CN108111535A - A kind of optimal attack path planing method based on improved Monte carlo algorithm - Google Patents
A kind of optimal attack path planing method based on improved Monte carlo algorithm Download PDFInfo
- Publication number
- CN108111535A CN108111535A CN201810031551.2A CN201810031551A CN108111535A CN 108111535 A CN108111535 A CN 108111535A CN 201810031551 A CN201810031551 A CN 201810031551A CN 108111535 A CN108111535 A CN 108111535A
- Authority
- CN
- China
- Prior art keywords
- path
- attack
- attack path
- host
- optimal
- 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.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/14—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
- H04L63/1433—Vulnerability analysis
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/20—Network architectures or network communication protocols for network security for managing network security; network security policies in general
Abstract
The present invention relates to a kind of optimal attack path planing methods based on improved Monte carlo algorithm, belong to field of information security technology.Concrete operation step is：Step 1: obtain network system situation information.Step 2: by improved Monte carlo algorithm, optimal attack path is obtained.It is proposed by the present invention based on improving the optimal attack paths planning method of Monte carlo algorithm compared with the prior art compared with haing the following advantages：1. the assessed value calculating based on weight vector avoids path and loses problem.2. algorithm space complexity reduces.
Description
Technical Field
The invention relates to an optimal attack path planning method based on an improved Monte Carlo algorithm, and belongs to the technical field of information security.
Background
When a network attack is carried out, an attacker always wants to attack a target host with the minimum attack cost so as to obtain data wanted by the attacker. The minimization of the attack cost mainly depends on an attack path, and an optimal attack path is often the key for attack success. The optimal attack path refers to the path with the least attack cost and the most returns. At present, the method for acquiring the optimal attack path mainly acquires all paths between a source node and a destination node through an attack graph and selects a path meeting the conditions as the optimal attack path.
Currently, the most popular network attack graphs are the attribute attack graph and the state attack graph, respectively. The problem with attack graphs based on the above two types is that: (1) the generation speed of the attack path is low; (2) the approach of defining a path to deal with the problem of state explosion leads to the absence of an attack path and the like.
In order to solve the above problems of the attack graph and reduce the influence of the problems on obtaining the optimal attack path, the hidden markov model and the attack graph are combined by the chinese scientist and the japanese scientist, and the optimal attack path is calculated by a probabilistic method by adopting an ant colony optimization algorithm. However, when facing a large-scale computer cluster, the method cannot rapidly calculate the optimal attack path due to the overhead problem.
The optimal attack path planning method based on Q learning proposed in the national invention patent (patent application number: 201710556319.6) mainly solves the following problems: (1) the proposed network model does not need to be trained, so that training data do not need to be collected; (2) the method can be used for online learning and determining the optimal attack paths corresponding to different network states at different moments in real time; (3) the learning rate uses an annealing model, so that the convergence is more accurate; (4) the method is suitable for large-scale computer clusters because the attack graph does not need to be generated. But the disadvantages are: (1) the algorithm space complexity is higher, so the occupied memory space is more. (2) There is a problem of path loss.
The existing algorithm employed in the present invention is the monte carlo algorithm.
The monte carlo algorithm is a process of gradually building an asymmetric search tree by randomly deriving a game. The method can be roughly divided into four steps of selection, expansion, simulation and back propagation. The Monte Carlo algorithm is applied to the optimal attack path planning problem, the optimal attack path can be planned better by combining the vulnerability risk degree, and meanwhile, the space complexity of the algorithm is reduced.
Disclosure of Invention
The invention aims to calculate the optimal attack path based on the Monte Carlo algorithm.
The purpose of the invention is realized by the following technical scheme.
The invention discloses an optimal attack path planning method based on an improved Monte Carlo algorithm, which comprises the following specific operation steps:
step one, acquiring state information of a network system.
Step 1.1: acquiring software applications of all hosts in a network system, and establishing a corresponding table of the software applications and the hosts.
The software application and host correspondence table comprises: a software application name and a host name.
Step 1.2: obtaining the session link between the hosts in the network system, and establishing a session link table between the hosts. The inter-host session link table includes: a source hostname and a target hostname.
Step 1.3: and acquiring the bugs existing in each host in the network system, and establishing a host bug state table. The host vulnerability state table comprises: host name, vulnerability ID, vulnerability type and vulnerability risk degree. The Vulnerability risk is obtained from Common Vulnerability Scoring System.
And step two, acquiring an optimal attack path through an improved Monte Carlo algorithm.
And on the basis of the operation of the step one, acquiring an optimal attack path through an improved Monte Carlo algorithm. The method comprises the following specific steps:
step 2.1: the optimal attack path sequence is denoted by the symbol L, and its initial value is null.
Step 2.2: initial values for the modified monte carlo model are set. The method specifically comprises the following steps:
step 2.2.1: two hosts are randomly selected in a network system, wherein one host is used as a source host, and the other host is used as a target host. All paths between the source host and the target host are taken as attack paths.
Step 2.2.2: a weight vector is set for all hosts involved in the attack path described in step 2.2.1. The weight vector includes a priority, a degree, and a vulnerability risk degree for each host. Wherein the priority is randomly assigned as a number between 1 and 10; the degree is the number of connections between each host and other hosts; and acquiring the vulnerability risk degree of each host according to a vulnerability scoring system.
Step 2.3: and calculating the evaluation value and the total vulnerability risk degree of the attack path through a formula (1) and a formula (2) by using the weight vector of each node in the attack path through a depth-first strategy, and calculating the ratio of the vulnerability risk degree of the attack path to the number of hosts by using a formula (3). Wherein the weight vector comprises a priority, a degree and a vulnerability risk degree of the host.
Wherein, X h An evaluation value representing the attack path; p is a radical of i Is the priority in each host weight vector in the attack path, which is a pre-random set value, p i ∈(1,10)，i∈[1,n]N represents the total number of nodes in the attack path; r is i Is the vulnerability risk degree in each host weight vector in the attack path; d is a radical of i Is the degree in each host weight vector in the attack path.
Wherein r is sum Is the vulnerability risk of the attack path.
Wherein r is rate Is the ratio of the vulnerability risk of the attack path to the total number of nodes in the attack path.
Step 2.4: and (5) electing. A threshold value alpha, alpha epsilon (1, 10) is set. For any one of the attack paths, if X h &Alpha, rejecting the attack path; otherwise, the attack path is reserved and put into the optimal attack path sequence L, so that the optimal attack path sequence L is obtained.
Step 2.5: and is propagated in the reverse direction. The specific operation is as follows:
step 2.5.1: for attack paths in the optimal attack path sequence L, according to X of each attack path h Sorting from big to small, then using symbols in turnAn evaluation value representing the sorted jth path; j is an element of [1, m ]]M is the number of paths, and m is a positive integer; then there isBy symbolsAnd representing the ratio of the vulnerability risk degree of the j-th ordered path to the total number of the nodes in the j-th ordered path.
Step 2.5.2: and comparing two attack paths which are ordered and adjacent in pairs in sequence, and processing the attack paths by the following 3 conditions:
case 1: if two adjacent attack paths satisfyAnd is provided withThen the evaluation value is given by equation (4)Priority p in the path of (1) i Modified to evaluate the value asPriority p in a path i A modification is made.
Case 2: if two adjacent attack paths satisfyAnd isThen the evaluation value is given by equation (6)Priority p in the path of (1) i Modified to evaluate the value asPriority p in a path i A modification is made.
Case 3: if two adjacent attack paths satisfyAnd isThen the evaluation value is given by equation (4)Priority p in the path of (1) i Modified to evaluate the value asPriority p in a path i A modification is made.
p sub ＝p large -δ (4)
Wherein p is sub Is evaluated asThe modified priority of the path of (1); p is a radical of large Is evaluated asPriority before path modification; delta is the maximum vulnerability risk degree in all paths in the optimal attack path sequence LDifference from minimum vulnerability risk; δ is calculated by equation (5).
δ＝ε(max{r sum }-min{r sum }) (5)
Wherein, epsilon is an adjusting coefficient, the initial value of epsilon is a preset random set value, epsilon belongs to (0, 1), and epsilon is used for normalization; max { r } sum Is the vulnerability risk degree r of all paths in the optimal attack path sequence L sum Maximum value of (1); min { r sum Is the vulnerability risk degree r of all paths in the optimal attack path sequence L sum Minimum value of (1).
p sum ＝p little +δ(6)
Wherein p is sum Is evaluated asThe modified priority of the path of (1); p is a radical of formula little Is evaluated asThe priority before modification of the path.
Step 2.6: and calculating the evaluation value of the attack path by using the weight vector of each node in the attack path through a depth-first strategy and by using the formula (1) again.
Wherein X h An evaluation value representing the attack path; p is a radical of i The priority in each host weight vector in the attack path is a pre-random set value, p i ∈(1,10)，i∈[1,n]N represents the total number of nodes in the attack path; r is i Is the vulnerability risk degree in each host weight vector in the attack path; d i Is the degree in each host weight vector in the attack path.
Step 2.7: and (4) electing. For any one of the attack paths, if the current X is h &Alpha, then the attack is rejectedA path; otherwise, the attack path is reserved and put into the optimal attack path sequence L, so that the optimal attack path sequence L is obtained.
The attack path in the optimal attack path sequence L is the optimal attack path from the source host to the target host.
Advantageous effects
Compared with the prior art, the optimal attack path planning method based on the improved Monte Carlo algorithm has the following advantages:
(1) the evaluation value calculation based on the weight vector avoids the path loss problem.
(2) The algorithm space complexity is reduced;
drawings
Fig. 1 is an operation flowchart of an optimal attack path planning method based on an improved monte carlo algorithm according to an embodiment of the present invention;
fig. 2 is a network topology diagram in an embodiment of the present invention.
Detailed Description
According to the technical scheme, the invention is described in detail by combining the drawings and the implementation examples.
The optimal attack path planning method based on the improved Monte Carlo algorithm provided by the invention is used for searching the optimal attack path in the network system, the operation flow is shown as figure 1, and the specific operation steps are as follows:
step one, acquiring network system state information.
Step 1.1: the network topology is shown in fig. 2. Acquiring software applications of all hosts in a network system, and establishing a corresponding table of the software applications and the hosts, as shown in table 1.
TABLE 1 software application to host mapping table
| Software name | Host name |
| IIS7.0 | H 2 ，H 3 |
| BIND 9 | H 1 ，H 2 ，H 5 |
| Sendmail 8.13 | H 3 ，H 4 ，H 5 ，H 7 |
| MySQL 5.7 | H 1 ，H 3 ，H 5 ，H 6 ，H 7 |
| Serv-U 10.5 | H 3 ，H 4 ，H 6 ，H 7 |
| IE6.0 | H 3 ，H 4 |
Step 1.2: session links between hosts in a network system are obtained, and an inter-host session link table is established, as shown in table 2.
Table 2 inter-host session linking table
In table 2, 1 indicates direct communication between the two hosts, and 0 indicates no direct communication between the two hosts.
Step 1.3: the vulnerabilities existing in each host in the network system are obtained, and a host vulnerability state table is established, as shown in table 3. The host vulnerability status table includes: host name, vulnerability ID, vulnerability type and vulnerability risk degree. The Vulnerability risk is obtained from Common Vulnerability Scoring System.
The vulnerability ID is the CVE (Common Vulnerabilities & Exposures) ID.
TABLE 3 host vulnerability State Table
And step two, acquiring an optimal attack path through an improved Monte Carlo algorithm.
And on the basis of the operation of the step one, acquiring an optimal attack path through an improved Monte Carlo algorithm. The method comprises the following specific steps:
step 2.1: the optimal attack path sequence is denoted by the symbol L, and its initial value is null.
Step 2.2: initial values for the modified monte carlo model are set. The method comprises the following specific steps:
step 2.2.1: randomly selecting H in network system 1 As source host, H 7 As a target host. Source host H 1 And target host H 7 All paths in between act as attack paths.
Step 2.2.2: for all hosts involved in the attack path described in step 2.2.1, a weight vector is set, as shown in table 4.
Table 4 initial host weight vector table
| H 1 | H 2 | H 3 | H 4 | H 5 | H 6 | H 7 | |
| Priority (p) i ) | 6 | 5 | 3 | 1 | 2 | 3 | 4 |
| Degree (d) i ) | 3 | 2 | 5 | 3 | 3 | 3 | 3 |
| Vulnerability risk degree (r) i ) | 18.7 | 13.7 | 13.9 | 15.0 | 17.5 | 14.3 | 19.4 |
Step 2.3: and calculating the evaluation value and the total vulnerability risk degree of the attack path through a formula (1) and a formula (2) by using the weight vector of each node in the attack path through a depth-first strategy, and calculating the ratio of the vulnerability risk degree of the attack path to the number of hosts by using a formula (3). Wherein the weight vector comprises a priority, a degree and a vulnerability risk degree of the host.
Wherein, X h An evaluation value representing the attack path; p is a radical of i Is the priority in each host weight vector in the attack path, i belongs to [1, n ]]N represents the total number of nodes in the attack path; r is i Is the vulnerability risk degree in each host weight vector in the attack path; d i Is the degree in each host weight vector in the attack path.
Wherein r is sum Is the vulnerability risk of the attack path.
Wherein r is rate Is the ratio of the vulnerability risk of the attack path to the total number of nodes in the attack path.
Step 2.4: and (5) electing. A threshold value alpha is set. For any one of the attack paths, if X h &Alpha, rejecting the attack path; otherwise, the attack path is reserved and put into the optimal attack path sequence L, so that the optimal attack path sequence L is obtained. In the present embodiment, the threshold α =4.5. At this time, there are 2 paths in the optimal attack path sequence L, which are: h 1 →H 3 →H 7 And H 1 →H 3 →H 6 →H 7 。
Step 2.5: and is propagated in the reverse direction. The specific operation is as follows:
step 2.5.1: for attack paths in the optimal attack path sequence L, according to X of each attack path h Sorting from large to small and then using symbols in sequenceAn evaluation value representing the sorted jth path; j is an element of [1, m ]]M =2; then there isBy means of symbolsAnd representing the ratio of the vulnerability risk degree of the j-th ordered path to the total number of the nodes in the j-th ordered path. The result after sorting is: (H) 1 →H 3 →H 6 →H 7 ，H 1 →H 3 →H 7 )。
Step 2.5.2: comparing two adjacent attack paths in sequence, because the two adjacent attack paths meet the situationIn case 1:and isTherefore, the evaluation value is given by equation (4)Priority p in the path of (1) i Modified to evaluate the value asPriority p in a path i A modification is made.
p sub ＝p large -δ (4)
Wherein p is sub Is evaluated asThe modified priority of the path of (1); p is a radical of large Is evaluated asPriority before path modification; delta is the difference between the maximum vulnerability risk degree and the minimum vulnerability risk degree in all the paths in the optimal attack path sequence L; δ is calculated by equation (5).
δ＝ε(max{r sum }-min{r sum }) (5)
Wherein, epsilon is an adjusting coefficient, the initial value of epsilon is a preset random set value, epsilon belongs to (0, 1), and epsilon is used for normalization; max { r } sum Is the vulnerability risk degree r of all paths in the optimal attack path sequence L sum Maximum value of (2); min { r sum Is the vulnerability risk degree r of all paths in the optimal attack path sequence L sum Minimum value of (1).
p sum ＝p little +δ (6)
Wherein p is sum Is evaluated asThe modified priority of the path of (1); p is a radical of formula little Is evaluated asPriority before modification of the path. In this embodiment, path H 1 →H 3 →H 6 →H 7 R of rate The value of the sum of the values is 16.5,a value of 5.0187; route H 1 →H 3 →H 7 R of rate The value of the sum of the measured values is 16.7,the value was 4.7427. When epsilon =0.1, delta =0.01 (97.5-52.0) =0.455.
Through the above steps, a modified host right vector table is obtained, as shown in table 5.
Table 5 modified host rights vector table
Step 2.6: and calculating the evaluation value of the attack path by using the weight vector of each node in the attack path through a depth-first strategy and by using the formula (1) again.
Wherein X h An evaluation value representing the attack path; p is a radical of i Is the priority in each host weight vector in the attack path, i ∈ [1, n [ ]]N representsThe total number of nodes in the attack path; r is a radical of hydrogen i Is the vulnerability risk degree in each host weight vector in the attack path; d is a radical of i Is the degree in each host weight vector in the attack path.
Step 2.7: and (5) electing. For any one of the attack paths, if the current X is h &If alpha, alpha =4.5, then eliminating the attack path; otherwise, the attack path is reserved and put into the optimal attack path sequence L, so that the optimal attack path sequence L is obtained. Through the operation of the step, the optimal attack path sequence L is obtained.
Only one attack path H in the optimal attack path sequence L 1 →H 3 →H 7 ，H 1 →H 3 →H 7 I.e., the optimal attack path from the source host to the target host, the evaluation value is 4.6597.
In order to illustrate the effectiveness of the method, under the same network environment, the optimal attack path obtained by using the optimal attack path planning method based on Q learning proposed by the patent (patent application number: 201710556319.6) is H 1 →H 3 →H 7 . The same results were obtained in both experiments, demonstrating the effectiveness of the method of the invention. In addition, the algorithm space complexity of the method is O (N) by comparison 2 ) The spatial complexity of the algorithm of patent 201710556319.6 is O (N) 3 ). Therefore, the method has higher calculation speed.
Claims (1)
1. An optimal attack path planning method based on an improved Monte Carlo algorithm is characterized in that: the specific operation steps are as follows:
step one, acquiring state information of a network system;
step 1.1: acquiring software applications of all hosts in a network system, and establishing a correspondence table between the software applications and the hosts;
the software application and host correspondence table comprises: a software application name and a host name;
step 1.2: acquiring session links among hosts in a network system, and establishing a session link table among the hosts; the inter-host session link table includes: a source hostname and a target hostname;
step 1.3: acquiring vulnerabilities existing in each host in a network system, and establishing a host vulnerability state table; the host vulnerability status table includes: host name, vulnerability ID, vulnerability type and vulnerability risk degree; the vulnerability risk degree is obtained from a general vulnerability scoring system;
step two, obtaining an optimal attack path through an improved Monte Carlo algorithm;
on the basis of the operation of the first step, obtaining an optimal attack path through an improved Monte Carlo algorithm; the method comprises the following specific steps:
step 2.1: the optimal attack path sequence is represented by a symbol L, and the initial value of the optimal attack path sequence is null;
step 2.2: setting an initial value of the improved Monte Carlo model; the method specifically comprises the following steps:
step 2.2.1: randomly selecting two hosts in a network system, wherein one host is used as a source host, and the other host is used as a target host; taking all paths between a source host and a target host as attack paths;
step 2.2.2: setting weight vectors for all hosts involved in the attack path in the step 2.2.1; the weight vector comprises the priority, the degree and the vulnerability risk degree of each host; wherein the priority is randomly assigned as a number between 1 and 10; the degree is the number of connections between each host and other hosts; acquiring the vulnerability risk degree of each host according to a vulnerability scoring system;
step 2.3: calculating an evaluation value and a total vulnerability risk degree of the attack path through a formula (1) and a formula (2) by using a weight vector of each node in the attack path through a depth-first strategy, and calculating a ratio of the vulnerability risk degree of the attack path to the number of hosts by using a formula (3); wherein the weight vector comprises the priority, the degree and the vulnerability risk degree of the host;
wherein, X h An evaluation value representing the attack path; p is a radical of i The priority in each host weight vector in the attack path is a pre-random set value, p i ∈(1,10)，i∈[1,n]N represents the total number of nodes in the attack path; r is i Is the vulnerability risk degree in each host weight vector in the attack path; d i Is the degree in each host weight vector in the attack path;
wherein r is sum Is the vulnerability risk of the attack path;
wherein r is rate The ratio of the vulnerability risk degree of the attack path to the total number of the nodes in the attack path;
step 2.4: electing; setting a threshold value alpha, alpha epsilon (1, 10); for any one of the attack paths, if X h &Alpha, rejecting the attack path; otherwise, the attack path is reserved and put into the optimal attack path sequence L, so that the optimal attack path sequence L is obtained;
step 2.5: backward propagation; the specific operation is as follows:
step 2.5.1: for attack paths in the optimal attack path sequence L, according to X of each attack path h Sorting from big to small, then using symbols in turnAn evaluation value representing the j-th path after sorting; j belongs to [1,m ]]M is the number of paths, and m is a positive integer; then there isBy symbolsRepresenting the ratio of the vulnerability risk degree of the j-th ordered path to the total number of the nodes in the j-th ordered path;
step 2.5.2: and comparing two attack paths which are ordered and adjacent in pairs in sequence, and processing the attack paths by the following 3 conditions:
case 1: if two adjacent attack paths satisfyAnd isThen the evaluation value is given by equation (4)Priority p in the path of (1) i Modified to evaluate the value asPriority p in a path i Modifying;
case 2: if two adjacent attack paths satisfyAnd isThen the evaluation value is given by equation (6)Priority p in the path of i Modified to evaluate the value asPriority p in a path i Modifying;
situation(s)3: if two adjacent attack paths satisfyAnd isThen the evaluation value is given by equation (4)Priority p in the path of (1) i Modified to evaluate the value asPriority p in a path i Modifying;
p sub ＝p large -δ (4)
wherein p is sub Is evaluated asThe modified priority of the path of (1); p is a radical of large Is evaluated asPriority before path modification; delta is the difference between the maximum vulnerability risk degree and the minimum vulnerability risk degree in all the paths in the optimal attack path sequence L; delta is calculated by formula (5);
δ＝ε(max{r sum }-min{r sum }) (5)
wherein epsilon is an adjusting coefficient, the initial value of epsilon is a preset random set value, epsilon belongs to (0, 1), and epsilon is used for normalization; max { r } sum Is the vulnerability risk degree r of all paths in the optimal attack path sequence L sum Maximum value of (2); min { r sum Is the vulnerability risk degree r of all paths in the optimal attack path sequence L sum Minimum value of (d);
p sum ＝p little +δ (6)
wherein p is sum Is evaluated asThe modified priority of the path of (1); p is a radical of little Is evaluated asPriority before path modification;
step 2.6: calculating the evaluation value of the attack path by using the weight vector of each node in the attack path and the formula (1) again through a depth-first strategy;
step 2.7: electing; for any one of the attack paths, if the current X is h &Alpha, rejecting the attack path; otherwise, the attack path is reserved and put into the optimal attack path sequence L, so as to obtain the optimal attack path sequence L;
and the attack path in the optimal attack path sequence L is the optimal attack path from the source host to the target host.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810031551.2A CN108111535A (en) | 2018-01-12 | 2018-01-12 | A kind of optimal attack path planing method based on improved Monte carlo algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810031551.2A CN108111535A (en) | 2018-01-12 | 2018-01-12 | A kind of optimal attack path planing method based on improved Monte carlo algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108111535A true CN108111535A (en) | 2018-06-01 |
Family
ID=62220012
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810031551.2A Pending CN108111535A (en) | 2018-01-12 | 2018-01-12 | A kind of optimal attack path planing method based on improved Monte carlo algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108111535A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109120646A (en) * | 2018-07-18 | 2019-01-01 | 北京理工大学 | Network optimum defense system construction method based on Monte Carlo graph search algorithm |
| CN112380532A (en) * | 2020-11-13 | 2021-02-19 | 深信服科技股份有限公司 | Host risk state determination method and device, electronic equipment and storage medium |
-
2018
- 2018-01-12 CN CN201810031551.2A patent/CN108111535A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109120646A (en) * | 2018-07-18 | 2019-01-01 | 北京理工大学 | Network optimum defense system construction method based on Monte Carlo graph search algorithm |
| CN112380532A (en) * | 2020-11-13 | 2021-02-19 | 深信服科技股份有限公司 | Host risk state determination method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113011602B (en) | Federal model training method and device, electronic equipment and storage medium | |
| CN109116841B (en) | Path planning smooth optimization method based on ant colony algorithm | |
| Ziavras | RH: a versatile family of reduced hypercube interconnection networks | |
| CN109064348B (en) | Method for locking rumor community and inhibiting rumor propagation in social network | |
| CN113422695B (en) | Optimization method for improving robustness of topological structure of Internet of things | |
| CN112165405B (en) | Method for testing big data processing capacity of supercomputer based on network topological structure | |
| CN108111535A (en) | A kind of optimal attack path planing method based on improved Monte carlo algorithm | |
| CN110061870A (en) | Efficiency estimation method is combined with side based on node in a kind of Tactical Internet | |
| CN110445654A (en) | A kind of social networks multi-source rumour source tracing method and system based on community's division | |
| Qiu et al. | Born this way: A self-organizing evolution scheme with motif for internet of things robustness | |
| CN110855654B (en) | Vulnerability risk quantitative management method and system based on flow mutual access relation | |
| CN109587080A (en) | A kind of network-on-chip fast mapping algorithm based on Topology partition | |
| Huang et al. | An improved biogeography-based optimization algorithm for flow shop scheduling problem | |
| Tang et al. | Fundamental matrix estimation by multiobjective genetic algorithm with Taguchi's method | |
| CN109120646B (en) | Network optimal defense system construction method based on Monte Carlo graph search algorithm | |
| Ning et al. | A cloud-supported cps approach to control decision of process manufacturing: 3D ONoC | |
| McClure | Toward a better understanding of species interactions through network biology | |
| Zhang et al. | On convergence rate for multi-agent consensus: a community detection algorithm | |
| WO2021012220A1 (en) | Evasion attack method and device for integrated tree classifier | |
| CN110351241A (en) | A kind of industrial network DDoS intruding detection system classification method based on GWA optimization | |
| CN115913749B (en) | Block chain DDoS detection method based on decentralization federation learning | |
| CN110910952A (en) | Method for predicting basic protein by using chemical reaction strategy | |
| Lin et al. | A double learning models-based multi-objective estimation of distribution algorithm | |
| Li et al. | Virtual network embedding based on multi-objective group search optimizer | |
| CN114386769B (en) | Power output determining method and device based on privacy protection in smart grid |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180601 |