CN106708923B - A kind of local cache sharing files method based on mobile collective intelligence network - Google Patents
A kind of local cache sharing files method based on mobile collective intelligence network Download PDFInfo
- Publication number
- CN106708923B CN106708923B CN201610985297.0A CN201610985297A CN106708923B CN 106708923 B CN106708923 B CN 106708923B CN 201610985297 A CN201610985297 A CN 201610985297A CN 106708923 B CN106708923 B CN 106708923B
- Authority
- CN
- China
- Prior art keywords
- user
- file
- alliance
- pattern
- cache
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
- G06F16/17—Details of further file system functions
- G06F16/176—Support for shared access to files; File sharing support
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
- G06F16/17—Details of further file system functions
- G06F16/172—Caching, prefetching or hoarding of files
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/06—Protocols specially adapted for file transfer, e.g. file transfer protocol [FTP]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Information Transfer Between Computers (AREA)
Abstract
The invention discloses a kind of local cache sharing files methods based on mobile collective intelligence network, it is primarily based on user coalitions pattern of the Cooperative reference building based on gunz sharing files network, the user coalitions pattern is the set that multiple user files that user is formed share alliance;According to user coalitions pattern is used, best file cache strategy and sharing files strategy are formulated for specific user alliance;According to file cache strategy and sharing files strategy, user is from the file of base station downloading corresponding proportion to local mobile device.Present invention is primarily based on collective intelligence networks to form the user coalitions that file cache is shared, and reduces transmission delay by locality connection;According to Cooperative reference, the distributed algorithm of the user coalitions pattern with individual rationality, Dhp stability and Contract Stability is obtained;Introduce the concept of file popularity, it is determined that the best cache policy for minimizing the total transmission delay of user coalitions.The present invention can reduce mobile network's transmission delay.
Description
Technical field
The invention belongs to mobile internet technical field more particularly to a kind of local cache texts based on mobile collective intelligence network
Part sharing method.
Background technique
Watching video by mobile device has become to become more and more popular.According to Cisco System Co. to global mobile data flow
Prediction, mobile video flow just exponentially increases, and expection in 2017 is up to 9.9 Chinese mugwort bytes/moon.End to 2020
Year, video flow will account for the 72% of total flow.It works independently by core network and would become hard to meet so huge demand, and
Somewhat expensive, while the transmission delay for shortening mobile video service is also imperative.Utilize the memory space of magnanimity edge device
Technological progress with short range communication can provide a solution.International mobile equipment and access point expection can reach 9,000,000,
The huge and inexpensive memory space using them, then the high-speed transfer by wireless terminal market realization file.Then, user
Cache file, such as the p- equipment of equipment-(Device-to-Device, D2D) communication, WiFi, indigo plant can be shared by short range communication
Tooth and ultra wide band.Two short range communication transmission rates between close user can be than user and core network base station
The fast several times of cellular transmission rate between (Base Stations, BSs) substantially shorten the transmission delay of user.
Recently, many concerns are had been obtained by the conception of home base stations or mobile terminal device cache file.With
Femto caches the proposition of concept, and welcome file has been buffered in small base station (Femtocells), downloads for surrounding user.
The scalability of cache file in user equipment is also all verified.However, user can only cache him due to the person's character of people's selfishness
Favorite file, it is not interested for other people cache files.Therefore it may result in repetition caching, to cause to store
The waste in space.There is researcher to recommend core network carrier with bonus policy, encourages user cache file and pass through D2D
Network is shared.Even so, but core network carrier is difficult to determine and quantify the money return of user cache behavior.Such as
Fruit core network carrier without centralized control, remove active cache and divide mutually by the not enough motivations of selfish user oneself
Enjoy file.
In order to solve this problem, distributed structure/architecture can be used, according to cooperative game theory, user's composition is based on gunz
The alliance of sharing files network, alliance internal members carry out sharing files, do not carry out sharing files with the member other than alliance.But
There are still challenges in caching system, the locality connection rate phase not to the utmost between the file preference of user, file size and user
Together, it therefore is difficult to entire alliance and determines best file cache strategy.Selfish user can only cache the text of most desired caching
Part, and alliance responsible person needs to consider the interests of this alliance, makes full use of the limited memory space of all user equipmenies.User
The formation of alliance has spontaneity, and user freely can be added or leave alliance.If coalizing causes service delay to be more than not add
Fashionable service delay, then selfish user will move out alliance.If other alliances, which are added, shortens current delay, user's meeting
It is detached from existing alliance.Determine that a kind of stable alliance sets up form and needs a model being easily achieved for user group.
Summary of the invention
In view of the deficiencies of the prior art, the present invention proposes a kind of novel local cache files based on collective intelligence network point
Enjoy method.
The technical scheme adopted by the invention is that: a kind of local cache sharing files method based on mobile collective intelligence network,
Characterized by comprising the following steps:
Step 1: the user coalitions pattern based on gunz sharing files network, the user are constructed based on Cooperative reference
Alliance's pattern is the set that multiple user files that user is formed share alliance;
Step 2: according to the user coalitions pattern formed using step 1, formulating best file cache for specific user alliance
Strategy and sharing files strategy;
Step 3: the file cache strategy and sharing files strategy obtained according to step 2, user download corresponding ratio from base station
Example file to local mobile device.
Preferably, the specific implementation of step 1 includes following sub-step:
Step 1.1: defining a kind of for measuring the superiority and inferiority relationship of alliance pattern S and alliance pattern R
Alliance pattern S is better than alliance's pattern R, all little and if only if service delay of all users at alliance pattern S
In the service delay at alliance pattern R, and service delay of at least one user at alliance pattern S strictly less than
Service delay under alliance pattern R, i.e., And it at least strictly keeps a kind of and is less than pass
System, wherein ci(S) service delay of the user i at alliance pattern S is indicated;
Step 1.2: defining the principle that alliance's pattern updates: combination principle, separation principle and addition principle;
Combination principle: if there is multiple alliances, alliance's pattern after merging is more excellent, i.e., Then merge the alliance { S of all dispersions1, S2..., SlBe
Separation principle: if there is an alliance, multiple alliances can be split as, and alliance's pattern after fractionation is more
It is excellent, i.e.,Then by allianceIt is decomposed into { S1, S2..., Sl};
Principle is added: if an alliance S1Middle there are a Ge Zi alliance S ', are detached from alliance S1Coalize S2Alliance afterwards
Pattern is more excellent, i.e.,Then by alliance S1Subset S ' coalize S2;
Step 1.3: initialization alliance's pattern is the alliance S0 comprising all isolated users;
Step 1.4: judging whether there is the alliance for meeting combination principle;
If it is present updating alliance's pattern according to combination principle, obtaining new alliance's pattern is S;And it connects circulation and holds
Row step 1.4;
If it does not exist, then executing following step 1.5;
Step 1.5: judging whether there is the alliance for meeting separation principle;
If it is present updating alliance's pattern according to separation principle, obtaining new alliance's pattern is S;And it connects circulation and holds
Row step 1.5;
If it does not exist, then executing following step 1.6;
Step 1.6: judging whether there is the alliance for meeting and principle being added;
If it is present updating alliance's pattern according to principle is added, obtaining new alliance's pattern is S;And it connects circulation and holds
Row step 1.6;
If it does not exist, then revolution executes the step 1.7.
Step 1.7: judging whether S0 is equal to S;
If it is not, revolution executes step 1.4;
If so, alliance's pattern that output is stable, this process terminate.
Preferably, the specific implementation of file cache strategy described in step 2 includes following sub-step:
Step 2.1: establishing user-file to (i, m) list
Step 2.2: being based on user preference file distribution, user's locality connection rate, user and base station transmission rate, calculate
Popularity of each user-file to (i, m)
Wherein RI, 0For the transmission rate between user i and base station, RJ, iLocal transmission speed between user i and user j
Rate,It is user i to the preference of file m, smFor the size of file m;
Step 2.3: from listIt is middle to remove user-file pair that all popularities are negative value;
Step 2.4: judging listIt whether is empty;
If so, this process terminates;
If it is not, then continuing to execute step 2.5;
Step 2.5: calculatingIn all user-files pair pre-cache ratio
Wherein γiFor the mobile device available space of user i, ymFor the ratio of the uncached part of file m;
Step 2.6: choosing listMiddle popularity and pre-cache ratio productMaximum user-file pair,
And from listMiddle removal this document pair;
Step 2.7: the ratio of designated user's i cache file m is
Step 2.8: the mobile device available space for updating user i is
Step 2.9: whether the mobile device available space for judging user i is 0;
If so, from listIt is middle to remove all user-files pair comprising user i;And execute following step 2.10;
If it is not, then executing following step 2.10;
Step 2.10: the ratio for updating the uncached part of file m is
Step 2.11: judging whether the ratio of the uncached part of file m becomes 0;
If so, from listIt is middle to remove all user-files pair comprising file m;And it turns round and executes the step
2.4;
The step 2.4 is executed if it is not, then turning round.
Preferably, the specific implementation process of sharing files strategy described in step 2 is: when user i needs file m,
First look in local mobile device whether the full content of storage file m;If it is not, being obtained by locality connection same
The ratio of the file m cached in user's j mobile device in alliance, the file m that user j is shared are its file m's cached
All parts;If the file m for not having caching whole in user's mobile device in same alliance, user i publish papers under base station
The remainder of part m.
Preferably, each alliance specifies an alliance responsible person, is counted by alliance responsible person after forming user coalitions pattern
Calculation and the best file cache strategy of implementation and supervision and sharing files strategy.
Preferably, if in user coalitions all users it is having the same to base station communication rate, locality connection rate,
File preference and mobile device memory space, then the file cache strategy of each user is identical as sharing files strategy;
The realization of file cache strategy A includes following sub-step:
Step A1: the popularity q of All Files m is calculatedm;
Wherein n is the number of users in alliance, smFor the size of file m, R has identical to base station for all users
Traffic rate, R0It is all two two users to the identical locality connection rate having, pmIt is identical right to have for all users
The preference of file m;
Step A2: all popularities are that the file of negative value will not be buffered, Lm={ m1, m2 ... } be for popularity
The listed files of positive value, with qmNon-increasing sequence;
Step A3: judge whether total memory space of the mobile device of user in alliance is less than LmThe size of middle All Files
Summation;
If so, caching LmIn preceding k file, k meetPreceding k-1 text
Part is all cached;The 1/n of k-1 file before each user cache, i.e., K-th of file quilt
Part caches, and the ratio of each user cache isWherein, γ is the phase that all users have
Same mobile device available space,For LmIn i-th of file size;
If it is not, then caching LmIn All Files, i.e.,
The realization of sharing files strategy A is including process: when user i needs file m, first looking at local mobile device
In whether the full content of storage file m;It is set if it is not, obtaining the movement of the user j in same alliance by locality connection
The ratio of the file m of standby middle caching, the file m that user j is shared for the file m of its caching all parts;If same
The remainder of the file m for not having caching whole in user's mobile device in alliance, user i from base station downloading file m.
Preferably, if in user coalitions all users be two classes, wherein it is every one kind in user all have it is identical
To base station communication rate, locality connection rate, file preference and mobile device memory space;And transmission speed in fellow users group group
Rate is higher, with RhIt indicates, and Rh> R0, R0Indicate that all users have identical to base station transmission rate;Inhomogeneity user group
Transmission rate is lower between group, with RlIt indicates;The then realization process of file cache strategy and sharing files strategy are as follows:
When intercorrelation is strong, that is, meet condition Rl> nRhR0/(2Rh+(n-2)R0), calculate All Files m popularity value
It is defined asIts file cache strategy and file cache strategy A principle phase
Together, sharing files strategy is identical as sharing files strategy A principle;
When intercorrelation is weak, that is, meet condition Rl< nRhR0/(2Rh+(n-2)R0)), calculate All Files m popularity value
It is defined asIts file cache strategy is identical as file cache strategy A principle, file point
Enjoy strategy be when user i needs file m, first look in local mobile device whether the full content of storage file m;If
No, then the file m cached in user's j mobile device in same alliance is obtained by locality connection, if user j and user
I is same group, then the ratio of the file m that user j is shared for the file m of its caching all parts, if user j and user
I is different groups, then user j does not share cached file m with user i;If do not had in user's mobile device in same alliance
The remainder of the file m for having caching whole, user i from base station downloading file m.
Compared with the existing technology, the beneficial effects of the present invention are: user is slow using mobile network's terminal progress local file
It deposits, and file-sharing is carried out by local high speed connection, the transmission for substantially reducing mobile network's business based on multimedia is prolonged
Late, the satisfaction of user is improved.
Detailed description of the invention
Fig. 1 is the flow chart of the embodiment of the present invention;
Fig. 2 is the best file cache strategic process figure of user coalitions of the embodiment of the present invention;
Fig. 3 is the schematic diagram of the embodiment of the present invention.
Specific embodiment
Understand for the ease of those of ordinary skill in the art and implement the present invention, with reference to the accompanying drawings and embodiments to this hair
It is bright to be described in further detail, it should be understood that implementation example described herein is merely to illustrate and explain the present invention, not
For limiting the present invention.
A kind of local cache sharing files method based on mobile collective intelligence network provided by the invention, comprising the following steps:
Step 1: the user coalitions pattern based on gunz sharing files network, user coalitions are constructed based on Cooperative reference
Pattern is the set that multiple user files that user is formed share alliance;Referring to Fig.1, its specific implementation includes following sub-step:
Step 1.1: defining a kind of for measuring the superiority and inferiority relationship of alliance pattern S and alliance pattern R
Alliance pattern S is better than alliance's pattern R, all little and if only if service delay of all users at alliance pattern S
In the service delay at alliance pattern R, and service delay of at least one user at alliance pattern S strictly less than
Service delay under alliance pattern R, i.e., And it at least strictly keeps a kind of and is less than pass
System, wherein ci(S) service delay of the user i at alliance pattern S is indicated;
Step 1.2: defining the principle that alliance's pattern updates: combination principle, separation principle and addition principle;
Combination principle: if there is multiple alliances, alliance's pattern after merging is more excellent, i.e., Then merge the alliance { S of all dispersions1, S2..., SlBe
Separation principle: if there is an alliance, multiple alliances can be split as, and alliance's pattern after fractionation is more
It is excellent, i.e.,Then by allianceIt is decomposed into { S1, S2..., Sl};
Principle is added: if an alliance S1Middle there are a Ge Zi alliance S ', are detached from alliance S1Coalize S2Alliance afterwards
Pattern is more excellent, i.e.,Then by alliance S1Subset S ' coalize S2;
Step 1.3: initialization alliance's pattern is the alliance S0 comprising all isolated users;
Step 1.4: judging whether there is the alliance for meeting combination principle;
If it is present updating alliance's pattern according to combination principle, obtaining new alliance's pattern is S;And it connects circulation and holds
Row step 1.4;
If it does not exist, then executing following step 1.5;
Step 1.5: judging whether there is the alliance for meeting separation principle;
If it is present updating alliance's pattern according to separation principle, obtaining new alliance's pattern is S;And it connects circulation and holds
Row step 1.5;
If it does not exist, then executing following step 1.6;
Step 1.6: judging whether there is the alliance for meeting and principle being added;
If it is present updating alliance's pattern according to principle is added, obtaining new alliance's pattern is S;And it connects circulation and holds
Row step 1.6;
If it does not exist, then revolution executes the step 1.7.
Step 1.7: judging whether S0 is equal to S;
If it is not, revolution executes step 1.4;
If so, alliance's pattern that output is stable, this process terminate.
See Fig. 3 (a), in the present embodiment, it is assumed that have 4 users, then the alliance's pattern being likely to form and corresponding user
Service delay it is as shown in the table:
According to superiority and inferiority relationshipIt is found thatBecause of the clothes of user 1 and user 2 in the case where { 12 }
Business delay is 64 and 43, the service delay 70 and 70 respectively less than in the case of { 1 }, { 2 }.
Embodiment realizes that the algorithm of Stable coalitions pattern is as follows:
1) initialization alliance's pattern is { 1 }, { 2 }, { 3 }, { 4 }.
2) alliance { 1 } and alliance { 2 } meet combination principle, merge into { 1,2 }, update alliance's pattern be { 1,2 }, { 3 },
{4}。
3) alliance { 3 } and alliance { 4 } meet combination principle, merge into { 3,4 }, and updating alliance's pattern is { 1,2 }, { 3,4 }.
4) there is no the alliances for meeting combination principle.
5) there is no the alliances for meeting disassembly principle.
6) alliance { 1,2 } and { 3,4 } meet addition principle, and updating alliance's pattern is { 1,2,4 }, { 3 }.
7) there is no meet the alliance that principle is added.
8) there is no the alliances for meeting combination principle.
9) alliance { 1,2,4 } meets disassembly principle, and updating alliance's pattern is { 1 }, { 2,4 }, { 3 }.
10) it there is no fractionation is met, is added or the alliance of combination principle, stable alliance's pattern is { 1 }, { 2,4 }, { 3 }.
Step 2: according to the user coalitions pattern formed using step 1, formulating best file cache for specific user alliance
Strategy and sharing files strategy;
See Fig. 2, the specific implementation of file cache strategy includes following sub-step:
Step 2.1: establishing user-file to (i, m) list
Step 2.2: being based on user preference file distribution, user's locality connection rate, user and base station transmission rate, calculate
Popularity of each user-file to (i, m)
Wherein RI, 0For the transmission rate between user i and base station, RJ, iLocal transmission speed between user i and user j
Rate,It is user i to the preference of file m, smFor the size of file m;
Step 2.3: from listIt is middle to remove user-file pair that all popularities are negative value;
Step 2.4: judging listIt whether is empty;
If so, this process terminates;
If it is not, then continuing to execute step 2.5;
Step 2.5: calculatingIn all user-files pair pre-cache ratio
Wherein γiFor the mobile device available space of user i, ymFor the ratio of the uncached part of file m;
Step 2.6: choosing listMiddle popularity and pre-cache ratio productMaximum user-file pair,
And from listMiddle removal this document pair;
Step 2.7: the ratio of designated user's i cache file m is
Step 2.8: the mobile device available space for updating user i is
Step 2.9: whether the mobile device available space for judging user i is 0;
If so, from listIt is middle to remove all user-files pair comprising user i;And execute following step 2.10;
If it is not, then executing following step 2.10;
Step 2.10: the ratio for updating the uncached part of file m is
Step 2.11: judging that whether the ratio of the uncached part of file m becomes into 0;
If so, from listIt is middle to remove all user-files pair comprising file m;And it turns round and executes the step
2.4;
The step 2.4 is executed if it is not, then turning round.
See Fig. 3 (b), in the present embodiment, it is assumed that there is a user coalitions { 1,2,3 }, caching shares file { a, b, c },
The parameter being related to is as follows:
ListFor (1, a), (1, b), (1, c), (2, a), (2, b), (2, c), (3, a), (3, b), (3, c), according to public affairs
Formula popularity calculated is as follows:
From listIt is middle remove (2, a) and (2, c).
Removing popularity is user-file of negative value to rear, listFor (1, a), (1, b), (1, c), (2, b),
(3, a), (3, b), (3, c).The product of pre-cache ratio and popularity and pre-cache ratio is as follows:
Select maximum user-file to (2, b), the ratio of 2 cache file b of designated user is 0.5, is removed from list
(2, b), the available space of user 2 become 0, and the ratio of the uncached part of file b becomes 0.2.Update pre-cache ratio and people
The product of gas index and pre-cache ratio is as follows:
Select maximum user-file to (3, c), the ratio of 3 cache file c of designated user is 1, is removed from list
The available space of (3, c), user 3 becomes 0, removed from list (3, a) become with (3, b), the ratio of the uncached part of file c
It is 0, removes (1, c) from list.The product for updating pre-cache ratio and popularity and pre-cache ratio is as follows:
Select maximum user-file to (1, b), the ratio of 1 cache file b of designated user is 0.15, is moved from list
Except (1, b), the available space of user 3 becomes 0, removes (1, c) from list, and the ratio of the uncached part of file b becomes
0.05.List is empty.Finally obtained best cache policy is that 1 cache file b ratio of user is 0.15,2 cache file b of user
Ratio is 0.8, and the ratio of 3 cache file c of user is 1.
Wherein the specific implementation process of sharing files strategy is: when user i needs file m, first looking at local movement
In equipment whether the full content of storage file m;It is moved if it is not, obtaining the user j in same alliance by locality connection
The ratio of the file m cached in dynamic equipment, the file m that user j is shared for the file m of its caching all parts;If same
The remainder of the file m for not having caching whole in user's mobile device in one alliance, user i from base station downloading file m.
In the present embodiment, when user 3 needs file c, first look at local mobile device caching, due to 3 cached it is complete
Portion file c, can directly use.When user 3 needs file b, check that local mobile device does not have a cache file b, user 3 from
User 1 and user 2 obtain 0.15 and 0.8 file b respectively, then remaining 0.05 file b is obtained from base station.
Step 3: the file cache strategy and sharing files strategy obtained according to step 2, user download corresponding ratio from base station
Example file to local mobile device.
The present embodiment is after forming user coalitions pattern, and each alliance specifies an alliance responsible person, by alliance responsible person
Calculate with the best file cache of implementation and supervision with share strategy.It can choose the user being located among user coalitions, or have
The user of larger mobile device memory space, or have very fast locally-attached user as alliance responsible person.Alliance is responsible for
People obtains the file preference of user by locality connection, mobile device memory space, and locality connection rate between each user is each to use
The information such as transmission rate are for calculating best file cache and sharing strategy between family and base station.
See Fig. 3 (b), the specific implementation process of the present embodiment is described as follows:
The memory space of user 1,2,3 is respectively 3,16,10, if selection has the use of larger mobile device memory space
Family is as alliance responsible person, then user 2 is alliance responsible person.The average local connection speed of user 1,2,3 is respectively 4,3.5,
4.5, if selecting have very fast locally-attached user as alliance responsible person, select user 3 for alliance responsible person.
In the present embodiment, if user coalitions, by closely located, file preference is similar, the similar user group of mobile device
At that is, all users are having the same to base station communication rate, and locality connection rate, file preference and mobile device storage are empty
Between, then the file cache strategy of each user is identical as sharing files strategy.
Wherein the realization process of file cache strategy A is: the popularity of calculating All Files m first isWherein n is the number of users in alliance.All popularities are negative value (qm< 0)
File will not be buffered, i.e.,Lm={ m1, m2... be popularity be positive value (qm> 0) listed files, with qmIt is non-
Sort ascending.If total memory space of the mobile device of user is less than L in alliancemThe summation of the size of middle All Files, then delay
Deposit LmIn preceding k file, k meetPreceding k-1 file is all cached, each
The 1/n of k-1 file before user cache, i.e.,K-th of file is partially cached, each user
The ratio of caching isIf total memory space of the mobile device of user is not less than L in alliancem
The summation of the size of middle All Files, then cache LmIn All Files, i.e.,
The realization of sharing files strategy A is including process: when user i needs file m, first looking at local mobile device
In whether the full content of storage file m;It is set if it is not, obtaining the movement of the user j in same alliance by locality connection
The ratio of the file m of standby middle caching, the file m that user j is shared for the file m of its caching all parts;If same
The remainder of the file m for not having caching whole in user's mobile device in alliance, user i from base station downloading file m.
See Fig. 3 (c), the specific implementation process of embodiment is described as follows:
Assuming that there is a user coalitions { 1,2,3,4,5,6 }, the complete homogeneity of user, caching shares file { a, b, c, d }, relates to
And the parameter arrived is as follows:
pa | 0.1 | pb | 0.2 | pc | 0.3 | pd | 0.4 |
R | 10 | R0 | 1 | γ | 20 | ||
sa | 10 | sb | 20 | sc | 90 | sd | 60 |
It is as follows according to formula popularity calculated:
qa | qb | qc | qd |
-13.5 | 3.0 | 58.5 | 72.0 |
File a will not be buffered, list LmFor { c, d, b }, preceding 2 files will be buffered, each user cache file c's
Ratio is 1/6=0.1667, and the ratio of each user cache file d is
When user 1 needs file c, its file c cached is obtained respectively from user 2~6, does not need to publish papers under base station
Part.When user 1 needs file d, its file d cached is obtained respectively from user 2~6, then obtain remaining ratio from base station
For 0.5 file d.When user 1 needs file a or b, need directly to download from base station.
In the present embodiment, if user coalitions can be divided into two user groups, every group is n/2 user, with the use in group
Family is fellow users mutually similar, and the user of different groups is user different each other.Transmission rate is (the same as in group in group
The rate of file is transmitted between any two users) it is higher, with RhIt indicates, and Rh> R0.Transmission rate (is appointed in difference group between group
The rate of file is transmitted between what two user) it is lower, with RlIt indicates.Due to the still well-balanced distribution of user, so each user
Cache policy is identical, and calculation is as follows.
When intercorrelation is strong, that is, meet condition Rl> nRhR0/(2Rh+(n-2)R0), calculate All Files m popularity value
It is defined asIts file cache strategy and file cache strategy A principle phase
Together, sharing files strategy is identical as sharing files strategy A principle;
See Fig. 3 (d), the specific implementation process of embodiment is described as follows:
Assuming that there is a user coalitions { 1,2,3,4,5,6 }, it is divided into two user groups { 1,2,3 } and { 4,5,6 }, caching point
File { a, b, c, d } is enjoyed, the parameter being related to is as follows:
pa | 0.1 | pb | 0.2 | pc | 0.3 | pd | 0.4 |
Rh | 10 | Rl | 5 | R0 | 1 | γ | 20 |
sa | 10 | sb | 20 | sc | 90 | sd | 60 |
It can be with design conditions Rl> nRhR0/(2Rh+(n-2)R0) set up, according to the following institute of formula popularity calculated
Show:
qa | qb | qc | qd |
-14.4 | 1.2 | 50.4 | 64.8 |
File a will not be buffered, list LmFor { d, c, b }, preceding 2 files will be buffered, each user cache file d's
Ratio is 1/6=0.1667, and the ratio of each user cache file c is
When user 1 needs file d, its file d cached is obtained respectively from user 2~6, does not need to publish papers under base station
Part.When user 1 needs file c, its file c cached is obtained respectively from user 2~6, then obtain remaining ratio from base station
For 0.3333 file c.When user 1 needs file a or b, need directly to download from base station.
When intercorrelation is weak, that is, meet condition Rl< nRhR0/(2Rh+(n-2)R0)), calculate All Files m popularity value
It is defined asIts file cache strategy is identical as file cache strategy A principle;File point
Enjoy strategy be when user i needs file m, first look in local mobile device whether the full content of storage file m;If
No, then the file m cached in user's i mobile device in same alliance is obtained by locality connection, if user i and user
I is same group, then the ratio of the file m that user i is shared for the file m of its caching all parts, if user i and user
I is different groups, then user i does not share cached file m with user i;If do not had in user's mobile device in same alliance
The remainder of the file m for having caching whole, user i from base station downloading file m.
Assuming that there is a user coalitions { 1,2,3,4,5,6 }, it is divided into two user groups { 1,2,3 } and { 4,5,6 }, caching point
File { a, b, c, d } is enjoyed, the parameter being related to is as follows:
pa | 0.1 | pb | 0.2 | pc | 0.3 | pd | 0.4 |
Rh | 10 | Rl | 2 | R0 | 1 | γ | 20 |
sa | 10 | sb | 20 | sc | 90 | sd | 60 |
It is as follows according to formula popularity calculated:
qa | qb | qc | qd |
-17.1 | -4.2 | 26.1 | 43.2 |
File a, b will not be buffered, list LmFor { d, c }, preceding 2 files will be buffered, each user cache file d's
Ratio is 1/6=0.1667, and the ratio of each user cache file c is
When user 1 needs file d, its file d cached is obtained respectively from user 2,3, user 1 cannot be from user 4~6
File d is obtained, user 1 obtains the file c that remaining ratio is 0.5 from base station again.When user 1 needs file c, from user 2,
3 obtain the file c of its caching respectively, and user 1 cannot obtain file c from user 4~6, and user 1 obtains remaining ratio from base station again
The file c that example is 0.6667.When user 1 needs file a or b, need directly to download from base station.
Present invention is primarily based on collective intelligence networks to form the user coalitions that file cache is shared, and is reduced and is transmitted by locality connection
Delay.This method devises one kind and obtains stablizing with individual rationality, Dhp stability and contract according to Cooperative reference
The distributed algorithm of the user coalitions pattern of property.Introduce the concept of file popularity, it is determined that transmit user coalitions always and prolong
The best cache policy minimized late.The local cache file sharing system based on mobile collective intelligence network obtained through the invention
Mobile network's transmission delay can be reduced.
It should be understood that the part that this specification does not elaborate belongs to the prior art.
It should be understood that the above-mentioned description for preferred embodiment is more detailed, can not therefore be considered to this
The limitation of invention patent protection range, those skilled in the art under the inspiration of the present invention, are not departing from power of the present invention
Benefit requires to make replacement or deformation under protected ambit, fall within the scope of protection of the present invention, this hair
It is bright range is claimed to be determined by the appended claims.
Claims (6)
1. a kind of local cache sharing files method based on mobile collective intelligence network, which comprises the following steps:
Step 1: the user coalitions pattern based on gunz sharing files network, the user coalitions are constructed based on Cooperative reference
Pattern is the set that multiple user files that user is formed share alliance;
The specific implementation of step 1 includes following sub-step:
Step 1.1: defining a kind of for measuring the superiority and inferiority relationship of alliance pattern S and alliance pattern R
Alliance pattern S is better than alliance's pattern R, is all not more than and if only if service delay of all users at alliance pattern S
Service delay under alliance pattern R, and service delay of at least one user at alliance pattern S is strictly less than in alliance
Service delay under pattern R, i.e., And at least strictly keep a kind of and be less than relationship,
In, ci(S) service delay of the user i at alliance pattern S is indicated;Refer to total user set;
Step 1.2: defining the principle that alliance's pattern updates: combination principle, separation principle and addition principle;
Combination principle: if there is multiple alliances, alliance's pattern after merging is more excellent, i.e.,
Then merge the alliance { S of all dispersions1,S2,…,SlBe
Separation principle: if there is an alliance, multiple alliances can be split as, and alliance's pattern after fractionation is more excellent, i.e.,Then by allianceIt is decomposed into { S1,S2,…,Sl};
Principle is added: if an alliance S1Middle there are a Ge Zi alliance S ', are detached from alliance S1Coalize S2Alliance's pattern afterwards
It is more excellent, i.e.,Then by alliance S1Subset S ' coalize S2;
Step 1.3: initialization alliance's pattern is the alliance S0 comprising all isolated users;
Step 1.4: judging whether there is the alliance for meeting combination principle;
If it is present updating alliance's pattern according to combination principle, obtaining new alliance's pattern is S;And it connects circulation and executes step
Rapid 1.4;
If it does not exist, then executing following step 1.5;
Step 1.5: judging whether there is the alliance for meeting separation principle;
If it is present updating alliance's pattern according to separation principle, obtaining new alliance's pattern is S;And it connects circulation and executes step
Rapid 1.5;
If it does not exist, then executing following step 1.6;
Step 1.6: judging whether there is the alliance for meeting and principle being added;
If it is present updating alliance's pattern according to principle is added, obtaining new alliance's pattern is S;And it connects circulation and executes step
Rapid 1.6;
If it does not exist, then executing following step 1.7;
Step 1.7: judging whether S0 is equal to S;
If it is not, revolution executes step 1.4;
If so, alliance's pattern that output is stable, this process terminate;
Step 2: according to the user coalitions pattern formed using step 1, formulating best file cache strategy for specific user alliance
With sharing files strategy;
Step 3: the file cache strategy and sharing files strategy obtained according to step 2, user download corresponding proportion from base station
File is to local mobile device.
2. the local cache sharing files method according to claim 1 based on mobile collective intelligence network, which is characterized in that step
The specific implementation of file cache strategy described in rapid 2 includes following sub-step:
Step 2.1: establishing user-file to (i, m) list
Step 2.2: being based on user preference file distribution, user's locality connection rate, user and base station transmission rate, calculate each
Popularity of the user-file to (i, m)
Wherein Ri,0For the transmission rate between user i and base station, Rj,iLocal transmission rate between user i and user j,
It is user i to the preference of file m, smFor the size of file m;
Step 2.3: from listIt is middle to remove user-file pair that all popularities are negative value;
Step 2.4: judging listIt whether is empty;
If so, this process terminates;
If it is not, then continuing to execute step 2.5;
Step 2.5: calculatingIn all user-files pair pre-cache ratio
Wherein γiFor the mobile device available space of user i, ymFor the ratio of the uncached part of file m;
Step 2.6: choosing listMiddle popularity and pre-cache ratio productMaximum user-file pair, and from column
TableMiddle removal this document pair;
Step 2.7: the ratio of designated user's i cache file m is
Step 2.8: the mobile device available space for updating user i is
Step 2.9: whether the mobile device available space for judging user i is 0;
If so, from listIt is middle to remove all user-files pair comprising user i;And execute following step 2.10;
If it is not, then executing following step 2.10;
Step 2.10: the ratio for updating the uncached part of file m is
Step 2.11: judging whether the ratio of the uncached part of file m becomes 0;
If so, from listIt is middle to remove all user-files pair comprising file m;And it turns round and executes the step 2.4;
The step 2.4 is executed if it is not, then turning round.
3. the local cache sharing files method according to claim 1 based on mobile collective intelligence network, which is characterized in that step
The specific implementation process of sharing files strategy described in rapid 2 is: when user i needs file m, first looking at local mobile device
In whether the full content of storage file m;It is set if it is not, obtaining the movement of the user j in same alliance by locality connection
The ratio of the file m of standby middle caching, the file m that user j is shared for the file m of its caching all parts;If same
The remainder of the file m for not having caching whole in user's mobile device in alliance, user i from base station downloading file m.
4. the local cache sharing files method based on mobile collective intelligence network according to claim 1 to 3,
Be characterized in that: after forming user coalitions pattern, each alliance specifies an alliance responsible person, is calculated and is supervised by alliance responsible person
Execute best file cache strategy and sharing files strategy.
5. the local cache sharing files method based on mobile collective intelligence network according to claim 1 to 3,
It is characterized in that: if all users are having the same to base station communication rate, locality connection rate, file preference in user coalitions
With mobile device memory space, then the file cache strategy of each user is identical as sharing files strategy;
Wherein, the realization of file cache strategy A includes following sub-step:
Step A1: the popularity q of All Files m is calculatedm;
Wherein n is the number of users in alliance, smFor the size of file m, R has identical to base station communication for all users
Rate, R0It is all two two users to the identical locality connection rate having, pmHave for all users identical to file
The preference of m;
Step A2: all popularities are that the file of negative value will not be buffered, Lm={ m1,m2... } and it be popularity is positive value
Listed files, with qmNon-increasing sequence;
Step A3: judge whether total memory space of the mobile device of user in alliance is less than LmThe size of middle All Files it is total
With;
If so, caching LmIn preceding k file, k meetPreceding k-1 file quilt
All cachings;The 1/n of k-1 file before each user cache, i.e., K-th of file is by part
Caching, the ratio of each user cache areWherein, γ has identical for all users
Mobile device available space,For LmIn i-th of file size;
If it is not, then caching LmIn All Files, i.e.,
Wherein, the realization of sharing files strategy A is including process: when user i needs file m, first looking at local movement and sets
In standby whether the full content of storage file m;If it is not, it is mobile to obtain the user j in same alliance by locality connection
The ratio for the file m that the file m cached in equipment, user j are shared for its caching file m all parts;If same
The remainder of the file m for not having caching whole in user's mobile device in alliance, user i from base station downloading file m.
6. the local cache sharing files method based on mobile collective intelligence network according to claim 5 any one, special
Sign is: if in user coalitions all users be two classes, wherein it is every one kind in user all have it is identical to base station communication
Rate, locality connection rate, file preference and mobile device memory space;And transmission rate is higher in fellow users group group, with
RhIt indicates, and Rh>R0, R0Indicate that all users have identical to base station transmission rate;Speed is transmitted between inhomogeneity user group group
Rate is lower, with RlIt indicates;The then realization process of file cache strategy and sharing files strategy are as follows:
When intercorrelation is strong, that is, meet condition Rl>nRhR0/(2Rh+(n-2)R0), it calculates All Files m popularity value and is defined asIts file cache strategy is identical as file cache strategy A principle, file
It is identical as sharing files strategy A principle to share strategy;
When intercorrelation is weak, that is, meet condition Rl<nRhR0/(2Rh+(n-2)R0)), calculate the definition of All Files m popularity value
ForIts file cache strategy is identical as file cache strategy A principle, sharing files plan
Slightly be when user i needs file m, first look in local mobile device whether the full content of storage file m;If not yet
Have, then the file m cached in user's j mobile device in same alliance is obtained by locality connection, if user j and user i
Be same group, then the ratio of the file m that user j is shared for the file m of its caching all parts, if user j and user i
It is different groups, then user j does not share cached file m with user i;If do not had in user's mobile device in same alliance
Cache whole file m, remainder of the user i from base station downloading file m.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610985297.0A CN106708923B (en) | 2016-11-09 | 2016-11-09 | A kind of local cache sharing files method based on mobile collective intelligence network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610985297.0A CN106708923B (en) | 2016-11-09 | 2016-11-09 | A kind of local cache sharing files method based on mobile collective intelligence network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106708923A CN106708923A (en) | 2017-05-24 |
CN106708923B true CN106708923B (en) | 2019-07-09 |
Family
ID=58940865
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610985297.0A Active CN106708923B (en) | 2016-11-09 | 2016-11-09 | A kind of local cache sharing files method based on mobile collective intelligence network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106708923B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109756908B (en) * | 2017-11-01 | 2021-10-15 | 上海交通大学 | Method/system for optimizing wireless network cache strategy, storage medium and equipment |
CN112597388B (en) * | 2020-12-18 | 2022-10-14 | 南京邮电大学 | Cache-enabled D2D communication joint recommendation and caching method |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103763334A (en) * | 2013-05-28 | 2014-04-30 | 大连理工大学 | Multimedia cooperative sharing method based on P2P-BT in VANET |
CN104967685A (en) * | 2015-06-25 | 2015-10-07 | 广东德诚网络科技有限公司 | Streaming media multi-level cache network acceleration method based on Flash P2P |
CN105791391A (en) * | 2016-01-29 | 2016-07-20 | 西安交通大学 | Method for calculating optimal cooperation distance of D2D fusion network based on file popularity |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8533507B2 (en) * | 2009-12-23 | 2013-09-10 | Apple Inc. | Efficient service advertisement and discovery in a peer-to-peer networking environment |
-
2016
- 2016-11-09 CN CN201610985297.0A patent/CN106708923B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103763334A (en) * | 2013-05-28 | 2014-04-30 | 大连理工大学 | Multimedia cooperative sharing method based on P2P-BT in VANET |
CN104967685A (en) * | 2015-06-25 | 2015-10-07 | 广东德诚网络科技有限公司 | Streaming media multi-level cache network acceleration method based on Flash P2P |
CN105791391A (en) * | 2016-01-29 | 2016-07-20 | 西安交通大学 | Method for calculating optimal cooperation distance of D2D fusion network based on file popularity |
Non-Patent Citations (1)
Title |
---|
Game theoretic Approaches for Wireless Proactive Caching;Zhiwen Hu;《IEEE》;20160831;第39-41页 |
Also Published As
Publication number | Publication date |
---|---|
CN106708923A (en) | 2017-05-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Wang et al. | Federated deep reinforcement learning for Internet of Things with decentralized cooperative edge caching | |
Wang et al. | Social data offloading in D2D-enhanced cellular networks by network formation games | |
Zhang et al. | Hierarchical caching for statistical QoS guaranteed multimedia transmissions over 5G edge computing mobile wireless networks | |
Zhong et al. | Deep multi-agent reinforcement learning based cooperative edge caching in wireless networks | |
Hu et al. | Public cloud storage-assisted mobile social video sharing: A supermodular game approach | |
Shivshankar et al. | An evolutionary game theory-based approach to cooperation in VANETs under different network conditions | |
CN101483670B (en) | Regional P2P computation data consistency maintenance method based on distributed hash table | |
Deng et al. | Hybrid data transmission scheme based on source node centrality and community reconstruction in opportunistic social networks | |
CN106708923B (en) | A kind of local cache sharing files method based on mobile collective intelligence network | |
CN107734482B (en) | The content distribution method unloaded based on D2D and business | |
Richerzhagen et al. | Bypassing the cloud: Peer-assisted event dissemination for augmented reality games | |
Sinky et al. | Responsive content-centric delivery in large urban communication networks: A LinkNYC use-case | |
Zhang et al. | Clustered device-to-device caching based on file preferences | |
Zhang et al. | A game-theoretic approach for optimal distributed cooperative hybrid caching in D2D networks | |
CN108810139B (en) | Monte Carlo tree search-assisted wireless caching method | |
Zhang et al. | Two time-scale caching placement and user association in dynamic cellular networks | |
Zhang et al. | Stochastic game based cooperative alternating Q-learning caching in dynamic D2D networks | |
Wang et al. | Interest‐driven avatar neighbor‐organizing for P2P transmission in distributed virtual worlds | |
Kouyoumdjieva et al. | Energy-aware opportunistic mobile data offloading for users in urban environments | |
CN105447188B (en) | A kind of reciprocity social networks document retrieval method of knowledge based study | |
CN109640298B (en) | Mobile D2D content sharing method and system based on alliance formation game | |
CN108596778B (en) | Community division method based on interest space | |
Thilakarathna et al. | User generated content dissemination in mobile social networks through infrastructure supported content replication | |
Shiroma et al. | A light-weight cooperative caching strategy by D2D content sharing | |
Somuyiwa et al. | Multicast-aware proactive caching in wireless networks with deep reinforcement learning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |