CN108848395A - Edge cooperation caching method for arranging based on drosophila optimization algorithm - Google Patents
Edge cooperation caching method for arranging based on drosophila optimization algorithm Download PDFInfo
- Publication number
- CN108848395A CN108848395A CN201810521001.9A CN201810521001A CN108848395A CN 108848395 A CN108848395 A CN 108848395A CN 201810521001 A CN201810521001 A CN 201810521001A CN 108848395 A CN108848395 A CN 108848395A
- Authority
- CN
- China
- Prior art keywords
- video
- drosophila
- cache
- individual
- glo
- 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
- 241000255581 Drosophila <fruit fly, genus> Species 0.000 title claims abstract description 122
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000005457 optimization Methods 0.000 title claims abstract description 30
- 230000009467 reduction Effects 0.000 claims abstract description 7
- 108090000623 proteins and genes Proteins 0.000 claims description 84
- 229910002056 binary alloy Inorganic materials 0.000 claims description 12
- 239000000796 flavoring agent Substances 0.000 claims description 12
- 235000019634 flavors Nutrition 0.000 claims description 12
- 230000000007 visual effect Effects 0.000 claims description 12
- 230000005540 biological transmission Effects 0.000 claims description 10
- 238000012216 screening Methods 0.000 claims description 7
- 239000012141 concentrate Substances 0.000 claims description 6
- 238000012804 iterative process Methods 0.000 claims description 6
- 230000008439 repair process Effects 0.000 claims description 6
- 230000035945 sensitivity Effects 0.000 claims description 6
- 230000003111 delayed effect Effects 0.000 claims 1
- 235000013399 edible fruits Nutrition 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 235000013305 food Nutrition 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
- H04N21/231—Content storage operation, e.g. caching movies for short term storage, replicating data over plural servers, prioritizing data for deletion
- H04N21/23106—Content storage operation, e.g. caching movies for short term storage, replicating data over plural servers, prioritizing data for deletion involving caching operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/21—Server components or server architectures
- H04N21/222—Secondary servers, e.g. proxy server, cable television Head-end
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
- H04N21/239—Interfacing the upstream path of the transmission network, e.g. prioritizing client content requests
- H04N21/2393—Interfacing the upstream path of the transmission network, e.g. prioritizing client content requests involving handling client requests
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/25—Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
- H04N21/251—Learning process for intelligent management, e.g. learning user preferences for recommending movies
- H04N21/252—Processing of multiple end-users' preferences to derive collaborative data
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
Abstract
The invention discloses a kind of edge cooperation caching method for arranging based on drosophila optimization algorithm, includes the following steps:(1) according to the historical requests information of area's intra domain user, the popular video collection and user demand vector in the region are obtained;(2) according to the popular video collection and user demand vector, objective optimisation problems are established as target to maximize total video propagation delay time reduction amount in region, and the objective optimisation problems are solved based on drosophila optimization algorithm, generates caching arrangement decision;It (3) is that each cache node distributes video cache task according to caching arrangement decision;(4) when user requests to reach on cache node, if the cache node does not cache corresponding contents, then from the content has been cached and the smallest neighbouring cache node of time delay downloads the content, if all cache nodes in the region all do not have cache responses content, downloaded from remote server.Cache hit rate can be improved in the present invention, reduces average video propagation delay time, improves user experience quality.
Description
Technical field
The present invention relates to edge cache technology more particularly to a kind of edge cooperation caching arrangements based on drosophila optimization algorithm
Method.
Background technique
With the rapid development of mobile communication business, the quantity of mobile device is significantly increased, the diversification of user demand trend,
High efficiency, mobile data flow also increase.Smart phone brings better web experience to user, as social application,
Audio-video service etc., wherein video playing service occupies wherein more than half data traffics.According to the investigation of Cisco, with 2012
Year is compared, and mobile data flow total amount increases by 13 times within 2017.From 2015 to the year two thousand twenty, mobile video flow will increase by 11
Times, account for the 75% of mobile data flow total amount.In order to cope with this phenomenon, edge cache technology is come into being.Caching technology is
By placing in most popular content to the node closer to request user, passback load is effectively reduced.But due to node
Limited storage space, how by these limited memory spaces more " flexible ", " clever " be used to cache popular content at
In order to need the problem of studying.
The main target of caching placement policies is to determine the content for needing to cache for the node of caching and these nodes.It passes
The caching placement policies of system are mainly single-point caching, and mode of operation is simple, but content redundancy is higher.Cooperation caching placement policies
This deficiency is improved by cooperation mode different between node, reduces delay.
Recently, a kind of emerging video coding technique scalable video (Scalable Video Coding, SVC)
It is widely used in applications such as network service, video storages, allows a variety of spatial resolutions (screen size), different frame speed
Rate or noise specific mass, make video be satisfied higher user experience quality.Using SVC, each video file is encoded
At a component layers, these layerings can achieve the video quality of user demand by combining.For example, in order to user's transmission quality
Grade is the video of Q, and all layers from the 1st layer to Q layers of the video require to be transmitted.For requesting a certain credit rating
The user of video, different layers content are received by him, are decoded and are played (rather than serial broadcasting) simultaneously.Under this configuration, it passes
Defeated time delay is determined by time delay is maximum from all layers of caching node-node transmission.Based on SVC, cache policy becomes significantly multiple
It is miscellaneous.
Summary of the invention
Goal of the invention:In view of the problems of the existing technology the present invention, provides a kind of edge based on drosophila optimization algorithm
Cooperation caching method for arranging, this method can the demand according to user to different video, same video different quality copy make
Corresponding cache decision reduces average video propagation delay time to improve cache hit rate, improves user experience quality.
Technical solution:Edge cooperation caching method for arranging of the present invention based on drosophila optimization algorithm includes following step
Suddenly:
(1) according to the historical requests information of area's intra domain user, obtain the region popular video collection be denoted as V=1,
2 ... } and user demand vector;
(2) according to the popular video collection and user demand vector, reduced with maximizing total video propagation delay time in region
Amount is target, establishes objective optimisation problems, and solve the objective optimisation problems based on drosophila optimization algorithm, generates caching arrangement and determines
Plan;
(3) according to the caching arrangement decision be each cache node distribute video cache task, cache node according to divide
The task buffer video matched;
(4) when user requests to reach on cache node, if the cache node does not cache corresponding contents, from caching
The content and time delay the smallest neighbouring cache node downloads the content, if all cache nodes in the region are all rung without caching
Content is answered, then is downloaded from remote server.
Further, (2) specifically include:
(2.1) to maximize total video propagation delay time reduction amount in region, as target, establishing objective optimisation problems is:
In formula,Total video propagation delay time in region when for no nodal cache content;NmTo own in current region m
The set of cache node, n are cache node serial number;V be demand video serial number, each video v there are Q layers of set L=1,
2 ..., Q }, ovlL layers of the size of the expression of > 0 video v;Q indicates user demand credit rating, and in order to transmit matter to user
It measures the video v that grade is q, all layers from the 1st layer to q layers of the video need to be required to be transmitted, i.e., it is shared
A byte, and ov1≥ov2≥....≥ovQ, λnvqIndicate in region n-th of node to v-th video in q credit rating
The average user demand of appearance, binary variable xnvlWhether will be placed on node n, the x if placing if indicating l layers of video vnvl
=1, otherwise xnvl=0;dnFor the unit time delay for being transferred to cache node n from content server, n* is to have cached required video layer
And it is transferred to the node ID that cache node n has minimum-time lag, dnn*To be transferred on cache node n from cache node n*
Unit time delay;CnFor the capacity of cache node n;The nodal cache strategy x of region mmIt is given by:
And each node cannot cache data more more than its capacity, i.e.,
(2.2) objective optimisation problems are solved by drosophila optimization algorithm, generates caching arrangement decision.
Further, step (2.2) specifically includes:
(2.2.1) spatial cache of each cache node is divided into local popular spatial cache and global popular caching is empty
Between two parts, and distinguished by setting parameter F ∈ [0,1], F represents global popular video file and accounts for the ratio integrally cached,
The part remaining 1-F caches local popular video file;
The popular spatial cache of the overall situation of all cache nodes in (2.2.2) layout area:Based on drosophila optimization algorithm to global flow
The higher video of row degree is screened, and the video layer content filtered out is put into for the highest node of video local demand
n#∈NmThe popular spatial cache of the overall situation in, generate the cache policy of video set contentAnd
Ensure at most FCnThe video content of+s size is buffered in node n#Place, wherein s is the full-size of all video layers;
The local popular spatial cache of all cache nodes in (2.2.3) layout area:It is screened based on drosophila optimization algorithm
The higher video of local popularity out, remaining spatial cache is filled, and is updatedAnd
The local popular spatial cache of each cache node does not repeat the video that caching is present in the global popular spatial cache of the node
Layer.
Further, the global higher video screening technique of popularity specifically includes in step (2.2.2):
(2.2.2.1) video layer is selected:The time delay for calculating each video layer of each video saves cost performance utiglo-vi,
And pick out uti in each videoglo-viMaximum video layer lv;Wherein:
utiglo-vi=valueglo-vi/weightglo-vi
In formula,It indicates to cache the video layer band in region
Come time delay saving, l(j∈{1...i})For target function, when subscript condition is true, l(j∈{1...i})=1, it is otherwise 0, and for l=
q,ovl+1=0;Indicate that popular video concentrates the size of the i-th layer video layer of v-th of video;
(2.2.2.2) initializes relevant parameter:Define the binary system drosophila gene sequence x that length is vglo-v, work as xglo-v=
When 1, the l of video vvLayer is placed on spatial cache, is not otherwise put into, and defines iteration total degree I, initializes the number of iterations i=
0, initialize probability vector p in iterative processivValue p0vDrosophila visual sensitivity b is arranged in=[0.5,0.5,0.5 ...];
(2.2.2.3) initializes center drosophila population:According to the probability vector p of i-th iterationiv, to xglo-vIt carries out X times
Random initializtion obtains X group binary system gene order, i.e., the center drosophila population being made of X drosophila individual;
The part (2.2.2.4) smell search:For each center drosophila individual, L are randomly choosed, and selected bits are taken
Instead, i.e., drosophila gene becomes 1 from 0 or becomes 0 from 1, repeats S times, generates S neighbours' drosophila individual;
(2.2.2.5) repairs compensating operation:It is individual for each center drosophila individual and neighbours drosophila, according to
utiglo-viEach gene is detected from small to large, if this is 1 and the corresponding weight of this gene orderglo>
weightglo-lim, then by the position, gene sets 0 and weightgloThe size that the gene position corresponds to video layer content is subtracted, until
weightglo> weightglo-lim;According to utiglo-viEach gene is detected from big to small, if the position is 0 and this gene order
Corresponding weightglo> weightglo-lim, then by the position, gene sets 1 and weightglo-viIn addition the corresponding view of the gene position
The size of frequency layer content, remains weight in the processglo> weightglo-lim;Wherein, weightglo-lim=NF
Cn, N is cache node number in region, weightgloFor the size of video layers all in gene order;
The search of (2.2.2.6) local visual:Compensated center drosophila individual and its neighbours are repaired for each, are found
The wherein maximum drosophila individual of individual flavor concentration, as local optimum individual, and replaces the center fruit with local optimum individual
Fly individual, the drosophila population updated;
The search of (2.2.2.7) overall Vision:For updated drosophila population, it is maximum to find wherein individual flavor concentration
Drosophila individual is denoted as x as global optimized individual, gene orderi,fg, other two drosophila individual in random selection population
Gene order is denoted as xi,f1、xi,f2, and global optimized individual cooperate modify probability vector together, updated probability vector isWherein Δi=xi,fg+0.5(xi,f1+xi,f2);
(2.2.2.8) determines current iteration optimal solution:Judge whether to reach maximum number of iterations, be, then exports optimal solution
xi,fg, and the video layer filtered out according to optimal solution, otherwise jump to (2.2.2.3).
Further, the local higher video screening technique of popularity specifically includes in step (2.2.3):
(2.2.3.1) video layer is selected:The time delay for calculating each video layer of each video saves cost performance utiloc-vi,
And pick out uti in each videoloc-viMaximum video layer lv;Wherein:
utiloc-vi=valueloc-vi/weightloc-vi
In formula,It indicates to cache the video layer in region
Time delay is brought to save, l(j∈{1...i})For target function, when subscript condition is true, l(j∈{1...i})=1, it is otherwise 0, and for l
=q, ovl+1=0;Indicate that popular video concentrates the size of the i-th layer video layer of v-th of video;
(2.2.3.2) initializes relevant parameter:Define the binary system drosophila gene sequence x that length is vloc-v, work as xloc-v=
When 1, the l of video vvLayer is placed on node n, is not otherwise put into, and defines iteration total degree I, initializes the number of iterations i=0, just
Probability vector p in beginningization iterative processivValue p0vDrosophila visual sensitivity b is arranged in=[0.5,0.5,0.5 ...];
(2.2.3.3) initializes center drosophila population:According to the probability vector p of i-th iterationiv, to xloc-vIt carries out X times
Random initializtion obtains X group binary system gene order, i.e., the center drosophila population being made of X drosophila individual;
The part (2.2.3.4) smell search:For each center drosophila individual, L are randomly choosed, and selected bits are taken
Instead, i.e., drosophila gene becomes 1 from 0 or becomes 0 from 1, repeats S times, generates S neighbours' drosophila individual;
(2.2.3.5) repairs compensating operation:It is individual for each center drosophila individual and neighbours drosophila, according to
utiloc-viEach gene is detected from small to large, if this is 1 and the corresponding weight of this gene orderloc>
weightloc-lim, then by the position, gene sets 0 and weightlocThe size that the gene position corresponds to video layer content is subtracted, until
weightloc> weightloc-lim;According to utiloc-viEach gene is detected from big to small, if the position is 0 and this gene order
Corresponding weightloc> weightloc-lim, then by the position, gene sets 1 and weightlocIn addition the gene position corresponds to video
The size of layer content, remains weight in the processloc> weightloc-lim, weightloc-lim=Cn-Cglo-n, Cglo-n
For the size for being buffered in all video layers in node n, weightlocFor the size of video layers all in gene order;
The search of (2.2.3.6) local visual:Compensated center drosophila individual and its neighbours are repaired for each, are found
The wherein maximum drosophila individual of individual flavor concentration, as local optimum individual, and replaces the center fruit with local optimum individual
Fly individual, the drosophila population updated;
The search of (2.2.3.7) overall Vision:For updated drosophila population, it is maximum to find wherein individual flavor concentration
Drosophila individual is denoted as x as global optimized individual, gene orderi,fg, other two drosophila individual in random selection population
Gene order is denoted as xi,f1、xi,f2, and global optimized individual cooperate modify probability vector together, updated probability vector isWherein Δi=xi,fg+0.5(xi,f1+xi,f2);
(2.2.3.8) determines current iteration optimal solution:Judge whether to reach maximum number of iterations, be, then exports optimal solution
xi,fg, and the video layer filtered out according to optimal solution, otherwise jump to (2.2.3.3).
Beneficial effect:Compared with prior art, the present invention its remarkable advantage is:
(1) the present invention is based on cache policy of the drosophila optimization algorithm to SVC video to carry out decision, average to minimize region
Video transmission delay.Primary dcreening operation is carried out to video layer content by time delay cost performance, MCK problem reduction is selected into knapsack for single choice and is asked
Topic, reduces the complexity of problem.Meanwhile drosophila optimization algorithm is the middle evolution algorithm deduced out of looking for food from drosophila, compared to
Traditional greedy algorithm, optimization process is simple, complexity is lower, is easy to Project Realization, while can obtain preferably caching plan
Slightly.
(2) present invention cooperating between consideration node in edge cache design, reduces region content redundancy, mentions
The high Buffer Utilization of spatial cache.
(3) present invention considers that the spatial cache of single node divides, in the popular video of caching the serviced user of this node
Hold and (optimize local demand) and is obtained between other node users popular video contents (optimizing global demand) of caching areal
Optimal tradeoff.
Detailed description of the invention
Fig. 1 is application scenarios schematic diagram applied by the present invention;
Fig. 2 is the flow diagram of one embodiment of the present of invention;
Fig. 3 is the flow diagram of drosophila optimization algorithm.
Specific embodiment
Attached drawing 1 is the typical case scene of the edge cooperation caching method based on drosophila optimization algorithm.Consideration one is general
The network architecture, in the architecture, K network operator (NOs) or network entity (constituting cache node set K) are to be distributed in
The user group (constituting user set M) of M geographic area provides Internet access service.For each region, in user to far
On the path for holding content server, each cache node is mounted with to cache in specific position.NOk ∈ K is represented in a certain region
Node, NkRepresent the set of all nodes (i.e. NOk) in the region.CnRepresent cache contents n ∈ NkSize, and Cn≥0
(byte).Video content composition set V=1,2 ... }.Assuming that in special time period (for example, a few houres or several days), to every
The average user demand of a video content is fixed, and assumes that user demand is known in advance (for example, first by analysis
The statistical information of user's request mode of preceding time may infer that following user's request, can using this learning method
Obtain user demand information).Definition set L={ 1,2 ..., Q }, each video can be transmitted with Q ∈ Z+ credit rating.
That is, each video content has one Q layers of set of quality classes L={ 1,2 ..., Q }, they are realized by accumulation
Different credit ratings:1 grade of quality may be implemented in layer 1 content itself, and 1 content of layer is just able to achieve 2 grades of matter in conjunction with 2 content of layer
Amount, and so on.L layers of the size o of video vvlIndicate ovl>0 (byte), and it is usually as l reduces, that is, ov1≥
ov2≥....≥ovQ.Node NkThe user serviced is when requesting video content, it may be necessary to different credit rating requirements.Example
Such as, there is Q=2 credit rating in total, there may be user's request of half to be required of the view of low definition quality (q=1)
Frequently, the other half then requires high definition quality (q=2).λnvqIndicate in region n-th of node to v-th video in q mass etc.
The average user demand of grade content.We can define respectively each cache node n request vector and NOk aggregate demand to
Amount:λn=(λnvq:v∈V,q∈Q),λk=(λn:n∈Nk)。
In order to provide a user video v q grades of credit ratings content, need to transmit video v from the first order to q grades
All credit ratings content, that is, need to transmit in totalByte.In stream video system, the segment of different layers
It is received simultaneously along each beam direction, decodes and be arranged broadcasting, rather than serially play.Under this configuration, transmission of video is transmitted to the end
The constraint of layer, therefore the time delay for transmitting entire video will be equal to the maximum video layer of propagation delay time.
A kind of situation is considered now, and the cache node in the same area can cooperate using respective spatial cache, different
Still (this model also contemplated another situation to the cache node in region, and network provides interior different content, but these independently of one another
Video content collection is overlapped in this case, and V represents the video content set of overlapping).In this case, a caching
Node can send video layer content to another cache node to meet the user demand of other side.Assuming that every in set K
The user that a NO is serviced requests in the same video content set V.It oneself is serviced nevertheless, each NO has
User, they may generate different user demands, i.e. λk1≠λk2.Define the set of all cache nodesWith
Their total user demands
Ideally, user wants to all layers required for receiving from local node n of content, to obtain
Minimum-time lag, without loss of generality, it is assumed that this is zero with reference to time delay.If can not find a certain user in local cache node n to ask
The video layer content asked, then n can download it from another node n ' of the same area buffered content.dnnShow this
The caused unit data time delay of transmission, in other words, the every layer of content requested from adjacent node will be with constant Mean Speed
Transmission, the Mean Speed can be expressed as 1/dnn.For example, this can be realized by using parallel TCP connection, every layer of company
It connects, each connection distributes fixed-bandwidth.While one can consider thatIf should without nodal cache in region
Content, then remote server can be with time delay dn>dnn(for the same area) transmitted video content.Obviously, Yong Huke
To download the video layer content of user demand from different cache nodes or server.Time delay when user demand is satisfied will wait
Maximum delay needed in varied situations.The target of NOs cooperation is to reduce average total content transmission time delay, all to meet
User's request.
With x=(xk:K ∈ K) indicate joint cache policy, then, overall delay can be expressed asWherein:
Here, Mn∈ M indicates the region where node n.If without the view of nodal cache user demand in one's respective area
Frequency layer content l ∈ { 1,2 ..., q }, i.e.,The content will be by remote server with dnUnit data
Time delay is transferred to local node n.Otherwise, content 1, which will have been cached it, in one's respective area can generate the node-node transmission of minimum-time lag.
When multiple cache nodes enter offline agreement cooperation, put together equal to their cache resources pond, so as to
Cache the layered video content that the user of other networks specifies.Determine the cache policy for the total user's time delay for minimizing all NO
Problem can be expressed as follows:
Wherein,
Since content can only be transmitted between the node of the same area, the above problem can be decomposed into M independent sons
Problem, each region m ∈ M.We useIndicate the node set for being located at region m.For specific region m, we are observed
To be to the total user's time delay not cached:Because all user's requests are required from distal end
Server downloads the 1st layer video content (this is maximum layer in all layers).By downloading required layer content from cache node,
It can reduce from remote server and download content bring overall delay.In a certain region m, video transmission delay will be maximized and subtracted
A small amount of problems (referred to as Rm) are expressed as follows:
ForIfI.e. the region does not have nodal cache
It, the l layer content of video v will be by content server with time delay ovldnIt is transferred to node n, otherwise, l layers of all cachings in one's respective area
In the node of content, the smallest node of time delay will transmit it.
Fig. 2 is the implementation flow chart of the edge cooperation caching method based on drosophila optimization algorithm, as shown, including following
Step:
(1) according to the historical requests information of area's intra domain user, obtain the region popular video collection be denoted as V=1,
2 ... } and user demand vector.
(2) according to the popular video collection and user demand vector, reduced with maximizing total video propagation delay time in region
Amount is target, establishes objective optimisation problems, and solve the objective optimisation problems based on drosophila optimization algorithm, generates caching arrangement and determines
Plan.
It (3) is that each cache node distributes video cache task according to caching arrangement decision;
(4) when user requests to reach on cache node, if the cache node does not cache corresponding contents, from caching
The content and time delay the smallest neighbouring cache node downloads the content, if all cache nodes in the region are all rung without caching
Content is answered, then is downloaded from remote server.
Further, (2) specifically include:
(2.1) to maximize total video propagation delay time reduction amount in region, as target, establishing objective optimisation problems is:
(2.2) objective optimisation problems are solved by drosophila optimization algorithm, generates caching arrangement decision.
Step (2.2) specifically includes:
(2.2.1) spatial cache of each cache node is divided into local popular spatial cache and global popular caching is empty
Between two parts, and distinguished by setting parameter F ∈ [0,1], F represents global popular video file and accounts for the ratio integrally cached,
The part remaining 1-F caches local popular video file;
The popular spatial cache of the overall situation of all cache nodes in (2.2.2) layout area:Based on drosophila optimization algorithm to the overall situation
The higher video of popularity is screened, and the video layer content filtered out is put into for the highest node of video local demand
n#∈NmThe popular spatial cache of the overall situation in, generate the cache policy of video set content
And ensure at most FCnThe video content of+s size is buffered in node n#Place, wherein s is the full-size of all video layers.
Wherein, buffered video screening problem can be mapped as more options knapsack (Multiple-choice Knapsack,
MCK) problem, i.e., the most video layer contents of selection (are mapped as in each video (the article class being mapped as in MCK problem)
Article item in MCK problem) it is put into spatial cache (being mapped as the knapsack in MCK problem), so that overall delay reduction amount (is mapped as
Total value in MCK problem) it maximizes, all buffered video layer content size weightgloIt (is mapped as in MCK problem
Total weight) it is no more than weightglo-lim(being mapped as the weight limitation in MCK problem).As shown in figure 3, specific overall situation popularity
Higher video screening technique includes:
(2.2.2.1) video layer is selected:The time delay for calculating each video layer of each video saves cost performance utiglo-vi,
And pick out uti in each videoglo-viMaximum video layer lv;Wherein:
utiglo-vi=valueglo-vi/weightglo-vi
In formula,It indicates to cache the video layer band in region
Come time delay saving, l(j∈{1...i})For target function, when subscript condition is true, l(j∈{1...i})=1, it is otherwise 0, and for l=
q,ovl+1=0;Indicate that popular video concentrates the size of the i-th layer video layer of v-th of video;
(2.2.2.2) initializes relevant parameter:Define the binary system drosophila gene sequence x that length is vglo-v, work as xglo-v=
When 1, the l of video vvLayer is placed on spatial cache, is not otherwise put into, and defines iteration total degree I, initializes the number of iterations i=
0, initialize probability vector p in iterative processivValue p0vDrosophila visual sensitivity b is arranged in=[0.5,0.5,0.5 ...];
(2.2.2.3) initializes center drosophila population:According to the probability vector p of i-th iterationiv, to xglo-vIt carries out X times
Random initializtion obtains X group binary system gene order, i.e., the center drosophila population being made of X drosophila individual;
The part (2.2.2.4) smell search:For each center drosophila individual, L are randomly choosed, and selected bits are taken
Instead, i.e., drosophila gene becomes 1 from 0 or becomes 0 from 1, repeats S times, generates S neighbours' drosophila individual;
(2.2.2.5) repairs compensating operation:It is individual for each center drosophila individual and neighbours drosophila, according to
utiglo-viEach gene is detected from small to large, if this is 1 and the corresponding weight of this gene orderglo>
weightglo-lim, then by the position, gene sets 0 and weightgloThe size that the gene position corresponds to video layer content is subtracted, until
weightglo> weightglo-lim;According to utiglo-viEach gene is detected from big to small, if the position is 0 and this gene order
Corresponding weightglo> weightglo-lim, then by the position, gene sets 1 and weightglo-viIn addition the corresponding view of the gene position
The size of frequency layer content, remains weight in the processglo> weightglo-lim;Wherein, weightglo-lim=NF
Cn, N is cache node number in region, weightgloFor the size of video layers all in gene order;
The search of (2.2.2.6) local visual:Compensated center drosophila individual and its neighbours are repaired for each, are found
The wherein maximum drosophila individual of individual flavor concentration, as local optimum individual, and replaces the center fruit with local optimum individual
Fly individual, the drosophila population updated;
The search of (2.2.2.7) overall Vision:For updated drosophila population, it is maximum to find wherein individual flavor concentration
Drosophila individual is denoted as x as global optimized individual, gene orderi,fg, other two drosophila individual in random selection population
Gene order is denoted as xi,f1、xi,f2, and global optimized individual cooperate modify probability vector together, updated probability vector isWherein Δi=xi,fg+0.5(xi,f1+xi,f2);
(2.2.2.8) determines current iteration optimal solution:Judge whether to reach maximum number of iterations, be, then exports optimal solution
xi,fg, and the video layer filtered out according to optimal solution, otherwise jump to (2.2.2.3).
The local popular spatial cache of all cache nodes in (2.2.3) layout area:It is screened based on drosophila optimization algorithm
The higher video of local popularity out, remaining spatial cache is filled, and is updatedAnd
The local popular spatial cache of each cache node does not repeat the video that caching is present in the global popular spatial cache of the node
Layer.
Similarly, the higher video screening technique of local popularity specifically includes:
(2.2.3.1) video layer is selected:The time delay for calculating each video layer of each video saves cost performance utiloc-vi,
And pick out uti in each videoloc-viMaximum video layer lv;Wherein:
utiloc-vi=valueloc-vi/weightloc-vi
In formula,It indicates to cache the video layer in region
Time delay is brought to save, l(j∈{1...i})For target function, when subscript condition is true, l(j∈{1...i})=1, it is otherwise 0, and for l
=q, ovl+1=0;Indicate that popular video concentrates the size of the i-th layer video layer of v-th of video;
(2.2.3.2) initializes relevant parameter:Define the binary system drosophila gene sequence x that length is vloc-v, work as xloc-v=
When 1, the l of video vvLayer is placed on node n, is not otherwise put into, and defines iteration total degree I, initializes the number of iterations i=0, just
Probability vector p in beginningization iterative processivValue p0vDrosophila visual sensitivity b is arranged in=[0.5,0.5,0.5 ...];
(2.2.3.3) initializes center drosophila population:According to the probability vector p of i-th iterationiv, to xloc-vIt carries out X times
Random initializtion obtains X group binary system gene order, i.e., the center drosophila population being made of X drosophila individual;
The part (2.2.3.4) smell search:For each center drosophila individual, L are randomly choosed, and selected bits are taken
Instead, i.e., drosophila gene becomes 1 from 0 or becomes 0 from 1, repeats S times, generates S neighbours' drosophila individual;
(2.2.3.5) repairs compensating operation:It is individual for each center drosophila individual and neighbours drosophila, according to
utiloc-viEach gene is detected from small to large, if this is 1 and the corresponding weight of this gene orderloc>
weightloc-lim, then by the position, gene sets 0 and weightlocThe size that the gene position corresponds to video layer content is subtracted, until
weightloc> weightloc-lim;According to utiloc-viEach gene is detected from big to small, if the position is 0 and this gene order
Corresponding weightloc> weightloc-lim, then by the position, gene sets 1 and weightlocIn addition the gene position corresponds to video
The size of layer content, remains weight in the processloc> weightloc-lim, weightloc-lim=Cn-Cglo-n, Cglo-n
For the size for being buffered in all video layers in node n, weightlocFor the size of video layers all in gene order;
The search of (2.2.3.6) local visual:Compensated center drosophila individual and its neighbours are repaired for each, are found
The wherein maximum drosophila individual of individual flavor concentration, as local optimum individual, and replaces the center fruit with local optimum individual
Fly individual, the drosophila population updated;
The search of (2.2.3.7) overall Vision:For updated drosophila population, it is maximum to find wherein individual flavor concentration
Drosophila individual is denoted as x as global optimized individual, gene orderi,fg, other two drosophila individual in random selection population
Gene order is denoted as xi,f1、xi,f2, and global optimized individual cooperate modify probability vector together, updated probability vector isWherein Δi=xi,fg+0.5(xi,f1+xi,f2);
(2.2.3.8) determines current iteration optimal solution:Judge whether to reach maximum number of iterations, be, then exports optimal solution
xi,fg, and the video layer filtered out according to optimal solution, otherwise jump to (2.2.3.3).
Above disclosed is only a preferred embodiment of the present invention, and the right model of the present invention cannot be limited with this
It encloses, therefore equivalent changes made in accordance with the claims of the present invention, is still within the scope of the present invention.
Claims (5)
1. the edge cooperation caching method for arranging based on drosophila optimization algorithm, it is characterised in that:Include the following steps:
(1) according to the historical requests information of area's intra domain user, the popular video collection for obtaining the region is denoted as V={ 1,2 ... }, with
And user demand vector;
(2) according to the popular video collection and user demand vector, it is to maximize total video propagation delay time reduction amount in region
Target establishes objective optimisation problems, and solves the objective optimisation problems based on drosophila optimization algorithm, generates caching arrangement decision;
It (3) is that each cache node distributes video cache task according to caching arrangement decision, cache node is according to distribution
Task buffer video;
(4) when user requests to reach on cache node, if the cache node does not cache corresponding contents, from having cached this
The content and the smallest neighbouring cache node of time delay downloads the content, if all cache nodes in the region are all without in cache responses
Hold, is then downloaded from remote server.
2. the edge cooperation caching method for arranging according to claim 1 based on drosophila optimization algorithm, it is characterised in that:
(2) it specifically includes:
(2.1) to maximize total video propagation delay time reduction amount in region, as target, establishing objective optimisation problems is:
In formula,Total video propagation delay time in region when for no nodal cache content;NmFor cachings all in current region m section
The set of point, n are cache node serial number;V be demand video serial number, each video v there are Q layers of set L=1,2 ...,
Q }, ovlL layers of the size of the expression of > 0 video v;Q indicates user demand credit rating, and in order to user's transmission quality grade
For the video v of q, all layers from the 1st layer to q layers of the video need to be required to be transmitted, i.e., it is sharedA byte,
And ov1≥ov2≥....≥ovQ, λnvqIndicate that n-th of node is to being averaged in q credit rating content of v-th of video in region
User demand, binary variable xnvlWhether will be placed on node n, the x if placing if indicating l layers of video vnvl=1, otherwise
xnvl=0;dnFor the unit time delay for being transferred to cache node n from content server, n* is to have cached required video layer and be transferred to
Cache node n has the node ID of minimum-time lag, dnn*To be transferred to the unit time delay on cache node n from cache node n*;
CnFor the capacity of cache node n;The nodal cache strategy x of region mmIt is given by:
And each node cannot cache data more more than its capacity, i.e.,
(2.2) objective optimisation problems are solved by drosophila optimization algorithm, generates caching arrangement decision.
3. the edge cooperation caching method for arranging according to claim 2 based on drosophila optimization algorithm, it is characterised in that:Step
Suddenly (2.2) specifically include:
The spatial cache of each cache node is divided into local popular spatial cache and global popular spatial cache two by (2.2.1)
Part, and distinguished by setting parameter F ∈ [0,1], the video file that F represents global prevalence accounts for the ratio integrally cached, remaining
The part 1-F cache local popular video file;
The popular spatial cache of the overall situation of all cache nodes in (2.2.2) layout area:Based on drosophila optimization algorithm to global flow
The higher video of row degree is screened, and the video layer content filtered out is put into for the highest node of video local demand
n#∈NmThe popular spatial cache of the overall situation in, generate the cache policy of video set content
And ensure at most FCnThe video content of+s size is buffered in node n#Place, wherein s is the full-size of all video layers;
The local popular spatial cache of all cache nodes in (2.2.3) layout area:This is filtered out based on drosophila optimization algorithm
The higher video of ground popularity, remaining spatial cache is filled, and is updatedAnd it is each
The local popular spatial cache of cache node does not repeat the video layer that caching is present in the global popular spatial cache of the node.
4. the edge cooperation caching method for arranging according to claim 3 based on drosophila optimization algorithm, it is characterised in that:Step
Suddenly the global higher video screening technique of popularity specifically includes in (2.2.2):
(2.2.2.1) video layer is selected:The time delay for calculating each video layer of each video saves cost performance utiglo-vi, and choose
Select uti in each videoglo-viMaximum video layer lv;Wherein:
utiglo-vi=valueglo-vi/weightglo-vi
In formula,Indicate that the video layer is cached in region brings time delay
It saves, l(j∈{1...i})For target function, when subscript condition is true, l(j∈{1...i})=1, it is otherwise 0, and for l=q, ovl+1
=0;Show that popular video concentrates the size of the i-th layer video layer of v-th of video;
(2.2.2.2) initializes relevant parameter:Define the binary system drosophila gene sequence x that length is vglo-v, work as xglo-vWhen=1,
The l of video vvLayer is placed on spatial cache, is not otherwise put into, and defines iteration total degree I, initializes the number of iterations i=0, just
Probability vector p in beginningization iterative processivValue p0vDrosophila visual sensitivity b is arranged in=[0.5,0.5,0.5 ...];
(2.2.2.3) initializes center drosophila population:According to the probability vector p of i-th iterationiv, to xglo-vProgress X times random first
Beginningization obtains X group binary system gene order, i.e., the center drosophila population being made of X drosophila individual;
The part (2.2.2.4) smell search:For each center drosophila individual, L are randomly choosed, and selected bits are negated,
I.e. drosophila gene becomes 1 from 0 or becomes 0 from 1, repeats S times, generates S neighbours' drosophila individual;
(2.2.2.5) repairs compensating operation:For each center drosophila individual and neighbours' drosophila individual, according to utiglo-viFrom
It is small to detecting each gene greatly, if this is 1 and the corresponding weight of this gene orderglo> weightglo-lim, then will
The position gene sets 0 and weightgloThe size that the gene position corresponds to video layer content is subtracted, until weightglo>
weightglo-lim;According to utiglo-viDetect each gene from big to small, if this be 0 and this gene order it is corresponding
weightglo> weightglo-lim, then by the position, gene sets 1 and weightglo-viIn addition the gene position corresponds to video layer content
Size, remain weight in the processglo> weightglo-lim;Wherein, weightglo-lim=NFCn, N is area
Cache node number in domain, weightgloFor the size of video layers all in gene order;
The search of (2.2.2.6) local visual:Compensated center drosophila individual and its neighbours are repaired for each, are found wherein
The individual maximum drosophila individual of flavor concentration, replaces the center drosophila as local optimum individual, and with local optimum individual
Body, the drosophila population updated;
The search of (2.2.2.7) overall Vision:For updated drosophila population, the wherein maximum drosophila of individual flavor concentration is found
Individual is denoted as x as global optimized individual, gene orderi,fg, the gene of other two drosophila individual in random selection population
Sequence is denoted as xi,f1、xi,f2, and global optimized individual cooperate modify probability vector together, updated probability vector isWherein Δi=xi,fg+0.5(xi,f1+xi,f2);
(2.2.2.8) determines current iteration optimal solution:Judge whether to reach maximum number of iterations, be, then exports optimal solution xi,fg,
And the video layer filtered out according to optimal solution, otherwise jump to (2.2.2.3).
5. the edge cooperation caching method for arranging according to claim 3 based on drosophila optimization algorithm, it is characterised in that:Step
Suddenly the local higher video screening technique of popularity specifically includes in (2.2.3):
(2.2.3.1) video layer is selected:The time delay for calculating each video layer of each video saves cost performance utiloc-vi, and choose
Select uti in each videoloc-viMaximum video layer lv;Wherein:
utiloc-vi=valueloc-vi/weightloc-vi
In formula,Indicate that the video layer is cached in region to be brought
Time delay is saved, l(j∈{1...i})For target function, when subscript condition is true, l(j∈{1...i})=1, it is otherwise 0, and for l=q,
ovl+1=0;Indicate that popular video concentrates the size of the i-th layer video layer of v-th of video;
(2.2.3.2) initializes relevant parameter:Define the binary system drosophila gene sequence x that length is vloc-v, work as xloc-vWhen=1,
The l of video vvLayer is placed on node n, is not otherwise put into, and defines iteration total degree I, initializes the number of iterations i=0, initialization
Probability vector p in iterative processivValue p0vDrosophila visual sensitivity b is arranged in=[0.5,0.5,0.5 ...];
(2.2.3.3) initializes center drosophila population:According to the probability vector p of i-th iterationiv, to xloc-vProgress X times random first
Beginningization obtains X group binary system gene order, i.e., the center drosophila population being made of X drosophila individual;
The part (2.2.3.4) smell search:For each center drosophila individual, L are randomly choosed, and selected bits are negated,
I.e. drosophila gene becomes 1 from 0 or becomes 0 from 1, repeats S times, generates S neighbours' drosophila individual;
(2.2.3.5) repairs compensating operation:For each center drosophila individual and neighbours' drosophila individual, according to utiloc-viFrom
It is small to detecting each gene greatly, if this is 1 and the corresponding weight of this gene orderloc> weightloc-lim, then will
The position gene sets 0 and weightlocThe size that the gene position corresponds to video layer content is subtracted, until weightloc>
weightloc-lim;According to utiloc-viDetect each gene from big to small, if this be 0 and this gene order it is corresponding
weightloc> weightloc-lim, then by the position, gene sets 1 and weightlocIn addition the gene position corresponds to video layer content
Size remains weight in the processloc> weightloc-lim, weightloc-lim=Cn-Cglo-n, Cglo-nTo be delayed
There are the size of all video layers in node n, weightlocFor the size of video layers all in gene order;
The search of (2.2.3.6) local visual:Compensated center drosophila individual and its neighbours are repaired for each, are found wherein
The individual maximum drosophila individual of flavor concentration, replaces the center drosophila as local optimum individual, and with local optimum individual
Body, the drosophila population updated;
The search of (2.2.3.7) overall Vision:For updated drosophila population, the wherein maximum drosophila of individual flavor concentration is found
Individual is denoted as x as global optimized individual, gene orderi,fg, the gene of other two drosophila individual in random selection population
Sequence is denoted as xi,f1、xi,f2, and global optimized individual cooperate modify probability vector together, updated probability vector isWherein Δi=xi,fg+0.5(xi,f1+xi,f2);
(2.2.3.8) determines current iteration optimal solution:Judge whether to reach maximum number of iterations, be, then exports optimal solution xi,fg,
And the video layer filtered out according to optimal solution, otherwise jump to (2.2.3.3).
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810521001.9A CN108848395B (en) | 2018-05-28 | 2018-05-28 | Edge cooperative cache arrangement method based on fruit fly optimization algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810521001.9A CN108848395B (en) | 2018-05-28 | 2018-05-28 | Edge cooperative cache arrangement method based on fruit fly optimization algorithm |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108848395A true CN108848395A (en) | 2018-11-20 |
CN108848395B CN108848395B (en) | 2020-08-25 |
Family
ID=64213608
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810521001.9A Expired - Fee Related CN108848395B (en) | 2018-05-28 | 2018-05-28 | Edge cooperative cache arrangement method based on fruit fly optimization algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108848395B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109974360A (en) * | 2019-03-19 | 2019-07-05 | 东南大学 | A kind of refrigeration system temperature optimization control method based on drosophila algorithm |
CN110505293A (en) * | 2019-08-15 | 2019-11-26 | 东南大学 | Cooperation caching method based on improved drosophila optimization algorithm in a kind of mist wireless access network |
CN111372096A (en) * | 2020-03-12 | 2020-07-03 | 重庆邮电大学 | D2D-assisted video quality adaptive caching method and device |
CN111432270A (en) * | 2020-03-09 | 2020-07-17 | 重庆邮电大学 | Real-time service delay optimization method based on layered cache |
CN112954026A (en) * | 2021-01-29 | 2021-06-11 | 重庆邮电大学 | Multi-constraint content collaborative cache optimization method based on edge calculation |
CN115865712A (en) * | 2022-11-15 | 2023-03-28 | 重庆邮电大学 | Edge data storage method based on aging intimacy model |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140310761A1 (en) * | 2007-07-24 | 2014-10-16 | Time Warner Cable Enterprises Llc | Generation, distribution and use of content metadata in a network |
US20160031083A1 (en) * | 2014-07-31 | 2016-02-04 | Siemens Industry Software Ltd. | Method and apparatus for industrial robotic energy saving optimization using fly-by |
CN105407004A (en) * | 2015-12-08 | 2016-03-16 | 清华大学深圳研究生院 | Method and device for performing content distribution based on edge wireless hotspots |
CN105491156A (en) * | 2016-01-08 | 2016-04-13 | 华中科技大学 | SD-RAN-based whole network collaborative content caching management system and method |
CN106658598A (en) * | 2016-12-08 | 2017-05-10 | 南京邮电大学 | Service migration method based on content caching and network state awareness |
CN106686655A (en) * | 2017-03-20 | 2017-05-17 | 重庆邮电大学 | Heterogeneous network joint user correlation and content cache method |
CN106790617A (en) * | 2016-12-30 | 2017-05-31 | 北京邮电大学 | Collaborative content cache control system and method |
CN107302571A (en) * | 2017-06-14 | 2017-10-27 | 北京信息科技大学 | Information centre's network route and buffer memory management method based on drosophila algorithm |
CN107483587A (en) * | 2017-08-21 | 2017-12-15 | 清华大学 | A kind of power telecom network cache optimization method of content oriented |
CN107909108A (en) * | 2017-11-15 | 2018-04-13 | 东南大学 | Edge cache system and method based on content popularit prediction |
-
2018
- 2018-05-28 CN CN201810521001.9A patent/CN108848395B/en not_active Expired - Fee Related
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140310761A1 (en) * | 2007-07-24 | 2014-10-16 | Time Warner Cable Enterprises Llc | Generation, distribution and use of content metadata in a network |
US20160031083A1 (en) * | 2014-07-31 | 2016-02-04 | Siemens Industry Software Ltd. | Method and apparatus for industrial robotic energy saving optimization using fly-by |
CN105407004A (en) * | 2015-12-08 | 2016-03-16 | 清华大学深圳研究生院 | Method and device for performing content distribution based on edge wireless hotspots |
CN105491156A (en) * | 2016-01-08 | 2016-04-13 | 华中科技大学 | SD-RAN-based whole network collaborative content caching management system and method |
CN106658598A (en) * | 2016-12-08 | 2017-05-10 | 南京邮电大学 | Service migration method based on content caching and network state awareness |
CN106790617A (en) * | 2016-12-30 | 2017-05-31 | 北京邮电大学 | Collaborative content cache control system and method |
CN106686655A (en) * | 2017-03-20 | 2017-05-17 | 重庆邮电大学 | Heterogeneous network joint user correlation and content cache method |
CN107302571A (en) * | 2017-06-14 | 2017-10-27 | 北京信息科技大学 | Information centre's network route and buffer memory management method based on drosophila algorithm |
CN107483587A (en) * | 2017-08-21 | 2017-12-15 | 清华大学 | A kind of power telecom network cache optimization method of content oriented |
CN107909108A (en) * | 2017-11-15 | 2018-04-13 | 东南大学 | Edge cache system and method based on content popularit prediction |
Non-Patent Citations (1)
Title |
---|
WEN等: ""Random Caching Based Cooperative Transmission in Heterogeneous Wireless Networks"", 《IEEE》 * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109974360A (en) * | 2019-03-19 | 2019-07-05 | 东南大学 | A kind of refrigeration system temperature optimization control method based on drosophila algorithm |
CN109974360B (en) * | 2019-03-19 | 2020-10-27 | 东南大学 | Fruit fly algorithm-based refrigeration system temperature optimization control method |
CN110505293A (en) * | 2019-08-15 | 2019-11-26 | 东南大学 | Cooperation caching method based on improved drosophila optimization algorithm in a kind of mist wireless access network |
CN110505293B (en) * | 2019-08-15 | 2021-11-12 | 东南大学 | Cooperative caching method based on improved drosophila optimization algorithm in fog wireless access network |
CN111432270A (en) * | 2020-03-09 | 2020-07-17 | 重庆邮电大学 | Real-time service delay optimization method based on layered cache |
CN111432270B (en) * | 2020-03-09 | 2022-03-11 | 重庆邮电大学 | Real-time service delay optimization method based on layered cache |
CN111372096A (en) * | 2020-03-12 | 2020-07-03 | 重庆邮电大学 | D2D-assisted video quality adaptive caching method and device |
CN111372096B (en) * | 2020-03-12 | 2022-02-18 | 重庆邮电大学 | D2D-assisted video quality adaptive caching method and device |
CN112954026A (en) * | 2021-01-29 | 2021-06-11 | 重庆邮电大学 | Multi-constraint content collaborative cache optimization method based on edge calculation |
CN115865712A (en) * | 2022-11-15 | 2023-03-28 | 重庆邮电大学 | Edge data storage method based on aging intimacy model |
Also Published As
Publication number | Publication date |
---|---|
CN108848395B (en) | 2020-08-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108848395A (en) | Edge cooperation caching method for arranging based on drosophila optimization algorithm | |
Pang et al. | Toward smart and cooperative edge caching for 5G networks: A deep learning based approach | |
US10986387B1 (en) | Content management for a distributed cache of a wireless mesh network | |
CN101800731B (en) | Network transmission management server, network transmission management method and network transmission system | |
CN111866601B (en) | Cooperative game-based video code rate decision method in mobile marginal scene | |
CN101431530A (en) | Method for caching content data packages in caching nodes | |
CN108600014A (en) | A kind of storage resource distribution method based on Stackelberg games | |
CN105979274A (en) | Distributive cache storage method for dynamic self-adaptive video streaming media | |
CN113282786B (en) | Panoramic video edge collaborative cache replacement method based on deep reinforcement learning | |
Sinky et al. | Responsive content-centric delivery in large urban communication networks: A LinkNYC use-case | |
Zhou et al. | QoE-aware 3D video streaming via deep reinforcement learning in software defined networking enabled mobile edge computing | |
Araldo et al. | Representation selection problem: Optimizing video delivery through caching | |
CN108769729A (en) | Caching arrangement system based on genetic algorithm and caching method | |
CN114863683B (en) | Heterogeneous Internet of vehicles edge computing unloading scheduling method based on multi-objective optimization | |
Baccour et al. | CE-D2D: Collaborative and popularity-aware proactive chunks caching in edge networks | |
CN109408211A (en) | A kind of equity network flow medium system data scheduling algorithm of multiple-objection optimization | |
Baccour et al. | Collaborative hierarchical caching and transcoding in edge network with CE-D2D communication | |
Yang et al. | Collaborative edge caching and transcoding for 360° video streaming based on deep reinforcement learning | |
Ye et al. | VRCT: A viewport reconstruction-based 360 video caching solution for tile-adaptive streaming | |
Nguyen et al. | An adaptive streaming method of 360 videos over HTTP/2 protocol | |
Shi et al. | CoLEAP: Cooperative learning-based edge scheme with caching and prefetching for DASH video delivery | |
Yang et al. | Enhancement or super-resolution: Learning-based adaptive video streaming with client-side video processing | |
CN106209952A (en) | Service node distribution method, device, CDN management server and system | |
Liang et al. | Adjustable Two‐Tier Cache for IPTV Based on Segmented Streaming | |
Zou et al. | Joint bandwidth allocation, data scheduling and incentives for scalable video streaming over peer-to-peer networks |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200825 |