Task demand sensitive auction model method of computing power block chain network

Abstract

The invention discloses a task demand sensitive auction model method of a computing power block chain network. Design and mathematical modeling are carried out for a computing power transaction system of the computing power block chain network. Establishing a computing power block chain network task confidentiality sensitive transaction model, a time delay sensitive transaction model and a price sensitive transaction model; integrating the three sub-models to quantify buyer utility; quantifying the computing power seller utility model; traversing all buyers and sellers, integrating the utility of the buyers and sellers of the system, and establishing a social welfare model of the system; all the utility models are integrated, and a task demand sensitive auction model of block chain network multi-objective optimization is established; an NSGA-II algorithm is improved based on a matching theory, and a fast auction solving algorithm is designed. According to the method, a series of challenges of low auction efficiency, unbalanced resource supply and demand, ineffective cost control, difficulty in meeting diversified task requirements and the like in existing block chain computing power transactions are solved, and the method has important theoretical significance and wide application prospects.

Description

TECHNICAL FIELD

[0001] The present application relates to a task demand sensitive auction model method of a computing power blockchain network, belonging to the technical field of blockchain. BACKGROUND

[0002] In recent years, with the rapid development and wide application of 5G technology, Internet technology and VR / AR technology, the connection scale, data processing capacity and intelligent level of the Internet of Things have been significantly improved, and the global Internet of Things device market size has achieved a leap-forward growth. Real-time data generated by a large number of terminal devices puts higher requirements on localized computing capacity. In order to complete the calculation in time, more and more devices with insufficient computing power resources purchase computing power from devices with idle computing resources, and a large amount of computing power resources are transferred and traded between different computing power devices. However, in the distributed Internet of Things environment, the supply and demand relationship of computing power resources is becoming more and more complex, and the traditional computing power trading mode is facing a series of challenges such as low trading efficiency, imbalance between supply and demand of resources, failure of cost control and difficulty in meeting diversified task requirements. At present, it is urgent to propose a resource trading method of a dynamic computing power network that can improve trading efficiency and meet the different needs of devices. SUMMARY

[0003] The present application provides a task demand sensitive auction model method of a computing power blockchain network, which further meets the diversified task requirements and improves the resource matching efficiency on the basis of ensuring the rationality, balance and authenticity of the transaction between buyers and sellers.

[0004] Technical scheme: In order to achieve the above-mentioned purpose, the technical scheme adopted by the present application is:

[0005] A task demand sensitive auction model method of a computing power blockchain network, comprising the following steps:

[0006] Step 1, collecting buyer and transaction information, and designing and mathematically modeling the computing power transaction system of the computing power blockchain network according to the buyer and transaction information.

[0007] Step 2, modeling the geographic location of the seller in the computing power transaction system to establish a dispersion factor, and establishing a task confidential sensitive transaction model of the computing power blockchain network.

[0008] Step 3, modeling the transmission delay and calculation delay based on the parallel computing mechanism in the computing power transaction system, and establishing a task delay sensitive transaction model of the computing power blockchain network.

[0009] Step 4, establishing a task price sensitive transaction model of the computing power blockchain network according to the cost, historical cooperation times and other factors in the computing power transaction system.

[0010] Step 5, integrate the buyer task confidentiality sensitive transaction model, the time delay sensitive transaction model and the price sensitive transaction model to quantify the buyer utility model.

[0011] Step 6, according to the unit computing power selling price of the seller in the computing power transaction system and the unit data calculation cost, quantify the computing power seller utility model.

[0012] Step 7, traverse all buyers and sellers, integrate the system buyer utility model and the seller utility model, and establish the system social welfare model.

[0013] Step 8, according to the buyer utility model, the seller utility model and the system social welfare model, establish the task demand sensitive auction model of the multi-objective optimization of the blockchain network.

[0014] Step 9, based on the matching theory, improve the NSGA-II algorithm, design a fast auction solving algorithm to solve the task demand sensitive auction model, and obtain the optimal solution of the computing power resource allocation.

[0015] Preferably, the task demand sensitive auction model is as follows:

[0016]

[0017] wherein, represents the task demand sensitive auction model, is the number of devices in the system applying to become a buyer, is the number of devices in the system applying to become a seller, represents the buyer e task confidentiality sensitive degree weight factor power, represents the buyer e task time delay sensitive degree weight factor power, represents the buyer e task price sensitive degree weight factor power, represents the buyer task confidentiality sensitive auction model, represents the buyer task time delay sensitive auction model, represents the buyer task price sensitive auction model, represents the unit computing power selling price of the seller to the buyer , is the computing power cost of the seller to the buyer per unit data calculation, represents the buyer allocation to the seller the order data volume, denotes the buyer the sum of the collaboration task data volume between the seller and the buyer the minimum requirement of the buyer for the task confidentiality, the maximum requirement of the buyer for the task latency, the maximum requirement of the buyer for the task price, denotes the element of the matching matrix .

[0018] Preferably, the system social welfare model is:

[0019]

[0020] wherein, denotes the social welfare of the whole system.

[0021] Preferably, the buyer utility model in step 5 is as follows:

[0022]

[0023] wherein, denotes the total utility of the buyer , denotes the buyer 's task confidentiality sensitivity weight factor , denotes the buyer 's task latency sensitivity weight factor , denotes the buyer 's task price sensitivity weight factor .

[0024] Preferably, the method of quantifying the computational power seller utility model in step 6 is as follows:

[0025] The revenue obtained by the seller from the transaction with the buyer is:

[0026]

[0027] The seller utility of a single transaction between the seller and the buyer is modeled as:

[0028]

[0029] The overall transaction utility of the seller is the sum of the individual seller utility after traversing all the buyers:

[0030]

[0031] wherein, represents the sum of the seller utility, represents the seller utility of a single transaction between the buyer and the seller.

[0032] Preferred: the method for establishing the task machine confidential sensitive transaction model of the computing power block chain network in step 2:

[0033] The position of the seller is ( , ), the position of the seller is ( , ), and the longitude difference and the latitude difference between the two are respectively:

[0034]

[0035]

[0036] The shortest surface distance is:

[0037]

[0038] wherein, is the average radius of the earth, and the shortest surface distance characterizes the spatial position relationship between the sellers.

[0039] The discriminant function is defined as:

[0040]

[0041] wherein, is the confidentiality ability of the seller to the buyer , is the confidentiality ability of the seller to the buyer .

[0042] The dispersion factor between the seller and the seller is:

[0043]

[0044] wherein, is the distance weight factor, is the buyer The order data volume allocated to the seller , The order data volume allocated to the seller , The order data volume allocated to the seller , the seller's privacy ability is the same, Encourage task allocation to sellers with greater distance.

[0045] The transaction model based on confidentiality sensitivity is:

[0046]

[0047] Maximize the confidentiality sensitive transaction model to get the final optimization model:

[0048]

[0049] Among them, Indicates the confidentiality sensitive transaction model of the buyer .

[0050] Preferably, the method for establishing the task delay sensitive transaction model of the block chain network in step 3 is:

[0051] The network bandwidth of the current seller Relative to the buyer Is , the communication signal to noise ratio is :

[0052]

[0053] Among them, Is the signal power at this time, Is the noise power at this time.

[0054] The data transmission rate between the buyer And the seller :

[0055]

[0056] The transmission time required for the task data to be sent from the buyer To the seller :

[0057]

[0058] The calculation formula of the data volume required by the seller After calculation and returned to the buyer :

[0059]

[0060] wherein, is a processing factor embodying the degree of compression or summarization of data.

[0061] The computing task data is from the seller to the buyer The transmission time required is:

[0062]

[0063] The single seller to the buyer The total latency of the transmission part is:

[0064]

[0065] Each seller node synchronously processes sub-tasks based on a parallel computing mechanism, the buyer allocates data to the seller The amount of data is processed in parallel, and in the current system, the seller may not only cooperate with a single buyer, but also may simultaneously and in parallel process data requests of multiple buyers, and the computing latency is:

[0066]

[0067] wherein, is the total number of parallel tasks of the seller is the number of parallel operations, is the processor performance of the seller itself. For a single seller , from accepting the data task of the buyer

[0068] to computing the data task to finally returning the result to the buyer , the total latency is the sum of the transmission latency and the computing latency:

[0069]

[0070]

[0071] The timeliness capability of the seller to the buyer is:

[0072]

[0073] wherein, is a latency weight factor representing the degree of influence of latency on timeliness capability.

[0074] ​​Based on the above conditions, the time delay sensitive transaction model is established as:

[0075]

[0076] For the buyer , the maximum minimum time delay sensitive transaction model is finally optimized as:

[0077]

[0078] Wherein, represents the buyer time delay sensitive transaction model.

[0079] Preferably, the method for establishing the computing power blockchain network task price sensitive transaction model in step 4 is:

[0080] The unit computing power selling price of the seller to the buyer is recorded as , and the unit computing power selling price base offer of the seller to the buyer is :

[0081]

[0082] Wherein, is a pricing coefficient adjustment factor, is the computing power cost of the seller to the buyer for each unit of data calculation.

[0083] The historical cooperation times of the seller and the buyer is , and the offer after being affected by the historical cooperation times is , wherein, represents the response coefficient of the historical cooperation times.

[0084] The influence of other factors except the computing cost and the historical cooperation times is quantified as , and the unit computing power selling price of the seller to the buyer is:

[0085]

[0086] Wherein, is a coefficient for controlling the influence degree of other influencing factors.

[0087] Based on the above conditions, after iterating through all sellers, a latency-sensitive transaction model is established as follows:

[0088]

[0089] in, It is a constant much smaller than 1.

[0090] From computing power buyers From this perspective, a trading model with minimizing price sensitivity as its core is constructed, and the final optimized model is as follows:

[0091]

[0092] in, Indicates buyer Price-sensitive trading models.

[0093] Preferred method: The method for mathematical modeling the computing power trading system of the computing power blockchain network in step 1:

[0094] The system has Each device applicant becomes a buyer, and the buyer group is... , Indicates the serial number is The total number of task requests from all buyers is [number]. Buyer The number of task requests is , It can be further divided into A resource demand group, for any buyer Buyer With the seller Data volume of collaborative tasks The sum, . , , , It is a requirement regarding confidentiality capabilities. It is a requirement in terms of timeliness and efficiency. It's a requirement regarding price. Buyer Weighting factors for the degree of sensitivity to mission secrecy. Buyer Minimum requirements for the level of confidentiality required for this mission. Buyer Weighting factors for the sensitivity of task latency. Buyer The highest requirements for the latency sensitivity of this task. Buyer a weight factor of the degree of sensitivity of the task to price, is the buyer the highest requirement of the degree of sensitivity of the task to price.

[0095] is the buyer the lowest requirement of the degree of sensitivity of the task to confidentiality is calculated by the following function:

[0096]

[0097] wherein, is a regulation factor of the degree of influence of personal willingness, is the quantified value of the personal willingness of the buyer, is a regulation factor of the degree of influence of the consequence of data leakage, is the influence of the lowest requirement of the consequence of data leakage to the confidentiality of the task, is a regulation factor of the degree of influence of industry standard, is the influence of the lowest requirement of the industry standard to the confidentiality of the task.

[0098] is the buyer the highest requirement of the degree of sensitivity of the task to time delay is calculated by the following function:

[0099]

[0100] wherein, is a regulation factor of the degree of influence of time tolerance, is the quantified value of the time tolerance of the buyer, is a regulation factor of the degree of influence of the urgency of the task, represents the urgency of the task.

[0101] is the buyer the highest requirement of the degree of sensitivity of the task to price is calculated by the following function:

[0102]

[0103] wherein, is a regulation factor of the degree of influence of the maximum acceptable purchase price of the individual, is the value of the maximum acceptable purchase price of the buyer.

[0104] There are devices in the system that apply to become a seller, and the set of sellers is , is the seller with serial number , and the seller The value of idle computing power available for trading is expected to be The total computing power value owned is The environmental error rate for quantifying the actual deliverable transaction value and idle value error is:

[0105]

[0106] Wherein, The environmental error rate for quantifying the actual deliverable transaction value and idle value error, The environmental impact weight, The environmental impact factor. The seller The actual deliverable computing power value Is:

[0107]

[0108] The buyer The task order data volume allocated to the seller Is affected by the actual deliverable computing power value of the seller , The greater, The greater, the mapping relationship is:

[0109]

[0110] Wherein, The adjustment factor for controlling the maximum value of the task order data volume, The parameter for controlling the nonlinear growth speed of the computing resource quantity to the computing resource, The parameter for describing the influence degree of the computing resource quantity to the data volume, The adjustment factor for controlling the influence degree of the logarithmic term, The adjustment factor for controlling the influence of the exponential decay term.

[0111] The seller The confidential ability of the buyer , the timeliness ability is , the unit computing power selling price is , in the current auction system, there are Possible pairing ways, the manager device records the transaction situation of the system buyer and seller as a Matching matrix , Each element In the matrix takes the value:

[0112]

[0113] When the buyer​​​ with the seller when a match is successful, when there is no match, .

[0114]

[0115] The matching matrix of the system is .

[0116] Another object of the present application is to provide a task demand sensitive auction system of a computing power blockchain network, which is used to realize a task demand sensitive auction model method of the computing power blockchain network, and comprises an input unit, a modeling unit, a confidential sensitive transaction model unit, a time delay sensitive transaction model unit, a price sensitive transaction model unit, a buyer utility model unit, a seller utility model unit, a system social welfare model unit, a task demand sensitive auction model unit, a solving unit, and an output unit.

[0117] The input unit is used to input the buyer and transaction information.

[0118] The modeling unit is used to design and mathematically model the computing power transaction system of the computing power blockchain network according to the buyer and transaction information.

[0119] The confidential sensitive transaction model unit is used to model the geographical location modeling dispersion factor of the seller in the computing power transaction system, and establish a task confidential sensitive transaction model of the computing power blockchain network.

[0120] The time delay sensitive transaction model unit is used to model the transmission time delay and the calculation time delay based on the parallel computing mechanism in the computing power transaction system, and establish a task time delay sensitive transaction model of the computing power blockchain network.

[0121] The price sensitive transaction model unit is used to establish a task price sensitive transaction model of the computing power blockchain network according to the cost, the number of historical cooperation times and other factors in the computing power transaction system.

[0122] The buyer utility model unit is used to integrate the buyer task confidential sensitive transaction model, the time delay sensitive transaction model and the price sensitive transaction model to quantify the buyer utility model.

[0123] The seller utility model unit is used to quantify the computing power seller utility model according to the unit computing power selling price and the unit data calculation cost in the computing power transaction system.

[0124] The system social welfare model unit is used to traverse all the buyers and sellers, comprehensively establish the system social welfare model by combining the buyer utility model and the seller utility model.

[0125] The task demand sensitive auction model unit is used for establishing a task demand sensitive auction model of multi-objective optimization of a blockchain network according to a buyer utility model, a seller utility model and a system social welfare model.

[0126] The solving unit is used for solving the task demand sensitive auction model by using a fast auction solving algorithm based on the improved NSGA-II algorithm based on the matching theory, so as to obtain an optimal solution of the computing power resource allocation.

[0127] The output unit is used for outputting the optimal solution of the computing power resource allocation.

[0128] Compared with the prior art, the present application has the following beneficial effects:

[0129] The present application provides a task demand sensitive auction model method of a computing power blockchain network. Firstly, a trusted transaction system based on end-edge-cloud collaboration is proposed. The system is composed of three core components: computing power buyer devices, computing power seller devices and computing power manager devices. Secondly, three transaction sub-models sensitive to task confidentiality, task delay and task price are constructed for different task demands. Thirdly, a comprehensive auction model sensitive to task demand is constructed by quantifying the buyer utility, the seller utility and the social welfare, which is used to grasp the importance of the buyer to the seller's confidentiality, timeliness and selling price, and to balance the interests of all parties. Finally, the improved NSGA-II algorithm is used to quickly solve the auction, efficiently solving the complex resource matching problem and ensuring the fairness and execution efficiency of the transaction. The present application constructs an innovative task demand sensitive auction model, provides an effective theoretical framework for resource transaction in the computing power blockchain network, and provides a feasible solution for dynamic resource allocation considering task demand. BRIEF DESCRIPTION OF DRAWINGS

[0130] Figure 1 The present application is a computing power transaction flowchart;

[0131] Figure 2 The present application is a computing power transaction scene diagram;

[0132] Figure 3 The present application is a social welfare comparison diagram;

[0133] Figure 4 The present application is a buyer utility comparison diagram. DETAILED DESCRIPTION

[0134] The present application will be further illustrated below in combination with the drawings and specific embodiments, and it should be understood that these examples are only used to illustrate the present application and not to limit the scope of the present application. After reading the present application, those skilled in the art can make various modifications to the equivalent forms of the present application, which fall within the scope defined by the appended claims.

[0135] This embodiment provides a task-demand-sensitive auction model method for a computing power blockchain network. It is a blockchain-driven, dual-feedback, trusted computing power transaction method for the Industrial Internet, designed and mathematically modeled for a computing power blockchain network's computing power transaction system. It establishes a task confidentiality-sensitive transaction model based on a dispersion factor modeled according to the seller's geographical location; a latency-sensitive transaction model based on parallel computing mechanisms modeling transmission and computation latency; a price-sensitive transaction model considering cost, historical cooperation frequency, and other factors; a buyer utility model integrating three sub-models: a task confidentiality-sensitive model, a latency-sensitive model, and a price-sensitive model; a seller utility model considering the seller's unit computing power price and unit data computation cost; a system social welfare model by traversing all buyers and sellers and comprehensively considering the system's buyer and seller utilities; and a multi-objective optimization task-demand-sensitive auction model for the blockchain network by integrating various utility models. Finally, a fast auction solution algorithm is designed based on an improved NSGA-II algorithm using matching theory. Figure 1 As shown, the specific steps include:

[0136] Step 1: Design and mathematically model the computing power trading system for the computing power blockchain network, such as... Figure 2 As shown, suppose the system has Each device applicant becomes a buyer, and the buyer group is... , Indicates the serial number is The total number of task requests from all buyers is [number]. Buyer The number of task requests is , It can be further divided into A resource demand group, for any buyer Buyer With the seller Data volume of collaborative tasks The sum, expressed as ; , and It is a collection of the buyer's needs in various aspects. It is a requirement regarding confidentiality capabilities. It is a requirement in terms of timeliness and efficiency. It's a requirement regarding price. , Buyer Weighting factors for the degree of sensitivity to mission secrecy. Buyer Minimum requirements for the level of confidentiality required for this mission; , Buyer a weight factor of the degree of sensitivity of the task to time delay, Buyer the highest requirement of the degree of sensitivity of the task to time delay; , Buyer a weight factor of the degree of sensitivity of the task to price, Buyer the highest requirement of the degree of sensitivity of the task to price;

[0137] affected by external factors such as the buyer's personal will, the consequences of data leakage, and industry standards, the calculation is performed by the following function:

[0138]

[0139] wherein, is a regulation factor for controlling the degree of influence of personal will, is the quantitative value of the buyer's personal will, is a regulation factor for controlling the degree of influence of the consequences of data leakage, is the minimum requirement of the influence of the consequences of data leakage on the sensitivity of the task to confidentiality, is a regulation factor for controlling the degree of influence of industry standards, is the minimum requirement of the influence of the industry standards involved in the task on the sensitivity of the task to confidentiality;

[0140] affected by the buyer's personal time tolerance and the urgency of the task, the calculation is performed by the following function:

[0141]

[0142] wherein, is a regulation factor for controlling the degree of influence of time tolerance, is the quantitative value of the buyer's personal time tolerance, is a regulation factor for controlling the degree of influence of the urgency of the task, represents the urgency of the task;

[0143] affected by the maximum purchase price acceptable to the buyer, the calculation is performed by the following function:

[0144]

[0145] wherein, is a regulation factor for controlling the degree of influence of the maximum purchase price acceptable to the buyer, is the value of the maximum purchase price acceptable to the buyer;

[0146] Let there be a system A device applies to become a seller, and a set of sellers is , The seller is a seller with a serial number The seller predicts that the idle computing power value available for trading is , the total computing power value owned is , and the environmental error rate used to quantify the actual deliverable transaction value and idle value error is:

[0147]

[0148] wherein, is the environmental impact weight, is the environmental impact factor. The seller actual deliverable computing power value is:

[0149]

[0150] The buyer allocates the task order data volume to the seller is affected by the actual deliverable computing power value of the seller , the greater, the greater, and the mapping relationship is:

[0151]

[0152] wherein, is the adjustment factor that controls the maximum value of the task order data volume, is a parameter that controls the nonlinear growth rate of the computing resource amount to the computing resource, is a parameter that describes the degree of influence of the computing resource amount on the data volume, is an adjustment factor that controls the degree of influence of the logarithmic term, is an adjustment factor that controls the influence of the exponential decay term;

[0153] The seller has a confidential ability to the buyer , a time-sensitive ability , and a unit computing power selling price . In the current auction system, there are possible pairing methods, and the manager device records the transaction situation of the system buyers and sellers as a matching matrix , each element of which takes a value: ​​

[0154]

[0155] When the buyer matches with the seller , no match, ;

[0156]

[0157] The matching matrix of the system is .

[0158] Step 2, according to the geographical location of the seller to model the dispersion factor, establish the computing power block chain network task machine confidential sensitive transaction model, the location of the seller is ( , ), the location of the seller is ( , ), the longitude difference and the latitude difference between the two are:

[0159]

[0160]

[0161] The shortest ground distance is:

[0162]

[0163] Where, is the average radius of the earth, and the shortest ground distance characterizes the spatial relationship between the sellers;

[0164] The discriminant function is defined as:

[0165]

[0166] Where, is the confidentiality ability of the seller to the buyer , and is the confidentiality ability of the seller to the buyer ;

[0167] The dispersion factor between the seller and the seller is:

[0168]

[0169] wherein, is a distance weight factor, is a buyer assigned to a seller , is a buyer assigned to a seller , and when the seller privacy capability is the same, the task is encouraged to be assigned to the seller with a larger distance;

[0170] The transaction model based on the confidentiality sensitivity is:

[0171]

[0172] The final optimization model is obtained by maximizing the confidentiality-sensitive transaction model:

[0173]

[0174] Constraint condition C1 specifies that the sum of the order data volume assignments should be equal to the total order data volume of the buyer, constraint condition C2 specifies that the system confidentiality sensitivity cannot be lower than the minimum confidentiality sensitivity value acceptable by the buyer, constraint condition C3 specifies that the maximum number of matches of the entire system should not exceed , and constraint condition C4 indicates that the number of seller cooperations of the buyer cannot exceed the maximum number of sellers in the system.

[0175] Step 3, based on the parallel computer mechanism, the transmission delay and the calculation delay are modeled, the task delay-sensitive transaction model of the computing power blockchain network is established, and the current seller is relative to the network bandwidth of the buyer , , and the communication signal-to-noise ratio is :

[0176]

[0177] wherein, is the signal power at this time, is the noise power at this time;

[0178] The data transmission rate between the buyer and the seller is:

[0179]

[0180] The transmission time required for the task data to be sent from the buyer to the seller is:

[0181]

[0182] Seller The data amount needed to be returned to the buyer after calculation The calculation formula is

[0183]

[0184] Wherein, is the processing factor reflecting the compression or aggregation degree of data;

[0185] The transmission time needed for the calculation task data to be returned and sent to the buyer from the seller The total time delay for the seller to transmit part of the buyer is:

[0186]

[0187] The total time delay for the single seller to transmit part of the buyer is:

[0188] Each seller node synchronously processes sub-tasks based on parallel computing mechanism, and the buyer

[0189] allocates the data amount to the seller for parallel processing. Since the seller may not only cooperate with a single buyer, but also may simultaneously and in parallel process data requests of multiple buyers in the current system, the calculation time delay is:

[0190] Wherein, is the total number of parallel tasks of the seller,

[0191] is the number of parallel operations, is the processor performance of the seller itself; For the single seller, the total time delay from accepting the data task of the buyer to calculating the data task to finally returning the result to the buyer is the sum of the transmission time delay and the calculation time delay:

[0192] The time efficiency ability of the seller to the buyer is:

[0193]

[0194]

[0195] The time efficiency ability of the seller to the buyer is: ​​​​​​​

[0196]

[0197] wherein, is a latency weight factor, representing the degree of influence of latency on timeliness capability;

[0198] The above conditions can be integrated to establish a latency-sensitive transaction model as follows:

[0199]

[0200] For the buyer, the degree of latency sensitivity depends on the latency of the slowest sub-task performed internally, and the smaller the total latency sensitivity, the more in line with the buyer's transaction expectations. Here, the maximum and minimum latency-sensitive transaction model is minimized, and the final optimization model is:

[0201]

[0202] Constraint condition C1 specifies that the sum of order data volume allocation should be equal to the total order data volume of the buyer, constraint condition C2 specifies that the system latency sensitivity cannot be higher than the highest latency sensitivity value that the buyer can accept, constraint condition C3 specifies that the maximum number of matches of the entire system should not exceed , and constraint condition C4 indicates that the number of sellers that the buyer cooperates with cannot exceed the maximum number of sellers in the system.

[0203] Step 4, considering cost, historical cooperation times, and other factors, a computing power blockchain network task price-sensitive transaction model is established, and the unit computing power selling price of the seller to the buyer is denoted as , and the unit computing power selling price base bid of the seller to the buyer is :

[0204]

[0205] wherein, is a pricing coefficient adjustment factor, is the computing power cost of the seller for each unit of data calculation of the buyer ;

[0206] The historical cooperation times of the seller with the buyer are , and the bid after being affected by the historical cooperation times can be expressed as , wherein, represents the response coefficient of historical cooperation times;

[0207] Quantify the influence of other factors in addition to the calculation cost and the number of historical cooperation times , considering all the above factors, the seller The unit computing power selling price of the buyer is:

[0208]

[0209] Among them, is the coefficient of the degree of influence of other influencing factors;

[0210] Based on the above conditions, all sellers are traversed to establish a transaction model based on latency sensitivity:

[0211]

[0212] Among them, is a constant much smaller than 1;

[0213] From the perspective of the computing power buyer , the goal is to obtain the most computing power resources at the lowest cost by cooperating with sellers with lower prices. Therefore, the smaller the price sensitivity value, the more in line with the buyer's transaction expectations. Based on this, the price sensitivity minimization is constructed as the core of the transaction model, and the final optimization model is:

[0214]

[0215] Constraint condition C1 specifies that the total order data volume allocation should be equal to the total order data volume of the buyer, constraint condition C2 specifies that the system price sensitivity cannot be higher than the highest price sensitivity value that the buyer can accept, constraint condition C3 specifies that the maximum number of matches in the entire system should not exceed , and constraint condition C4 indicates that the number of cooperating sellers cannot exceed the maximum number of sellers in the system.

[0216] Step 5, integrate the buyer task confidential sensitive model, latency sensitive model and price sensitive model three sub-models to quantify the buyer's utility, the total utility of a single buyer is:

[0217]

[0218] Step 6, considering the unit computing power selling price of the seller and the unit data calculation cost, quantifying the computing power seller utility model, the income of the seller from the transaction with the buyer is:

[0219]

[0220] ​The seller utility of a single transaction between a seller and a buyer is modeled as:

[0221]

[0222] In the transaction process proposed in the application, the seller can sell computing power resources to multiple buyers at the same time, cooperate with multiple buyers, and the overall transaction utility of the seller is the sum of the single seller utility after traversing all buyers, which is:

[0223]

[0224] Step 7, traverse all seller buyers, integrate system buyer seller utility, establish system social welfare model, the overall system social welfare is the sum of all buyer utility and all seller utility, traverse all buyers to get the buyer part of the social welfare:

[0225]

[0226] Traverse all seller devices to get the seller part of the social welfare:

[0227]

[0228] The overall system social welfare is:

[0229]

[0230]

[0231] Step 8, integrate various utility models to establish a task demand sensitive auction model for the multi-objective optimization of the blockchain network, and based on the above content, the task demand sensitive auction model for the computing power blockchain network can be established as:

[0232]

[0233] Constraint condition C1 indicates that the total sum of order data volume allocation should be equal to the total order data volume of the buyer, constraint condition C2 indicates that the system confidentiality sensitivity cannot be higher than the minimum confidentiality sensitivity value that the buyer can accept, constraint condition C3 indicates that the system latency sensitivity cannot be higher than the maximum latency sensitivity value that the buyer can accept, constraint condition C4 indicates that the system price sensitivity cannot be higher than the maximum price sensitivity value that the buyer can accept, constraint condition C5 indicates that the maximum number of matches in the entire system should not exceed , and constraint condition C6 indicates that the number of seller buyers cooperating with the buyer cannot exceed the maximum number of sellers in the system.

[0234] Step 9, improve the NSGA-II algorithm based on the matching theory and design a fast auction solution algorithm.

[0235] The solving algorithm comprises the following steps:

[0236] S1, randomly generating an initial population of the outer layer ;

[0237] S2, establishing a weight vector according to a weight factor of the confidentiality sensitivity , a weight factor of the time delay sensitivity , and a weight factor of the price sensitivity ; ;

[0238] S3, calculating a matching degree score according to the confidentiality capability , the time delay capability , and the unit computing power selling price ; ;

[0239] S4, generating a candidate matching list in descending order of the matching degree score under the condition that the computing capability of the seller is not exceeded, and obtaining an updated initial population;

[0240] S5, calculating a dispersion factor , a time delay capability , and a computing power selling price for the task quantity allocation result of each initial population, and then calculating , , ;

[0241] S6, judging whether the auction constraint conditions C2, C3 and C4 are met, and if all the conditions are met, retaining the allocation scheme, otherwise replacing the device and updating until all the conditions are met;

[0242] S7, calculating a buyer utility , and generating an inner layer Pareto frontier solution set;

[0243] S8, calculating a seller utility and a social welfare ;

[0244] S9, performing fast non-dominated sorting, calculating crowding degree, and selecting individuals with better fitness;

[0245] S10, constructing a two-parent individual task allocation bipartite graph , and comparing to form a maximum common matching subgraph ;

[0246] S11, adding normal distribution disturbance, updating the task allocation of the non-common part by using hybrid normal crossover, and selecting the optimal matching part and the hybrid of the non-common part from the parents to generate two offspring;

[0247] S12, the offspring variation generates an offspring population, and the parent and offspring populations are merged;

[0248] S13, screening the top 50% individuals from the merged population to form a new population, repeating the process of S6 to S12, constantly iterating until reaching a termination threshold or solution convergence, and outputting the optimal solution set;

[0249] S14, selecting the social welfare largest solution from the converged solution set to generate a matching matrix , a data allocation matrix .

[0250] In another embodiment of the present application, a task demand sensitive auction system of a computing power blockchain network is provided for implementing a task demand sensitive auction model method of the computing power blockchain network, comprising an input unit, a modeling unit, a confidential sensitive transaction model unit, a time delay sensitive transaction model unit, a price sensitive transaction model unit, a buyer utility model unit, a seller utility model unit, a system social welfare model unit, a task demand sensitive auction model unit, a solving unit, and an output unit, wherein:

[0251] The input unit is used to input buyer and transaction information.

[0252] The modeling unit is used to design and mathematically model the computing power transaction system of the computing power blockchain network according to the buyer and transaction information.

[0253] The confidential sensitive transaction model unit is used to model the dispersion factor of the geographic location of the seller in the computing power transaction system, and establish a task confidential sensitive transaction model of the computing power blockchain network.

[0254] The time delay sensitive transaction model unit is used to model the transmission time delay and the calculation time delay based on the parallel computing mechanism in the computing power transaction system, and establish a task time delay sensitive transaction model of the computing power blockchain network.

[0255] The price sensitive transaction model unit is used to establish a task price sensitive transaction model of the computing power blockchain network according to the cost, the number of historical cooperation times and other factors in the computing power transaction system.

[0256] The buyer utility model unit is used to integrate the buyer task confidential sensitive transaction model, the time delay sensitive transaction model and the price sensitive transaction model to quantify the buyer utility model.

[0257] The seller utility model unit is used to quantify the computing power seller utility model according to the unit computing power selling price and the unit data calculation cost of the seller in the computing power transaction system.

[0258] The system social welfare model unit is configured to traverse all buyers and sellers, integrate the system buyer utility model and the seller utility model, and establish a system social welfare model.

[0259] The task demand sensitive auction model unit is configured to establish a task demand sensitive auction model for multi-objective optimization of the blockchain network according to the buyer utility model, the seller utility model, and the system social welfare model.

[0260] The solving unit is configured to improve the NSGA-II algorithm based on the matching theory, design a fast auction solving algorithm to solve the task demand sensitive auction model, and obtain an optimal solution of the computing power resource allocation.

[0261] The output unit is configured to output the optimal solution of the computing power resource allocation.

[0262] As shown in the social welfare comparison of the embodiment Figure 3 As shown in the buyer utility comparison of the embodiment Figure 4 The present application comprehensively considers the dynamic characteristics of multi-dimensional task demand in computing power transaction, accurately depicts the core constraints of buyers by constructing differentiated sub-models, defines the computing power buyer and seller utility and social welfare, constructs a comprehensive auction model, realizes the Pareto optimality of computing power resource allocation, and balances the interests of both supply and demand sides by combining the improved NSGA-II algorithm. The present application solves a series of challenges in the existing blockchain computing power transaction, such as low auction efficiency, imbalance between supply and demand of resources, ineffective cost control, and difficulty in meeting diversified task demands, and has important theoretical significance and wide application prospect.

[0263] The above only describes the preferred embodiments of the present application, and it should be noted that for ordinary skilled persons in the art, several improvements and refinements can be made without departing from the principles of the present application, and these improvements and refinements should also be considered within the protection scope of the present application.

Claims

1. A method for a task demand sensitive auction model of a computing power blockchain network, characterized in that, The method comprises the following steps: Step 1, collecting buyer and transaction information, and designing and mathematically modeling a computing power transaction system of a computing power block chain network according to the buyer and transaction information; Step 2, modeling a dispersion factor of a seller geographic location in the computing power transaction system, and establishing a task secret sensitive transaction model of the computing power block chain network; Step 3, modeling transmission time delay and calculation time delay based on a parallel computing mechanism in the computing power transaction system, and establishing a task time delay sensitive transaction model of the computing power block chain network; Step 4, establishing a task price sensitive transaction model of the computing power block chain network according to a cost, a historical cooperation frequency and other factors in the computing power transaction system; Step 5, integrating a buyer utility model quantified by the task secret sensitive transaction model, the time delay sensitive transaction model and the price sensitive transaction model; Step 6, quantifying a computing power seller utility model according to a unit computing power selling price and a unit data calculation cost in the computing power transaction system; Step 7, traversing all buyers and sellers, comprehensively establishing a system social welfare model by combining the buyer utility model and the seller utility model; Step 8, establishing a task demand sensitive auction model of the block chain network according to the buyer utility model, the seller utility model and the system social welfare model; Step 9, improving an NSGA-II algorithm based on a matching theory, designing a fast auction solving algorithm to solve the task demand sensitive auction model, and obtaining an optimal solution of computing power resource allocation. 2.The method of claim 1, wherein the method further comprises: The task demand sensitive auction model is as follows: wherein, denotes a task requirement sensitive auction model, is the number of devices in the system applying to become buyers, is the number of devices in the system applying to become sellers, denotes a buyer of e a task confidentiality sensitivity weight factor to the power of, denotes a buyer of e a task latency sensitivity weight factor to the power of, denotes a buyer of e a task price sensitivity weight factor to the power of, denotes a buyer a task confidentiality sensitive auction model, denotes a buyer a task latency sensitive auction model, denotes a buyer a task price sensitive auction model, denotes a seller a unit computing power selling price to a buyer is a seller a computing power cost per unit of data calculation to a buyer denotes a buyer an order data volume allocated to a seller denotes a buyer a total of a collaborative task data volume between a buyer and a seller denotes a buyer a minimum requirement for the task confidentiality sensitivity denotes a buyer a maximum requirement for the task latency sensitivity denotes a buyer a maximum requirement for the task price sensitivity denotes an element of a matching matrix .​​​​ 3.The method of claim 2, wherein: The system social welfare model is as follows: where, represents the social welfare of the entire system.

4. The method of claim 3, wherein the method is a task demand sensitive auction model method for the computing power blockchain network. The buyer utility model in step 5 is as follows: wherein, total utility of the buyer, total utility of the buyer, total utility of the buyer, task confidentiality sensitivity weight factor power, total utility of the buyer, task latency sensitivity weight factor power, total utility of the buyer, task price sensitivity weight factor power.

5. The method of claim 4, wherein the method is a task demand sensitive auction model method for the computing power blockchain network. The method for quantifying the computing power seller utility model in step 6 is as follows: revenue earned by a seller from a transaction with a buyer for: The seller utility of a single transaction between the seller and the buyer is modeled as follows: The overall transaction utility of the seller is the sum of the single seller utility after traversing all buyers: wherein, represents the sum of the seller utilities, represents the seller utility of a single transaction between a seller and a buyer. between a seller and a buyer. 6.The method of claim 5, wherein the method further comprises: The method for establishing the task secret sensitive transaction model of the computing power block chain network in step 2 is as follows: Seller The location of the seller is ( , ) and the location of the buyer is ( , ). The longitude difference and latitude difference between the two are:​ Shortest surface distance is: wherein, is the earth's mean radius, the shortest surface distance characterizes the spatial position relationship between the sellers; defining a discriminant function is: wherein is a seller to a buyer of a secret capability, is a seller to a buyer of a secret capability; Seller and the seller between the seller and the buyer is: wherein, is a distance weight factor, is a buyer order data volume assigned to a seller, is a buyer order data volume assigned to a seller, is a buyer order data volume assigned to a seller, when the seller privacy capability is the same, encourage task assignment to a seller with a larger distance; The secret sensitive transaction model is as follows: The final optimization model obtained by maximizing the secret sensitive transaction model is as follows: wherein, represents a buyer confidential sensitive transaction model.

7. The method of claim 6, wherein the method is a task demand sensitive auction model method for the computing power blockchain network. The method for establishing the task time delay sensitive transaction model of the computing power block chain network in step 3 is as follows: Current seller Relative to buyer Network bandwidth is Communication signal-to-noise ratio is : wherein is the signal power at this time, is the noise power at this time; Data transfer rate between the buyer and the seller is: Task data from buyer to seller Required transmission time is: Seller Data amount to be returned to the buyer after calculation The calculation formula of the data amount is : wherein, is a processing factor embodying the degree of compression or summarization of the data; The computing task data is sent from the seller The return is sent to the buyer The required transmission time is: Single seller To the buyer The total latency of the transport section is: Each seller node synchronously processes sub-tasks based on a parallel computing mechanism, and the buyer is assigned a data volume to the seller node for parallel processing. In the current system, the seller may not only cooperate with a single buyer, but also may simultaneously process multiple buyers' data requests in parallel, and the calculation delay is: wherein, is the seller the total number of parallel tasks, is the number of parallel operations, is the seller's own processor performance; For a single seller The total latency from accepting the data task from the buyer to computing the data task to finally returning the result to the buyer is the sum of the transmission latency and the computation latency: Seller To the buyer The time limit is: wherein, is a latency weight factor representing the degree of influence of latency on time-critical capability; The time delay sensitive transaction model established based on the above conditions is as follows: For the buyer The final optimization model for the buyer, which minimizes the latency-sensitive transaction model, is: wherein, represents a buyer latency-sensitive transaction model. 8.The method of claim 7, wherein the method further comprises: determining a number of the plurality of computing resources based on the number of the plurality of computing resources and the number of the plurality of computing resources required by the plurality of computing tasks. The method for establishing the task price sensitive transaction model of the computing power block chain network in step 4 is as follows: The seller The unit computing power selling price of the buyer The unit computing power selling price of the buyer The unit computing power selling price of the buyer The unit computing power selling price of the buyer The unit computing power selling price of the buyer : wherein, is a pricing coefficient adjustment factor, is a seller to a buyer computational power cost per unit of data; Seller Number of historical collaborations with the buyer , the bid after the influence of the number of historical collaborations is represented by , where represents the response coefficient of the number of historical collaborations;​ Quantify the influence of other factors in addition to the calculation cost and the number of historical cooperation times , considering all the above factors, the seller The unit computing power selling price of the buyer is: ​ wherein, is a coefficient controlling the degree of influence of other influencing factors; The time delay sensitive transaction model established based on the above conditions is as follows: wherein is a constant much smaller than 1; From the perspective of the computing power buyer The final optimization model is constructed by minimizing the price sensitivity, which is the core of the transaction model: wherein representing a buyer price sensitive transaction model. 9.The method of claim 8, wherein the method further comprises: The method for mathematically modeling the computing power transaction system of the computing power block chain network in step 1 is as follows: There are devices in the system applying to be buyers, and the set of buyers is , denotes the buyer with the serial number , the total number of task requests of all buyers is , the number of task requests of buyer is , The buyer can be further divided into resource requirement groups, for any buyer, is the sum of the amount of collaborative task data between buyer and seller , ; ; , , , is the requirement in the aspect of confidential ability, is the requirement in the aspect of timeliness ability, is the requirement in the aspect of price, is the weight factor of the buyer sensitivity to the task confidentiality, is the minimum requirement of the buyer sensitivity to the task confidentiality; is the weight factor of the buyer sensitivity to the task delay, is the maximum requirement of the buyer sensitivity to the task delay; is the weight factor of the buyer sensitivity to the task price, is the maximum requirement of the buyer sensitivity to the task price; Buyer The minimum requirement for the sensitivity of the task to the secret The calculation is made by the following function: wherein, is a regulation factor controlling the degree of influence of the individual's will, is a quantitative value of the buyer's individual will, is a regulation factor controlling the degree of influence of the consequences of data leakage, is the influence of the minimum requirement of the consequences of data leakage on the task's confidentiality, is a regulation factor controlling the degree of influence of the industry standard, is the influence of the minimum requirement of the industry standard involved in the task on the task's confidentiality; Buyer the highest degree of sensitivity to the task latency The calculation is made by the following function: wherein, is a regulation factor that controls the degree of influence of the time tolerance, is a quantitative value of the time tolerance of the buyer individual, is a regulation factor that controls the degree of influence of the task urgency, represents the urgency of the task; Buyer The highest requirement for the price sensitivity of the task Calculated by the following function: wherein, is a regulation factor that controls the degree of influence of the maximum purchase price acceptable by the individual, is the value of the maximum purchase price acceptable by the buyer; There are devices in the system apply to become a seller, and the seller set is , The seller with the serial number is a seller, and the seller is expected to have idle computing power values available for trading , and the total computing power value owned is , and the environmental error rate used to quantify the actual deliverable transaction value and idle value error is wherein, is the environmental error rate for actual deliverable transaction value and idle value error, is the environmental impact weight, is the environmental impact factor; seller is the actual deliverable hashpower value is: Buyer Assigned to seller Task order data volume Seller Actual deliverable computing power Influence, The larger, The larger the value, the more accurate the mapping becomes: wherein, is an adjustment factor controlling the maximum value of the data volume of the task order, is a parameter controlling the non-linear growth rate of the amount of computing resources to the amount of computing resources, is a parameter describing the degree of influence of the amount of computing resources on the amount of data, is an adjustment factor controlling the degree of influence of the logarithmic term, is an adjustment factor controlling the influence of the exponential decay term; Seller To the buyer The secret ability is , the timeliness ability is , the unit computing power selling price is , in the current auction system, there are Possible pairing ways, the manager device records the transaction of the system buyer and seller as a Matching matrix , Each element The value is: When the buyer With the seller On successful match, On no match, ; The matching matrix of the system is .

10. An auction system for implementing the task demand sensitive auction model method of the computing power blockchain network of claim 1, characterized in that: The method comprises an input unit, a modeling unit, a secret sensitive transaction model unit, a time delay sensitive transaction model unit, a price sensitive transaction model unit, a buyer utility model unit, a seller utility model unit, a system social welfare model unit, a task demand sensitive auction model unit, a solving unit and an output unit, wherein: The input unit is used for inputting buyer and transaction information; The modeling unit is used for designing and mathematically modeling a computing power transaction system of a computing power block chain network according to the buyer and transaction information; The secret sensitive transaction model unit is used for modeling a dispersion factor of a seller geographic location in the computing power transaction system, and establishing a task secret sensitive transaction model of the computing power block chain network; The time delay sensitive transaction model unit is used for modeling transmission time delay and calculation time delay based on a parallel computing mechanism in the computing power transaction system, and establishing a task time delay sensitive transaction model of the computing power block chain network; The time-sensitive transaction model unit is used for modeling transmission time delay and calculation time delay based on a parallel computing mechanism in a computing power transaction system, and establishing a time-sensitive transaction model of a computing power blockchain network task; The price-sensitive transaction model unit is used for establishing a price-sensitive transaction model of a computing power blockchain network task according to cost, historical cooperation times and other factors in the computing power transaction system; The buyer utility model unit is used for integrating the buyer task secret-sensitive transaction model, the time-sensitive transaction model and the price-sensitive transaction model to quantify a buyer utility model; The seller utility model unit is used for quantifying a computing power seller utility model according to a unit computing power selling price and a unit data calculation cost in the computing power transaction system; The system social welfare model unit is used for traversing all buyers and sellers, comprehensively considering the buyer utility model and the seller utility model, and establishing a system social welfare model; The task demand-sensitive auction model unit is used for establishing a task demand-sensitive auction model of a multi-objective optimization of a blockchain network according to the buyer utility model, the seller utility model and the system social welfare model; The solving unit is used for improving an NSGA-II algorithm based on a matching theory, designing a fast auction solving algorithm to solve the task demand-sensitive auction model, and obtaining an optimal solution of computing power resource allocation; The output unit is used for outputting the optimal solution of computing power resource allocation.

Images