CN109711176A - A Q-Learning-based smart contract validity detection method - Google Patents
A Q-Learning-based smart contract validity detection method Download PDFInfo
- Publication number
- CN109711176A CN109711176A CN201811515288.0A CN201811515288A CN109711176A CN 109711176 A CN109711176 A CN 109711176A CN 201811515288 A CN201811515288 A CN 201811515288A CN 109711176 A CN109711176 A CN 109711176A
- Authority
- CN
- China
- Prior art keywords
- intelligent contract
- promoter
- contract
- external
- decrypt
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 8
- 238000009826 distribution Methods 0.000 claims abstract description 18
- 238000000034 method Methods 0.000 claims abstract description 12
- 238000005457 optimization Methods 0.000 claims abstract description 4
- 230000008569 process Effects 0.000 claims abstract description 4
- 230000006870 function Effects 0.000 claims description 24
- 239000000470 constituent Substances 0.000 claims description 18
- 230000008450 motivation Effects 0.000 claims description 9
- 241000208340 Araliaceae Species 0.000 claims description 4
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims description 4
- 235000003140 Panax quinquefolius Nutrition 0.000 claims description 4
- 239000004744 fabric Substances 0.000 claims description 4
- 235000008434 ginseng Nutrition 0.000 claims description 4
- 230000015572 biosynthetic process Effects 0.000 claims description 2
- 239000012141 concentrate Substances 0.000 claims description 2
- 230000003247 decreasing effect Effects 0.000 claims description 2
- 235000013399 edible fruits Nutrition 0.000 claims description 2
- 238000002474 experimental method Methods 0.000 claims description 2
- 238000010200 validation analysis Methods 0.000 abstract description 6
- 230000008901 benefit Effects 0.000 description 11
- 230000009471 action Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000003993 interaction Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000001186 cumulative effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The intelligent contract validation checking method based on Q-Learning that the invention discloses a kind of, it is intended to which changing currently is mostly the status based on program code loophole to the detection of intelligent contract, meets needs of the intelligent contract as practical application on block chain.It is characterized in that, regard the data feeding process in contract intelligent on block chain as a random distribution process, the effectiveness for defining intelligent contract participant is the function of these random distributions, proposes a kind of new intelligent contract validation checking method；And using the parameter in Q-Learning optimization random distribution, achieve the purpose that optimize participant's utility function.This method quickly and efficiently, has high accuracy, Gao Youhua intensity and robustness, and the present invention is suitable for validity check and secret protection of the intelligent contract as electronic contract.
Description
Technical field
The invention belongs to field of information security technology, it is related to examining using Q-learning algorithm optimization intelligence contract validity
Survey technology.
Background technique
A kind of program code of the intelligent contract as automatic running on block chain, there are many loopholes for itself.Needle at present
Detection method to intelligent contract validity (including correctness and fairness) is mostly based on program code Hole Detection, externally
The feeding research of portion's data is less.In fact, the implementation procedure of intelligent contract is largely dependent upon the feedback of external data
Send the triggering with condition.However there is very big enchancement factor, the shadow of execution and validity for intelligent contract in external data
Sound is very crucial, therefore becomes asking for urgent need to resolve for the validation checking problem of the intelligent contract with external data feeding
One of topic.
These uncertainties can be regarded as being distributed by some stochastic variables and constitute, and these stochastic variables are as data
It is fed to intelligent contract participant.They become the condition for triggering intelligent contract, can influence the operation of intelligent contract, because it
Be constitute participant's effectiveness important component.In intelligent contract, participant needs enough economic motivations, also
It is to say that they only just participate in the execution of intelligent contract in the case where keeping making a profit.It is dynamic for participant in intelligent contract at present
The work of machine research is lacking, and lacks a kind of effective method detection participant's motivation.Therefore, by utility function with
The random distribution of data feeding combines, and utilizes Q-Learning technology, studies the detection of intelligent contract validity and excellent
Change is one of the difficult point in current intelligent contract validation checking technology.
Summary of the invention
The detection method for the intelligent contract validity based on Q-Learning that the object of the present invention is to provide a kind of, feature
Be by specific intelligent contract building, data feeding, parameter optimization and etc. come what is realized, detailed process is as follows:
Step 1: a file is divided into s parts by intelligent contract promoter, and to every a encryption, if it is desired to restore entire file,
Need to decrypt each one's share of expenses for a joint undertaking file；Spending required for decryption is v, because classified document has a timeliness；I.e. in secrecy text
Before part is not announced, value is higher, but after the decryption, value is decreased obviously；Therefore, it is necessary to use one random point
Cloth spends to indicate that its is decrypted, and cannot be indicated using a constant, and Weibull distribution can be used to indicate decryption cost
(value of film)；
Step 2: external constituents download s parts of encryption subfiles, it is intended to decrypt file；
Step 3: intelligent contract randomly selects a random number of s ' from s, wherein s ' < s；S ' is sent to intelligent conjunction by intelligent contract
About promoter allows him to decrypt these subfiles；
Step 4: if these subfiles of intelligent contract promoter successful decryption, intelligent contract waits external data feeding, continues
It executes；If intelligent contract promoter's successful decryption subfile, it will be able to prove that the subfile that remaining is not decrypted is also correct
's；Otherwise, intelligent termination of contract；
Step 5: each external constituents if it is desired to decrypt entire file, needs part payment amount of money m > 1, but this is golden
Volume is less than the cost for decrypting entire file；Decided whether to pay m according to personal preference, if the probability of payment obeys binomial point
Cloth, wherein k is the total number of persons of donations, and n is experiment number, and p is the probability donated every time；
Step 6: the value of m is also to obey random distribution herein, each participant want to decrypt entire file wish and itself
Financial resources condition is different；Because block chain network has small world, i.e., 80% wealth concentrates in 20% manpower, so, m
Pareto distribution is obeyed, the amount of donation m of this distribution expression most people (such as 80%) is 1 between b；
Step 7: each external constituents, after deciding whether donations, intelligent contract collects these amounts of donation km, as
Data are fed in intelligent contract program；
Step 8: the promoter of intelligent contract decides whether to decrypt entire file according to its income, if km > v, illustrate promoter
As soon as income be a positive value, he has the motivation for decrypting entire file；Gap between km and v is bigger, and the motivation of promoter is got over
Greatly；After entire file decryption, sum participant of all downloading encryption subfiles can obtain decryption file；It is worth noting
Be sum >=k, that is to say, that the external constituents for not paying m can sit idle and enjoy the fruits of others' work, and formation is hitchhiked phenomenon；This point be by
What intelligent contract determined, it, can only be them as one because the promoter of intelligent contract can not identify the identity of external constituents
A entirety is treated；If km≤v, illustrate that intelligent contract promoter is profitless, therefore he does not decrypt the motivation of entire file,
He can select non-decrypting；
Step 9: defining the utility function of intelligent contract promoter and external constituents according to the distribution situation of external data；Intelligence
The utility function of energy contract promoter are as follows: km-v, the utility function of external constituents are as follows:, wherein sum is all
External constituents' number,Value be that 0 or 1,0 expression this external constituents do not pay m, 1 indicates this external ginseng
M is paid with person；
Step 10: repeating above-mentioned steps, using Q-Learning algorithm, optimize parameter p, b, a, c therein.
This method quickly and efficiently, has high accuracy, Gao Youhua intensity and robustness, and the present invention reaches following effect: will
The random distribution characteristic of outside feeding data is dissolved into intelligent contract participant utility function, determines the ginseng of each random distribution
Number, and Q-Learning Optimal Parameters are utilized, it realizes that participant's utility function maximizes, stablizes participant and execute intelligent contract
Motivation.The invention is suitable for validity check and secret protection of the intelligent contract as electronic contract.
Detailed description of the invention
The relationship between the algorithm flow and parameters of intelligent contract is described in detail in Fig. 1.
The number of success of dealer under Fig. 2 difference donations rate
Fig. 3 difference averagely donates the number of success of dealer under intensity
The benefit distribution situation of dealer when Fig. 4 donations rate p=0.1
Specific embodiment:
Step 1: realizing the status change during contract circulation using finite state machine model A, state includes states
= ['s1', 's2', 's3', 's4', 'Sfail', 'Ssucc', 'Sinc'].A is used to emulate the spectators that contract is added.
Step 2: defining conditional jump function, including donations wish function d, donations amount function m and decoding failure function
f；Function d and f is portrayed using the sampling function that the bernoulli that mean value is q=0.1 is distributed, and m uses pareto sampling function
It portrays,, wherein b=1,2 be the Shape parameter in Pareto distribution.
Step 3: defining the cost function v of film, which is distributed exponweib using index W eibull；, wherein parameter a is used to control mean value, and a is bigger, and sampling mean value is bigger；C is used
Sample variance is controlled, c is bigger, and variance is smaller, a=0.49 in test, c=1.9.
Step 4: defining the state space of Q learning algorithm, motion space and benefiting function table, Q table is defined.
Step 5: algorithm scans for state, action schedule using Greedy strategy.
Step 6: defining environment function interaction function step, which arrives state and movement according to the observation, completes a secondary ring
Border interaction.Specifically, step function can check the donations situation of current film, if donations amount deficiency is expected, obtain
Benefit is 0.Otherwise benefit to be donation amount.When movement instruction continues waiting for, then step function operation automatic machine A, the number for running A make
It is distributed with possionIt portrays, intensity λ=200. is reached in test
Step 7: selecting a movement according to strategic function, which is passed into environment function step, and observation state becomes
Change, benefit reward and whether terminate this calculating.
best_next_action = np.argmax(Q[next_state])
td_target = reward + discount_factor * Q[next_state][best_next_action]
td_delta = td_target - Q[state][action]
Q[state][action] += alpha * td_delta
Step 8: using time difference technology innovation Q table.First look for the best movement under current state, and by the state and
Benefit under best movement is cumulative multiplied by the reward in discount factor discount_factor=0.9, with the 7th step, as
Current target benefits td_target.
Step 9: by the state of td_target and Q table, the difference of movement benefit multiplied by learning rate alpha=0.5, as Q table
The new value that state action benefits is updated into Q table.Q table segment are as follows:
(0, 0):( [ 0.005,0. ])
(206, 299)([ 0.005,0.])
(433, 695):( [0.005,0.])
(625, 1): ( 4, 1])
(1016, 574): ([ 0.005,0.],
(1191, 866)([ 0.005,0.])
(1398, 214): ([ 0.005,0.])
…
Step 10: the 5th to the 9th step is repeated, until reaching preset the number of iterations.
Step 11: drawing the relational graph benefited with iteration time, and parameter is modified, observation donations rate, donations amount ginseng
The influence that the state of several couples of A reaches, the influence to intelligent contract promoter income.
Validation verification of the present invention
In order to prove effectiveness of the invention, we have studied dealer under different donations rates, donations intensity to test at 1000 times
In number of success.Figure it is seen that dealer is under the guidance that Q learns immediately, it is intended to which study is to his optimal film
Strategy is launched, but the donations rate raising of participant does not ensure that dealer obtains higher success.Fig. 3 gives different donations
Strength conditions, the acquisition success rate of dealer, Fig. 3 illustrate that influence of the donations intensity to donations intensity to success rate is significant.
In the case that Fig. 4 is donations rate 0.1, the benefit situation of dealer, the mean value that dealer benefits is that 56.9, Fig. 4 is said
Bright dealer in the case where continuous study schedule, benefits also far below expection immediately.
Claims (1)
1. a kind of detection method of the intelligent contract validity based on Q-Learning, it is characterised in that closed by specific intelligence
About building, data feeding, parameter optimization and etc. come what is realized, detailed process is as follows:
Step 1: a file is divided into s parts by intelligent contract promoter, and to every a encryption, if it is desired to restore entire file,
Need to decrypt each one's share of expenses for a joint undertaking file；Spending required for decryption is v, because classified document has a timeliness；I.e. in secrecy text
Before part is not announced, value is higher, but after the decryption, value is decreased obviously；Therefore, it is necessary to use one random point
Cloth spends to indicate that its is decrypted, and cannot be indicated using a constant, and Weibull distribution can be used to indicate decryption cost
(value of film)；
Step 2: external constituents download s parts of encryption subfiles, it is intended to decrypt file；
Step 3: intelligent contract randomly selects a random number of s ' from s, wherein s ' < s；S ' is sent to intelligent conjunction by intelligent contract
About promoter allows him to decrypt these subfiles；
Step 4: if these subfiles of intelligent contract promoter successful decryption, intelligent contract waits external data feeding, continues
It executes；If intelligent contract promoter's successful decryption subfile, it will be able to prove that the subfile that remaining is not decrypted is also correct
's；Otherwise, intelligent termination of contract；
Step 5: each external constituents if it is desired to decrypt entire file, needs part payment amount of money m > 1, but this is golden
Volume is less than the cost for decrypting entire file；Decided whether to pay m according to personal preference, if the probability of payment obeys binomial point
Cloth, wherein k is the total number of persons of donations, and n is experiment number, and p is the probability donated every time；
Step 6: the value of m is also to obey random distribution herein, each participant want to decrypt entire file wish and itself
Financial resources condition is different；Because block chain network has small world, i.e., 80% wealth concentrates in 20% manpower, so, m
Pareto distribution is obeyed, the amount of donation m of this distribution expression most people (such as 80%) is 1 between b；
Step 7: each external constituents, after deciding whether donations, intelligent contract collects these amounts of donation km, as
Data are fed in intelligent contract program；
Step 8: the promoter of intelligent contract decides whether to decrypt entire file according to its income, if km > v, illustrate promoter
As soon as income be a positive value, he has the motivation for decrypting entire file；Gap between km and v is bigger, and the motivation of promoter is got over
Greatly；After entire file decryption, sum participant of all downloading encryption subfiles can obtain decryption file；It is worth noting
Be sum >=k, that is to say, that the external constituents for not paying m can sit idle and enjoy the fruits of others' work, and formation is hitchhiked phenomenon；This point be by
What intelligent contract determined, it, can only be them as one because the promoter of intelligent contract can not identify the identity of external constituents
A entirety is treated；If km≤v, illustrate that intelligent contract promoter is profitless, therefore he does not decrypt the motivation of entire file,
He can select non-decrypting；
Step 9: defining the utility function of intelligent contract promoter and external constituents according to the distribution situation of external data；Intelligence
The utility function of energy contract promoter are as follows: km-v, the utility function of external constituents are as follows:, wherein sum is all
External constituents' number,Value be that 0 or 1,0 expression this external constituents do not pay m, 1 indicates this external ginseng
M is paid with person；
Step 10: repeating above-mentioned steps, using Q-Learning algorithm, optimize parameter p, b, a, c therein.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811515288.0A CN109711176B (en) | 2018-12-12 | 2018-12-12 | A method for detecting the validity of smart contracts based on Q-Learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811515288.0A CN109711176B (en) | 2018-12-12 | 2018-12-12 | A method for detecting the validity of smart contracts based on Q-Learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109711176A true CN109711176A (en) | 2019-05-03 |
| CN109711176B CN109711176B (en) | 2023-08-25 |
Family
ID=66255627
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811515288.0A Active CN109711176B (en) | 2018-12-12 | 2018-12-12 | A method for detecting the validity of smart contracts based on Q-Learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109711176B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112039864A (en) * | 2020-08-25 | 2020-12-04 | 华北电力大学 | Method for analyzing cross-layer security risk of electric power CPS |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018076762A1 (en) * | 2016-10-27 | 2018-05-03 | 上海亿账通区块链科技有限公司 | Block chain-based transaction verification method and system, electronic device, and medium |
| CN108418783A (en) * | 2017-09-01 | 2018-08-17 | 矩阵元技术（深圳）有限公司 | A method and medium for protecting the privacy of blockchain smart contracts |
| GB201813459D0 (en) * | 2018-08-17 | 2018-10-03 | Uvue Ltd | Transaction system and method of operation thereof |
| CN108985936A (en) * | 2018-07-09 | 2018-12-11 | 北京中电普华信息技术有限公司 | A kind of auditing method and system based on intelligent contract |
-
2018
- 2018-12-12 CN CN201811515288.0A patent/CN109711176B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018076762A1 (en) * | 2016-10-27 | 2018-05-03 | 上海亿账通区块链科技有限公司 | Block chain-based transaction verification method and system, electronic device, and medium |
| CN108418783A (en) * | 2017-09-01 | 2018-08-17 | 矩阵元技术（深圳）有限公司 | A method and medium for protecting the privacy of blockchain smart contracts |
| CN108985936A (en) * | 2018-07-09 | 2018-12-11 | 北京中电普华信息技术有限公司 | A kind of auditing method and system based on intelligent contract |
| GB201813459D0 (en) * | 2018-08-17 | 2018-10-03 | Uvue Ltd | Transaction system and method of operation thereof |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112039864A (en) * | 2020-08-25 | 2020-12-04 | 华北电力大学 | Method for analyzing cross-layer security risk of electric power CPS |
| CN112039864B (en) * | 2020-08-25 | 2021-09-14 | 华北电力大学 | Method for analyzing cross-layer security risk of electric power CPS |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109711176B (en) | 2023-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Janssen et al. | Head-enders as stationary bandits in asymmetric commons: Comparing irrigation experiments in the laboratory and the field | |
| Ke et al. | An uncertain random programming model for project scheduling problem | |
| Bull et al. | Reconciling multiple counterfactuals when evaluating biodiversity conservation impact in social‐ecological systems | |
| Lee et al. | Evidence or confidence: What is really monitored during a decision? | |
| Shadowen et al. | Participatory governance in smart cities: Future scenarios and opportunities | |
| Jaworski et al. | Go West Young Man: Self‐Selection and Endogenous Property Rights | |
| Shattock et al. | In the interests of time: improving HIV allocative efficiency modelling via optimal time‐varying allocations | |
| Buscemi et al. | Do LLMs trust AI regulation? Emerging behaviour of game-theoretic LLM agents | |
| CN109711176A (en) | A Q-Learning-based smart contract validity detection method | |
| Mahalingam et al. | Predicting financial savings decisions using sigmoid function and information gain ratio | |
| Liu et al. | Comparing proactive and reactive management: managing a transboundary fish stock under changing environment | |
| Dennis et al. | Options and Motivations for International AI Benefit Sharing | |
| Webb et al. | Waiting when both certainty and magnitude are increasing: Certainty overshadows magnitude | |
| Irfan et al. | Causal strategic inference in a game—theoretic model of multiplayer networked microfinance markets | |
| Ahn et al. | Caution as a Response to Scientific Uncertainty | |
| Kyriakou | Technology and sustainable growth: Towards a synthesis | |
| Garcia | Homo Silicus is Hyper-Rational: Why LLM Agents Fail to Replicate Attention-Driven Trading | |
| Gaie et al. | Public services data analytics using artificial intelligence solutions derived from telecommunications systems | |
| Wang et al. | Group‐guided artificial bee colony algorithm with elastic adjustment strategy | |
| Zhang | Data-Driven Agent-based Modeling of Innovation Diffusion (Doctoral Consortium) | |
| Mitra et al. | Enhancing coordination in water management through communication tools: Results from experimental games in Coastal Bangladesh | |
| Daimari et al. | Firefly based Unit Commitment | |
| Katsikas | Game theoretic models of networks security | |
| Garibaldi et al. | Enhancing Bold Collective Action: A Variable Geometry and Incentives Regime | |
| Karova et al. | A genetic algorithm basic approach for software management project |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230920 Address after: Room 903-904, Building 3, No. 17 Huaguoshan Avenue, High tech Zone, Lianyungang City, Jiangsu Province, 222000 Patentee after: Lianyungang Micro-Painting Hall Network Technology Co.,Ltd. Address before: 273165 Jingxuan West Road, Qufu City, Jining, Shandong Province, No. 57 Patentee before: QUFU NORMAL University |