustment unit automatically adjusts auction parameters based on real-time market supply and demand data. The dynamic parameter adjustment unit includes: Reserve price: dynamically set based on the median of historical transaction prices; Bidding time window: shorten or extend according to the urgency of the transaction; Transaction fee rate: Inversely correlated with the user's credit score, with lower fees for users with high credit rating.

Images