CN108966241A - A kind of optimization method of adaptive impovement fish-swarm algorithm - Google Patents
A kind of optimization method of adaptive impovement fish-swarm algorithm Download PDFInfo
- Publication number
- CN108966241A CN108966241A CN201810685277.0A CN201810685277A CN108966241A CN 108966241 A CN108966241 A CN 108966241A CN 201810685277 A CN201810685277 A CN 201810685277A CN 108966241 A CN108966241 A CN 108966241A
- Authority
- CN
- China
- Prior art keywords
- node
- pixel
- fish
- behavior
- visual
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/18—Network planning tools
- H04W16/20—Network planning tools for indoor coverage or short range network deployment
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/02—Arrangements for optimising operational condition
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W84/00—Network topologies
- H04W84/18—Self-organising networks, e.g. ad-hoc networks or sensor networks
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Farming Of Fish And Shellfish (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of optimization methods of adaptive impovement fish-swarm algorithm, belong to sensor network Covering domain.By increased jump behavior, speed and efficiency that the shoal of fish jumps out local optimum are improved；In conjunction with K neighboring mean value thought, new food concentration determination method is set, the probability for falling into local optimum is reduced；Add rebuffed behavior, the perfect critical processing to coverage situation at network boundary；By decay factor θ, the adaptive step-length for adjusting fish and the visual field ensure convergent stability.The present invention is under the premise of guaranteeing the network coverage, fish-swarm algorithm is solved the problems, such as to severe local optimum delay of response, reduce the probability for falling into local optimum simultaneously, the perfect critical processing to coverage situation at network boundary has ensured convergent stability.
Description
Technical field
The present invention relates to a kind of optimization methods of adaptive impovement fish-swarm algorithm, belong to sensor network Covering domain.
Background technique
In wireless sensor network performance optimization problem, the network coverage has become an important research direction.Greatly
Quantity sensor node is placed in monitoring environment by way of shedding.Since node density is big and the randomness of deployment, usually
There can be a large amount of redundant node, lead to network coverage inefficiency.Intelligent algorithm is generallyd use, mobile sensor node is passed through
Mode, increase the effective area of covering, improve covering performance.
During carrying out node optimizing using intelligent algorithm, often fall into local optimum (node is gathered closely together on a large scale)
Unfavorable situation, influence algorithm performance, and then reduce Searching efficiency.In order to promote covering performance, while solving local optimum and asking
Topic, generally use improvement or many algorithms in conjunction with mode cope with.Have a kind of fictitious force at present and population combines
Algorithm (Virtual Force Diminishing Particle Swarm Optimization, VFDPSO), passes through the whole network model
Whole optimizing in enclosing obtains high coverage rate.VFDPSO, which realizes automatic optimal and the network coverage, to be improved, but due to failing to overcome
Typical particle group optimizing method (Particle Swarm Optimization, PSO) is easily trapped into the chronic illness of local optimum, because
The stability of this covering performance is not high.In view of this, " the wireless sensor network coverage optimization based on artificial fish-swarm algorithm " (is published
Source: computer application research, 2013,30 (02): 554-556) by fish-swarm algorithm introduce sensor network covering problem research
In, by the behaviors such as bunch, knock into the back of the simulation shoal of fish, objective function is promoted to jump out local optimum.Although AFSA solves part
Optimal problem, but algorithm takes a long time.Algorithm timeliness is improved in order to solve annual reporting law, " wireless sensor network covers efficiency
Optimization Simulation " (publish source: Computer Simulation, 2017,34 (08): 297-301+345) propose a kind of fish for merging fictitious force
Group's algorithm merges fictitious force by introducing, improves the speed for jumping out local optimum, but the algorithm still has to severe part
Peak optimization reaction is slow, and later period coverage rate promotes the defects of unobvious.
It is above-mentioned have been proven that fish-swarm algorithm really and can be applied to solve local optimum trap, but jump out the speed of trap
Degree, reaction and algorithm in face of local optimum need to mention in the promotion of covering performance and the processing of border issue
It is high.
Summary of the invention
In order to solve the problems, such as that presently, there are fish-swarm algorithms to severe local optimum delay of response, the present invention provides it is a kind of from
Adapt to improve the optimization method of fish-swarm algorithm, the technical solution is as follows:
Assuming that shedding N number of isomorphism sensor node S={ S at random in detected region I1(x1,y1),S2(x2,y2),
..Sk(xk,yk),...,SN(xN,yN), wherein k=1,2,3 ..., N；Node is all made of the Boolean sense model that radius is r；
For ease of description, turning to l × l pixel collection M={ M for region I is discrete simultaneously(1,1),M(1,2),...,
M(x,y),...,M(l,l), wherein x, y=1,2,3 ..., N, and the node for setting random placement all falls within some pixel in I
Point on；
Step 1: each node Sk5 behaviors of looking for food, bunch, repel, jump and be rebuffed are executed respectively, and in notice board
The position after behavior optimizing is recorded in Billboard；
Step 2: calculating network coverage Y after front-wheel optimizing；
Step 3: if being improved before gained network coverage Y is relative to optimizing after front-wheel optimizing, node set S is pressed
It is moved according to the location point of Billboard record；Otherwise according to whether reaching local optimal searching upper limit Cth, judge whether to need to decline
Subtract visual field visual and step-length step；
Wherein, step indicates that fish executes the step-length once moved, and step meets step=step0× rand, rand are indicated
(0,1) random number between；step0Indicate fish execute a behavior it is mobile apart from radix；
Step 4: judging Y > YthOr visual < VthIt is whether true, algorithm is exited if set up；If not, return to step
Rapid 2；Wherein YthAnd VthIndicate the end threshold value of visual and Y.
Optionally, the jump behavior is defined as follows:
In node SkVisual in, appoint take some pixel M(x,y)；
If node SkMeet: Densityk> QthAnd pk≥Pth, then SkJump behavior is executed, pixel ArgMax is jumped to
(T(x,y))；
Wherein, pixel M(x,y)MeetIf | ArgMax (T(x,y)) | > 1, then SkAt random
Jump to ArgMax (T(x,y)) any of pixel；ArgMax(T(x,y)) indicate T(x,y)When obtaining maximum value corresponding (x, y)
Coordinate；
Definition node SkThe sum of coverage rate of pixel is node S in sensing range rkCrowding Densityk:
M(e,f)MeetWhereinExpression is rounded downwards r；
Food concentration T(x,y)Is defined as: pixel M(x,y)With parameter K, then pixel M(x,y)Food concentration T(x,y)
With M(x,y)K neighborhood in pixel coverage rate it is related:
K neighboring mean value method is applied to T(x,y)In calculating, realize M(x,y)The covering of pixel in surrounding K neighborhood
Situation is included in T(x,y)Calculating scope；By setting different weight coefficients for each etale neighborhoodCharacterize M(x,y)Neighborhood
Influence difference between interior each layer；
K neighborhood indicates M(x,y)The K layer neighbour of surrounding, optional K=3.
Optionally, the rebuffed behavior is defined as follows:
If node SkVertical range d between the boundary I of region meetsThen node SkIt is moved towards boundary opposite direction
It is dynamic, and moving distance is rand × (r-d)；
Update visual and step: given CthAfter the optimizing of wheel, if convergence increment deficiency occurs in Y, i.e. Y is sought through excessive wheel
Do not meet increment index after excellent, then update visual using decay factor θ and reduces the visual field and step-length；More new formula is as follows:
Visual=visual × θ (3)
Step=step × θ (4)
Wherein, decay factor θ=0.8
Optionally, the network coverage Y is defined as:
Wherein, as pixel M(x,y)With any SkThe distance between be less than r when, then it is assumed that M(x,y)It can be covered by S, i.e. network
S is in point M(x,y)Coverage rate are as follows:
Wherein, as pixel M(x,y)With sensor node SkThe distance between be less than r when, then it is assumed that SkIn M(x,y)Perception
Rate is 1, is otherwise 0, it may be assumed that
Optionally, described look for food, bunch, repulsion behavior is defined respectively as:
Foraging behavior: in node SkVisual in, appoint take some pixel M(x,y)If M(x,y)MeetThen think point M(x,y)Food concentration be higher than SkThe food concentration of position,
SkTo M(x,y)Mobile step；If given choose under constraint, fail to find point M(x,y), then current fish SkStart random walk；
Wherein, D () indicates the Euclidean distance between two elements；
Bunch behavior: if node SkMeetWhen, SkTo the mobile step in current nearest fish direction；
Repulsion behavior: if node SkMeetWhen, SkIt is mobile to current fish opposite direction recently
step；
Wherein, α and β is the distance between node, i.e. α=2r when sensing range boundary that α value is two nodes is tangent
=16m.Repel threshold value beta and differentiation setting, setting are carried out according to experiment demand
Optionally, before the step 1, further includes:
Parameters are initialized.
Optionally, the method is applied in the optimization of wireless sensor network covering performance.
Optionally, described to be applied in the optimization of wireless sensor network covering performance be specially to be applied to lake blue algae to detect
In.
Optionally, described to be applied in the optimization of wireless sensor network covering performance be specially to be applied to home for destitute promptly to search
In seeking.
The medicine have the advantages that
By increased jump behavior, speed and efficiency that the shoal of fish jumps out local optimum are improved；Think in conjunction with K neighboring mean value
Think, sets new food concentration determination method, reduce the probability for falling into local optimum；Add rebuffed behavior, it is perfect to net
The critical processing of network boundary coverage situation；By decay factor θ, the adaptive step-length for adjusting fish and the visual field ensure and receive
The stability held back.The present invention solves fish-swarm algorithm and reacts slow to severe local optimum under the premise of guaranteeing the network coverage
Slow problem, while reducing the probability for falling into local optimum, the perfect critical part to coverage situation at network boundary
Reason, has ensured convergent stability.
Detailed description of the invention
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for
For those of ordinary skill in the art, without creative efforts, it can also be obtained according to these attached drawings other
Attached drawing.
Fig. 1 is that figure is compared in influence of the network size to the network coverage in algorithms of different；
Fig. 2 is algorithms of different coverage rate and mobile number relationship analysis figure in local optimum, wherein (a) Test=50,
(b) Test=40, (c) Test=30, (d) Test=20；
Fig. 3 is the AIFS coverage rate speedup analysis result figure in the case of different local optimums.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with attached drawing to embodiment party of the present invention
Formula is described in further detail.
Embodiment one:
The present embodiment provides a kind of optimization methods of adaptive impovement fish-swarm algorithm, and it is as follows to arrange experiment scene: given
Detected region I in, shed N number of isomorphism sensor node S={ S at random1(x1,y1),S2(x2,y2),..Sk(xk,
yk),...,SN(xN,yN), node is all made of the Boolean sense model that radius is r, wherein k=1,2,3 ..., N.For convenient for retouching
It states, while turning to l × l pixel collection M={ M for region I is discrete(1,1),M(1,2),...,M(x,y),...,M(l,l), wherein
X, y=1,2,3 ..., N, and the node for setting random placement is all fallen on some pixel in I.
A kind of optimization method of adaptive impovement fish-swarm algorithm provided by the invention, it is included the step of it is as follows:
Step 1: parameters initialization.
Step 2: each node Sk5 behaviors of looking for food, bunch, repel, jump and be rebuffed are executed respectively, and in notice
The position after behavior optimizing is recorded in plate Billboard.
Step 3: Y after front-wheel optimizing is calculated.
Step 4: if when front-wheel optimizing gained network coverage Y before optimizing relative to being improved, node set S will
It is moved according to the location point that Billboard is recorded；Otherwise according to whether reaching local optimal searching upper limit Cth, judge whether to need
Decay visual field visual and step-length step, to reduce the visual field and step-length, into new round optimizing.
Step indicates that fish executes the step-length once moved, and step meets step=step0× rand, rand indicate (0,1)
Between a random number.step0Indicate fish execute a behavior it is mobile apart from radix.
Step 5: judge Y > YthOr visual < VthIt is whether true, algorithm is exited if set up.If not, return to step
Rapid 2.Wherein YthAnd VthIndicate the end threshold value of visual and Y.
Behavior explanation
Before explaining in detail behavior, first to several key definitions in the present invention, agree as follows:
Define 1 (pixel perception rate): as pixel M(x,y)With sensor node SkThe distance between be less than r when, then recognize
For SkIn M(x,y)Perception rate be 1, be otherwise 0, it may be assumed that
Define 2 (pixel dot coverages): as pixel M(x,y)With any SkThe distance between be less than r when, then it is assumed that M(x,y)
It can be covered by S, i.e. network S is in point M(x,y)Coverage rate are as follows:
It defines 3 (network coverages): defining ratio of the sum of the coverage rate of all pixels point with pixel sum in network,
For network coverage Y, it may be assumed that
By formula (6) and (5) it is found that meeting P (M(x,y), S) and=1 pixel quantity is more, and it can more effective raising
Coverage rate Y of the network to detection zone I.Therefore the target of sensor network covering performance optimization, exactly mobile there are redundancies to cover
The node of lid increases meet P (M as far as possible(x,y), S) and=1 pixel.
In step 2, the behavior in basic fish-swarm algorithm is executed, including look for food, bunched, and repelled；Three behaviors are such as
Under:
Behavior 1 (foraging behavior): in current some node SkVisual in, appoint take some pixel M(x,y)If M(x,y)
MeetThen think point M(x,y)Food concentration be higher than SkThe food of position
Concentration, SkTo M(x,y)Mobile step.If given choose under constraint, fail to find point M(x,y), then current fish SkStart random walk.
Wherein, D () indicates the Euclidean distance between two elements.
Behavior 2 (behavior of bunching): if current SkMeetWhen, SkIt is moved to current nearest fish direction
Dynamic step.
Behavior 3 (repels behavior): if current SkMeetWhen, SkIt is moved to current fish opposite direction recently
Dynamic step.
Wherein, α and β is the distance between node, i.e. α=2r when sensing range boundary that α value is two nodes is tangent
=16m.Repel threshold value beta and differentiation setting, setting are carried out according to experiment demand
Other than basic fish school behavior, increase by two kinds of behaviors, is jump behavior and rebuffed behavior respectively.
Behavior 4 (jump behavior): if node SkMeet: Densityk> QthAnd pk≥Pth, then SkJump behavior is executed, is jumped
It jumps to pixel ArgMax (T(x,y)), wherein meetingIf | ArgMax (T(x,y)) | > 1, then Sk
Random skip is to ArgMax (T(x,y)) any of pixel.Wherein, ArgMax (T(x,y)) indicate T(x,y)Obtain maximum value when pair
(x, the y) coordinate answered.
It defines 4 (crowdings): defining some node SkThe sum of the coverage rate of pixel in sensing range r, for node Sk's
Crowding Densityk
M(e,f)It should meetWhereinExpression is rounded downwards r.DensitykIt embodies
SkThe degree of crowding between neighbor node is to measure SkIt whether there is the important parameter of redundant node in sensing range.
Define 5 (food concentration T(x,y)): given pixel point M(x,y)With parameter K, then pixel M(x,y)Food concentration
T(x,y)With M(x,y)K neighborhood in pixel coverage rate it is related, be defined as follows:
K neighboring mean value method is applied to T(x,y)In calculating, realize M(x,y)The covering of pixel in surrounding K neighborhood
Situation is included in T(x,y)Calculating scope.By setting different weight coefficients for each etale neighborhoodCharacterize M(x,y)Neighborhood
Influence difference between interior each layer.
Behavior 5 (rebuffed behavior): if node SkVertical range d between the boundary I of region meetsThen sensor section
Point is mobile towards boundary opposite direction, and moving distance is rand × (r-d).
In order to further enhance convergence, when convergence increment deficiency, by adjusting step-length and view
Open country, to improve algorithm overall precision.
Visual and step updates: given CthAfter the optimizing of wheel, if convergence increment deficiency occurs in Y, i.e. Y is sought through excessive wheel
Do not meet increment index after excellent, then update visual using decay factor θ and reduces the visual field and step-length.More new formula is as follows:
Visual=visual × θ (3)
Step=step × θ (4)
Reduce the value of visual and step, in order to promote low optimization accuracy, it is increased to improve network coverage acquisition
Possibility facilitates the operation of algorithm continuous and effective.
A kind of effect of optimization of the optimization method of the adaptive fish-swarm algorithm proposed for the present invention will be described in detail, it is following to borrow
Experimental data is helped to be illustrated:
This experiment adjusts network size by adjusting number of nodes N value, to investigate AIFS, particle group optimizing method
A kind of (Particle Swarm Optimization, PSO) and coverage optimization algorithm (Coverage based on sample information
Optimization algorithm based on Sampling for Homogeneous WSNs, COSH) the network coverage
Rate Y is by network size N effect.Experiment setting N ∈ [30,65] is sought with 20 wheel of spacing progress that increases of 5 interstitial contents
It is excellent.
The result of comparative experiments is as shown in Figure 1.Wherein, origin indicates the Y calculated in primal environment.As seen from the figure, with
The increase of N, the effect of optimization of each algorithm pair all promoted, as number of nodes N > 40, promoted effect it is more significant.Although
Y after each algorithm optimization increased, but promotion of the intelligence class algorithm PSO and AIFS to Y, hence it is evident that calculate better than non intelligent class
Method COSH, and AIFS optimizing effect is better than PSO.In particular, as the further increasing of network size (N > 50) is compared with PSO,
The promotion gain of the network coverage of AIFS is more prominent, and reason is to increase with N, and the crowding between each node also exists
Different degrees of increase possesses jump and repels the AIFS of behavior, embodied better constringency performance.
In order to verify the ability that AIFS jumps out local optimum, i.e. sensor node can quickly find to fall into the clustering of node,
And quickly take off the ability of group.In region, the case where local optimum, is simulated in the center of I, and node collection S is all arranged in I at random
Center position side length is wherein Test=50 in the square of Test, 40,30,20.Investigate AIFS, fish-swarm algorithm AFSA
(Artificial Fish Swarm Algorithm) and population fictitious force combination algorithm VFDPSO (Virtual Force
Diminishing Particle Swarm Optimization), when node each round is mobile, the variation of network coverage Y
Situation.Experimental result is as shown in Figure 2.
As shown in Fig. 2, when node falls into local optimum, by multiple movement, AIFS, AFSA and tri- kinds of VFDPSO
Algorithm can raising Y in various degree so that sensor network jumps out local optimum trap.But it is mentioned compared to VSDPSO coverage rate Y
The Y ascending curve of the local irregularities risen, AIFS and AFSA is more smooth, therefore when cope with local optimum, AIFS with
AFSA has the more stable ability for jumping out local optimum.The speed that correlation curve rises, AIFS are jumping out local optimum
In speed, hence it is evident that be better than AFSA, this advantage with local optimum deterioration (Test ↓), and it is more obvious.
In order to which more detail analysis AIFS jumps out the ability of local optimum, node motion number is indicated with times, with DY
It indicates speedup of the Y relative to previous movement, extracts AIFS respectively in four kinds of local optimums, 5 time points (times=2,4,
6,8,10) DY when, statistical data are as shown in Figure 3；After node motion times=10 times, the coverage rate increment of VFDPSO and AFSA
Value DY is gradually reduced with the reduction of Test.It is worth noting that in the case of for any one given local optimum,
AIFS can obtain the coverage rate gain of DY=7~14% in earlier time point (times=2), it is meant that most severe when facing
Local optimum when, AIFS is at algorithm initial stage, so that it may acceleration jump out local optimum, embody stronger adaptability.
The method is applied in lake blue algae detection, for example is applied in the blue algae monitoring of Taihu Lake:
If to be monitored cyanobacteria, need to dispense in sensor on the water surface, but through the water surface after a period of time
After fluctuation, the position of sensor changes at random, carries out re-moving deployment to sensor using AIFS algorithm at this time, can achieve
Cover the effect of more waters surface.
The method is applied to during home for destitute promptly searches,
Such as: it needs at a time to carry out emergency searching to old man in home for destitute, in traditional indoor positioning algorithms
In the case where not can guarantee accurate search rate, the sensor of signal projector with old man can will be perceived in certain distance
It is placed in nurse, when executing flash request, by AIFS algorithm, can carry out providing the random shield in commands direct position
Work is moved, as much as possible covering home for destitute region, to search the locating region of target old man.
By above-mentioned experimentation it is found that the present invention improves the shoal of fish and jump out local optimum by increased jump behavior
Speed and efficiency；In conjunction with K neighboring mean value thought, new food concentration determination method is set, reduces and falls into the general of local optimum
Rate；Add rebuffed behavior, the perfect critical processing to coverage situation at network boundary；By decay factor θ, adaptively
Step-length and the visual field for adjusting fish, ensure convergent stability.The present invention solves the shoal of fish under the premise of guaranteeing the network coverage
The problem of algorithm is to severe local optimum delay of response, while reducing the probability for falling into local optimum, it is perfect to network edge
The critical processing of coverage situation, has ensured convergent stability at boundary.
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all in spirit of the invention and
Within principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (9)
1. a kind of optimization method of adaptive impovement fish-swarm algorithm, which is characterized in that the described method includes:
Assuming that shedding N number of isomorphism sensor node S={ S at random in detected region I1(x1,y1),S2(x2,y2),..Sk
(xk,yk),...,SN(xN,yN), wherein k=1,2,3 ..., N；Node is all made of the Boolean sense model that radius is r；
For ease of description, turning to l × l pixel collection M={ M for region I is discrete simultaneously(1,1),M(1,2),...,
M(x,y),...,M(l,l), wherein x, y=1,2,3 ..., N, and the node for setting random placement all falls within some pixel in I
Point on；
Step 1: each node Sk5 behaviors of looking for food, bunch, repel, jump and be rebuffed are executed respectively, and record behavior optimizing
Position afterwards；
Step 2: calculating network coverage Y after front-wheel optimizing；
Step 3: if gained network coverage Y is improved relative to before front-wheel optimizing after front-wheel optimizing, node set S
It is moved according to the location point of record；Otherwise according to whether reaching local optimal searching upper limit Cth, judge whether to need the visual field of decaying
Visual and step-length step；
Wherein, step indicates that fish executes the step-length once moved, and step meets step=step0× rand, rand indicate (0,1)
Between a random number；step0Indicate fish execute a behavior it is mobile apart from radix；
Step 4: judging Y > YthOr visual < VthIt is whether true, algorithm is exited if set up；If not, return step 2；Its
Middle YthAnd VthIndicate the end threshold value of visual and Y.
2. the method according to claim 1, wherein the jump behavior is defined as follows:
In node SkVisual in, appoint take some pixel M(x,y)；
If node SkMeet: Densityk> QthAnd pk≥Pth, then SkJump behavior is executed, pixel ArgMax is jumped to
(T(x,y))；
Wherein, pixel M(x,y)MeetIf | ArgMax (T(x,y)) | > 1, then SkRandom skip is extremely
ArgMax(T(x,y)) any of pixel；ArgMax(T(x,y)) indicate T(x,y)Obtain corresponding (x, y) coordinate when maximum value；
Definition node SkThe sum of coverage rate of pixel is node S in sensing range rkCrowding Densityk:
M(e,f)MeetWhereinExpression is rounded downwards r；
Food concentration T(x,y)Is defined as: pixel M(x,y)With parameter K, then pixel M(x,y)Food concentration T(x,y)With
M(x,y)K neighborhood in pixel coverage rate it is related:
K neighboring mean value method is applied to T(x,y)In calculating, realize M(x,y)The coverage condition of pixel in surrounding K neighborhood,
It is included in T(x,y)Calculating scope；By setting different weight coefficients for each etale neighborhoodCharacterize M(x,y)Each layer in neighborhood
Between influence difference.
3. according to the method described in claim 2, it is characterized in that, the rebuffed behavior is defined as follows:
If node SkVertical range d between the boundary I of region meetsThen node SkIt is mobile towards boundary opposite direction, and
Moving distance is rand × (r-d)；
Update visual and step: given CthAfter the optimizing of wheel, if convergence increment deficiency occurs in Y, i.e. Y is after excessively wheel optimizing
Do not meet increment index, then update visual using decay factor θ and reduces the visual field and step-length；More new formula is as follows:
Visual=visual × θ (3)
Step=step × θ (4)
Wherein, decay factor θ=0.8.
4. the method according to claim 1, wherein the network coverage Y is defined as:
Wherein, as pixel M(x,y)With any SkThe distance between be less than r when, then it is assumed that M(x,y)It can be covered by S, i.e. network S exists
Point M(x,y)Coverage rate are as follows:
Wherein, as pixel M(x,y)With sensor node SkThe distance between be less than r when, then it is assumed that SkIn M(x,y)Perception rate be
1, it is otherwise 0, it may be assumed that
5. according to the method described in claim 3, it is characterized in that, it is described look for food, bunch, repulsion behavior is defined respectively as:
Foraging behavior: in node SkVisual in, appoint take some pixel M(x,y)If M(x,y)MeetThen think point M(x,y)Food concentration be higher than SkThe food concentration of position,
SkTo M(x,y)Mobile step；If given choose under constraint, fail to find point M(x,y), then current fish SkStart random walk；
Wherein, D () indicates the Euclidean distance between two elements；
Bunch behavior: if node SkMeetWhen, SkTo the mobile step in current nearest fish direction；
Repulsion behavior: if node SkMeetWhen, SkTo the current mobile step of fish opposite direction recently；
Wherein, α and β is the distance between node when sensing range boundary that α value is two nodes is tangent.
6. the method according to claim 1, wherein before the step 1, further includes:
Parameters are initialized.
7. the method according to claim 1, which is characterized in that the method is applied to wireless sensor network
In covering performance optimization.
8. the method according to the description of claim 7 is characterized in that the wireless sensor network covering performance that is applied to optimizes
In be specially be applied to lake blue algae detection in.
9. the method according to the description of claim 7 is characterized in that the wireless sensor network covering performance that is applied to optimizes
In be specially to be applied to during home for destitute promptly searches.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810685277.0A CN108966241B (en) | 2018-06-28 | 2018-06-28 | Optimization method for self-adaptively improving fish swarm algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810685277.0A CN108966241B (en) | 2018-06-28 | 2018-06-28 | Optimization method for self-adaptively improving fish swarm algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108966241A true CN108966241A (en) | 2018-12-07 |
| CN108966241B CN108966241B (en) | 2020-09-25 |
Family
ID=64487366
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810685277.0A Active CN108966241B (en) | 2018-06-28 | 2018-06-28 | Optimization method for self-adaptively improving fish swarm algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108966241B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112446457A (en) * | 2020-12-02 | 2021-03-05 | 中国计量大学 | Gas leakage source positioning method based on improved artificial fish swarm algorithm |
| CN115048274A (en) * | 2022-08-11 | 2022-09-13 | 中国通信建设第三工程局有限公司 | Operation and maintenance system based on big data |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106845623A (en) * | 2016-12-13 | 2017-06-13 | 国网冀北电力有限公司信息通信分公司 | A kind of electric power wireless private network base station planning method based on artificial fish-swarm algorithm |
| CN107103355A (en) * | 2017-04-21 | 2017-08-29 | 北京工业大学 | A kind of intelligent vehicle SLAM data correlation methods based on improvement artificial fish-swarm algorithm |
-
2018
- 2018-06-28 CN CN201810685277.0A patent/CN108966241B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106845623A (en) * | 2016-12-13 | 2017-06-13 | 国网冀北电力有限公司信息通信分公司 | A kind of electric power wireless private network base station planning method based on artificial fish-swarm algorithm |
| CN107103355A (en) * | 2017-04-21 | 2017-08-29 | 北京工业大学 | A kind of intelligent vehicle SLAM data correlation methods based on improvement artificial fish-swarm algorithm |
Non-Patent Citations (3)
| Title |
|---|
| JUNBO XIA: "《2016 International Conference on Smart City and Systems Engineering (ICSCSE)》", 31 December 2016 * |
| 刘彦君: "鱼群算法及在无线传感器网络覆盖优化中的应用", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 李志武: "人工鱼群算法的改进及在无线传感器网络覆盖优化的应用", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112446457A (en) * | 2020-12-02 | 2021-03-05 | 中国计量大学 | Gas leakage source positioning method based on improved artificial fish swarm algorithm |
| CN112446457B (en) * | 2020-12-02 | 2023-07-18 | 中国计量大学 | Gas leakage source positioning method based on improved artificial fish swarm algorithm |
| CN115048274A (en) * | 2022-08-11 | 2022-09-13 | 中国通信建设第三工程局有限公司 | Operation and maintenance system based on big data |
| CN115048274B (en) * | 2022-08-11 | 2022-11-18 | 中国通信建设第三工程局有限公司 | Operation and maintenance system based on big data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108966241B (en) | 2020-09-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| With et al. | Dispersal success on fractal landscapes: a consequence of lacunarity thresholds | |
| CN102279973B (en) | Sea-sky-line detection method based on high gradient key points | |
| CN108966241A (en) | A kind of optimization method of adaptive impovement fish-swarm algorithm | |
| CN109492561A (en) | A kind of remote sensing image Ship Detection based on improved YOLO V2 model | |
| CN1932850A (en) | Remoto sensing image space shape characteristics extracting and sorting method | |
| Lagona et al. | A hidden Markov model for the analysis of cylindrical time series | |
| CN107203761B (en) | Road width estimation method based on high-resolution satellite image | |
| CN104778477A (en) | Optimized ABC (artificial bee colony) clustering method | |
| CN110751172B (en) | Method and system for deducing category of pathological full-film image for weak supervision learning | |
| CN104080154B (en) | Industrial wireless sensor network node dormancy method based on cellular automaton | |
| CN107918953A (en) | The extracting method and device of laser scanning power line point cloud based on three dimensions | |
| Joshi et al. | Hybrid DSSCS and convolutional neural network for peripheral blood cell recognition system | |
| CN108337685A (en) | Wireless sensor network data fusion method based on sub-clustering DGM | |
| Yang et al. | Research on subway pedestrian detection algorithms based on SSD model | |
| Xiao | A clustering algorithm based on artificial fish school | |
| CN105574528A (en) | Synechia cell image segmenting method based on polyphase mutual exclusion level set | |
| CN108876062A (en) | A kind of big data method and device of crime dramas intelligent predicting | |
| Jiang et al. | An improved ant colony clustering algorithm | |
| Rodríguez‐Tricot et al. | From Hutchinsonian ratios to spatial scaling theory: the interplay among limiting similarity, body size and landscape structure | |
| Lobón‐Cerviá et al. | Two‐phase self‐thinning in stream‐living juveniles of lake‐migratory brown trout Salmo trutta L. Compatibility between linear and non‐linear patterns across populations? | |
| Essian et al. | Predicting effects of water management on breeding abundance of three wading bird species | |
| Tang et al. | Intrusive tumor growth inspired optimization algorithm for data clustering | |
| CN117041996A (en) | Secondary deployment method for mobile node of hybrid sensor network | |
| Park et al. | Reliable and flexible detection of large-scale phenomena on wireless sensor networks | |
| Dogra et al. | Identifying indicators of sustainable smart agriculture driven by big data using modified total interpretive structural modeling (mtism) |
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 |