CN106488482B - Wireless sensor network optimizing method based on multi-Agent evolutionary Algorithm - Google Patents
Wireless sensor network optimizing method based on multi-Agent evolutionary Algorithm Download PDFInfo
- Publication number
- CN106488482B CN106488482B CN201610817374.1A CN201610817374A CN106488482B CN 106488482 B CN106488482 B CN 106488482B CN 201610817374 A CN201610817374 A CN 201610817374A CN 106488482 B CN106488482 B CN 106488482B
- Authority
- CN
- China
- Prior art keywords
- wireless sensor
- intelligent body
- sensor network
- energy
- wireless
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- 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)
- Mobile Radio Communication Systems (AREA)
- Arrangements For Transmission Of Measured Signals (AREA)
Abstract
The present invention discloses a kind of wireless sensor network optimizing method based on multi-Agent evolutionary Algorithm, and its step are as follows: (1) constructing wireless sensor network；(2) maximum number of iterations is set；(3) Agent Grid is constructed；(4) using the Monte Carlo method intelligent computing agent energy with punishment；(5) neighborhood contention operation；(6) mutation operation；(7) local optimum intelligent body is determined；(8) self study operates；(9) fictitious force operates；(10) judge whether cyclic algebra reaches maximum number of iterations, if so, executing step (11), otherwise, cyclic algebra adds 1, executes step (3)；(11) wireless sensor network optimum results are exported.The present invention accelerates the speed of optimization wireless sensor network, reduces the interference between wireless sensor, is suitable for optimizing large-scale wireless sensor network.
Description
Technical field
The invention belongs to fields of communication technology, further relate to one of wireless communication technology field and are based on mostly intelligently
The wireless sensor network optimizing method of body evolution algorithm.The present invention can be used for optimizing wireless sensor in wireless sensor network
Distribution, be monitored with region of the wireless sensor to certain scale, make wireless sensor to monitoring region coverage area
Reach maximum.
Background technique
Wireless sensor network is a kind of wireless network that can effectively obtain information.Wireless sensor network is extensive
Apply in military, social production and life, such as battlefield supervision, disaster relief, target identification, weather monitoring, disaster are done
In advance, medical treatment and nursing etc..Wireless sensor network generally refers to for the wireless sensor of certain amount to be distributed in a scale
Biggish region is monitored.Therefore, how to go to the position for being laid out these wireless sensors that coverage area is made to reach maximum is one
A critically important technical problem.When wireless sensor number is less, can artificially be placed, when wireless sensor number compared with
When big, it is necessary to consider how reasonably to place wireless sensor.When wireless sensor placement is unreasonable, between wireless sensor
It will cause interference, so that information transmission is unreliable.Traditional optimization method be only applicable to wireless sensor number it is less when, work as nothing
When line number of sensors is more, calculating speed is slow, and time-consuming.Wireless sensor network optimization based on multi-Agent evolutionary Algorithm
Method can not only reduce cost, accelerate calculating speed, moreover it is possible to enhance the stability of wireless sensor network.
Paper " the An Efficient Genetic Algorithm for that Yourim Yoon et al. is delivered at it
Maximum Coverage Development in Wireless Sensor Networks”(《IEEE Transactions
On Cybernetics " article number: 2168-2267 (2013)) in disclose a kind of wireless biography based on effective genetic algorithm
Sensor network coverage optimization method.Method for normalizing is added on the basis of genetic algorithm in this method, scans for solution space,
With in population individual string come to wireless sensor position and covering radius encode, by the area coverage of wireless sensor
Fitness function is solved with Monte Carlo method as fitness function, by successive ignition, retains in population and fits
The maximum individual of response functional value stops until meeting termination condition.Shortcoming existing for this method is to work as wireless sensor
When number is more, speed of searching optimization is slow, and time-consuming, and search space is larger, and convergence rate is slow, and the number of iterations is more.
Institutes Of Technology Of Nanjing its application patent " based on the wireless sense network coverage optimization method of fictitious force algorithm be
It discloses and a kind of is calculated based on fictitious force in system " (application number: 201410579915.2, application publication number: 104333866 A of CN)
The wireless sense network coverage optimization method and system of method.Method includes the following steps: the monitoring section of 1. setting wireless sense networks
The detection range of domain range, sensor；2. wireless sensor dispensing in monitoring regional scope；3. determining wireless sensor node
The coordinate value of point；4. calculating distance between each node, the stress condition that each node is calculated to node coordinate matrix is stored；5. sentencing
Whether disconnected node motion meets constraint condition: if it is exported using current node coordinate matrix as node location data,
Otherwise enter next step, judge whether the distance between the boundary of node location and monitoring regional scope is more than distance threshold:
If be less than, make normal movement after joints, if it does, node is then made to stop the boundary side to monitoring regional scope
To movement, spring back second distance threshold value, then return step 3.Shortcoming existing for this method is that calculating speed is slow, consumption
Duration, wireless sensor are smaller to the coverage area in monitoring region, and the covering overlapping between wireless sensor is larger, make wireless sensing
It is interfered between device, information transmission is unreliable.
Summary of the invention
It is an object of the invention to overcoming the shortcomings of above-mentioned prior art, provide a kind of based on multi-Agent evolutionary Algorithm
Wireless sensor network optimizing method, so that wireless sensor reaches maximum to the coverage area in monitoring region.
The present invention comprises the following steps that
(1) wireless sensor network is constructed:
The working region range of (1a) input wireless sensor network；
(1b) inputs the covering radius of three kinds of wireless sensors；
Three kinds of wireless sensors are randomly dispersed among the working region of wireless sensor network by (1c), complete a nothing
The building of line sensor network；
(2) maximum number of iterations is set:
In the range of [0,500], according to the optimum results of multi-Agent evolutionary Algorithm, multi-Agent evolutionary Algorithm is set
Maximum number of iterations；
(3) Agent Grid is constructed:
Using a wireless sensor network as an intelligent body, being built into size with 81 intelligent bodies is 9 × 9 intelligent bodies
Grid；
(4) using the Monte Carlo method with punishment, each intelligent body in the Agent Grid that size is 9 × 9 is calculated
Energy；
(5) Agent Grid for being 9 × 9 to size executes neighborhood contention operation；
(6) Agent Grid for being 9 × 9 to size executes mutation operation；
(7) determine that size is the local optimum intelligent body of 9 × 9 Agent Grid:
Using the Monte Carlo method with punishment, size is each intelligent body in 9 × 9 Agent Grids after calculating variation
Energy, using the maximum intelligent body of energy as 9 × 9 local optimum intelligent bodies；
(8) self study operates:
(8a) according to the optimum results of multi-Agent evolutionary Algorithm, is arranged what self study operated in the range of [0,100]
Maximum number of iterations；
(8b) using each wireless sensor network as an intelligent body, constructing a size with 25 intelligent bodies is 5
× 5 Agent Grid；
(8c) calculates each intelligent body in the Agent Grid that size is 5 × 5 using the Monte Carlo method with punishment
Energy；
The Agent Grid that (8d) is 5 × 5 to size executes neighborhood contention operation；
The Agent Grid that (8e) is 5 × 5 to size executes mutation operation；
(8f) calculates the energy of each intelligent body in the Agent Grid after making a variation using the Monte Carlo method with punishment
Amount, finds out the maximum intelligent body of energy as 5 × 5 local optimum intelligent bodies；
The energy of 9 × 9 local optimum intelligent bodies is compared by (8g) with the energy of 5 × 5 local optimum intelligent bodies, if 5
When the energy of × 5 local optimum intelligent bodies is greater than the energy of 9 × 9 local optimum intelligent bodies, more with 5 × 5 local optimum intelligent bodies
New 9 × 9 local optimum intelligent body；
(8h) is using updated 9 × 9 local optimum intelligent body as optimal wireless sensor network；
(8i) judges whether the cyclic algebra of current self study operation reaches maximum number of iterations, if so, thening follow the steps
(9), it after the cyclic algebra that self study operates otherwise, is added 1, executes step (8d)；
(9) fictitious force operates:
(9a) arbitrarily chooses a wireless sensor from optimal wireless sensor network；
(9b) according to the following formula, it is wireless with each in addition to selected wireless sensor respectively to calculate selected wireless sensor
Distance between sensor:
Wherein, dmIndicate m-th of wireless sensor in addition to m-th of wireless sensor n-th of wireless sensor it
Between distance, xmAnd ymRespectively indicate the cross, ordinate position of m-th of wireless sensor, xnAnd ynIt respectively indicates in addition to m-th
The cross of n-th of wireless sensor, ordinate position except wireless sensor；
(9c) according to the following formula, it is wireless with each in addition to selected wireless sensor respectively to calculate selected wireless sensor
The sum of covering radius of sensor:
Lm=Rm+Rn
Wherein, LmIndicate m-th of wireless sensor and n-th of wireless sensor in addition to m-th of wireless sensor
The sum of covering radius, RmIndicate the covering radius of m wireless sensor, RnIndicate n-th in addition to m-th of wireless sensor
The covering radius of wireless sensor；
(9d) judges the distance d between wireless sensormWhether covering radius the sum of L is less thanm, if so, to m-th of wireless biography
Sensor applies repulsive force, otherwise, applies attraction to m-th of wireless sensor；
(9e) judges whether each wireless sensor has been selected in optimal wireless sensor network, if so, executing
Step (10) otherwise executes step (9a)；
(10) judge whether the cyclic algebra of current multi-Agent evolutionary Algorithm reaches maximum number of iterations, if so, holding
Row step (11) otherwise after the cyclic algebra of multi-Agent evolutionary Algorithm is added 1, executes step (3)；
(11) wireless sensor network distribution results are exported.
The invention has the following advantages over the prior art:
First, since the present invention uses the Monte Carlo method with punishment, each intelligent body in computational intelligence volume mesh
Energy reduces the number of iterations of searching process, has quickly found optimal wireless sensor network, has overcome in the prior art
Conventional method convergence rate is slow, the disadvantage that time-consuming, the number of iterations is more.So that the present invention accelerates optimization wireless sensor
The convergence rate of network shortens the time for finding optimal wireless sensor network.
Second, since the present invention executes neighborhood contention operation to Agent Grid, reduce wireless sensor network optimizing
Search space, accelerate the searching process of wireless sensor network, overcome conventional method search space in the prior art
Greatly, computationally intensive disadvantage.So that the present invention reduces search space, the calculation amount of searching process is greatly reduced.
Third is suitable for large-scale wireless sensor network since the present invention executes self study operation to Agent Grid
Network optimization problem, time-consuming too long when overcoming conventional method in the prior art to solving the problems, such as extensive, speed is too slow to be lacked
Point.When so that the present invention solving large-scale wireless sensor network optimization problem, optimal solution can be quickly found out by appointing.
4th, since the present invention is operated using fictitious force, the area coverage of wireless sensor network is increased, nothing is reduced
Overlapping coverage between line sensor overcomes conventional method wireless sensor in the prior art to the covering model in monitoring region
Enclose smaller, covering between wireless sensor overlapping is larger, makes to interfere between wireless sensor, and information transmission is insecure to ask
Topic.So that increasing the reliability of information transmission The present invention reduces the interference between wireless sensor.
Detailed description of the invention
Fig. 1 is flow chart of the invention；
Fig. 2 is analogous diagram of the invention.
Specific embodiment
The present invention will be further described with reference to the accompanying drawing.
Referring to attached drawing 1, the specific steps of the present invention are as follows.
Step 1, wireless sensor network is constructed.
Input the working region range of wireless sensor network.
The working region range of the wireless sensor network is 100 × 100 meters.
Input the covering radius of three kinds of wireless sensors.
The covering radius of the wireless sensor is respectively r1, r2, r3, wherein and r3=0.8 × r2 meters, r2=0.8 ×
R1 meters.
Three kinds of wireless sensors are randomly dispersed among the working region of wireless sensor network, a wireless biography is completed
The building of sensor network.
Step 2, maximum number of iterations is set.
In the range of [0,500], according to the optimum results of multi-Agent evolutionary Algorithm, multi-Agent evolutionary Algorithm is set
Maximum number of iterations, set 300 for the maximum number of iterations of multi-Agent evolutionary Algorithm in the embodiment of the present invention.
Step 3, Agent Grid is constructed.
Using a wireless sensor network as an intelligent body, being built into size with 81 intelligent bodies is 9 × 9 intelligent bodies
Grid.
Step 4, using the Monte Carlo method with punishment, calculating size is each intelligent body in 9 × 9 Agent Grids
Energy.
Specific step is as follows for the Monte Carlo method with punishment:
The first step calculates the area coverage of each wireless sensor network corresponding with each intelligent body according to the following formula:
Wherein, SkIndicate the area coverage of p-th of wireless sensor network corresponding with k-th of intelligent body, the number of k and p
It is worth equal, N indicates population of the uniformly dispersing in working region, and q indicates to fall in the work covered by wireless sensor network
Population in region, A indicate the area of the working region of wireless sensor network.
Second step, according to the following formula, calculating has overlapping in p-th of wireless sensor network corresponding with k-th of intelligent body
Penalty value between all wireless sensors of covering:
φk(C)=λ × ∑ S'(Ci,Cj)
Wherein, φk(C) indicate that there is overlapping covering in p-th of wireless sensor network corresponding with k-th of intelligent body
Penalty value between all wireless sensors, λ indicate that the penalty factor that numerical value is greater than zero, ∑ indicate sum operation, S'(Ci,Cj) table
Show i-th of wireless sensor C in wireless sensor networkiWith j-th of wireless sensor CjOverlapping coverage, work as wireless sensing
Device CiWith CjWhen overlapping covering, S'(Ci,Cj) value 1, as wireless sensor CiWith CjWhen non-overlapping covering, S'(Ci,Cj) value 0.
Third step, according to the following formula, the energy of each intelligent body in computational intelligence volume mesh:
Energyk=Sk+φk(C)
Wherein, EnergykIndicate the energy of k-th of intelligent body in Agent Grid, SkIt indicates corresponding with k-th of intelligent body
P-th of wireless sensor network area coverage, φk(C) p-th of wireless sensor corresponding with k-th of intelligent body is indicated
There is the penalty value between all wireless sensors of overlapping covering in network.
Step 5, the Agent Grid for being 9 × 9 to size executes neighborhood contention operation.
Specific step is as follows for the neighborhood contention operation:
The first step, an optional intelligent body are found out in four neighborhoods from four neighborhoods up and down of selected intelligent body
The maximum intelligent body of energy.
The energy of energy maximum intelligent body is compared by second step with the energy of selected intelligent body, if energy most pansophy
When the energy of energy body is greater than the energy of selected intelligent body, selected intelligent body is updated with the maximum intelligent body of energy, after obtaining update
Intelligent body.
Step 6, the Agent Grid for being 9 × 9 to size executes mutation operation.
The mutation operation refers to, the random perturbation that one meets Gaussian Profile is added on updated intelligent body,
Intelligent body after being made a variation.
Step 7, determine that size is the local optimum intelligent body of 9 × 9 Agent Grid.
Using the Monte Carlo method with punishment, size is each intelligent body in 9 × 9 Agent Grid after calculating variation
Energy, find out the maximum intelligent body of energy as 9 × 9 local optimum intelligent bodies.
Specific step is as follows for the Monte Carlo method with punishment:
The first step calculates the area coverage of each wireless sensor network corresponding with each intelligent body according to the following formula:
Wherein, SkIndicate the area coverage of p-th of wireless sensor network corresponding with k-th of intelligent body, the number of k and p
It is worth equal, N indicates population of the uniformly dispersing in working region, and q indicates to fall in the work covered by wireless sensor network
Population in region, A indicate the area of the working region of wireless sensor network；
Second step, according to the following formula, calculating has overlapping in p-th of wireless sensor network corresponding with k-th of intelligent body
Penalty value between all wireless sensors of covering:
φk(C)=λ × ∑ S'(Ci,Cj)
Wherein, φk(C) indicate that there is overlapping covering in p-th of wireless sensor network corresponding with k-th of intelligent body
Penalty value between all wireless sensors, λ indicate that the penalty factor that numerical value is greater than zero, ∑ indicate sum operation, S'(Ci,Cj) table
Show i-th of wireless sensor C in wireless sensor networkiWith j-th of wireless sensor CjOverlapping coverage, work as wireless sensing
Device CiWith CjWhen overlapping covering, S'(Ci,Cj) value 1, as wireless sensor CiWith CjWhen non-overlapping covering, S'(Ci,Cj) value 0；
Third step, according to the following formula, the energy of each intelligent body in computational intelligence volume mesh:
Energyk=Sk+φk(C)
Wherein, EnergykIndicate the energy of k-th of intelligent body in Agent Grid, SkIt indicates corresponding with k-th of intelligent body
P-th of wireless sensor network area coverage, φk(C) p-th of wireless sensor corresponding with k-th of intelligent body is indicated
There is the penalty value between all wireless sensors of overlapping covering in network.
Step 8, self study operates.
(8a) according to the optimum results of multi-Agent evolutionary Algorithm, is arranged what self study operated in the range of [0,100]
Maximum number of iterations, the maximum number of iterations in the embodiment of the present invention by self study operation are set as 20；
(8b) using each wireless sensor network as an intelligent body, constructing a size with 25 intelligent bodies is 5
× 5 Agent Grid.
(8c) calculates each intelligent body in the Agent Grid that size is 5 × 5 using the Monte Carlo method with punishment
Energy.
Specific step is as follows for the Monte Carlo method with punishment:
The first step calculates the area coverage of each wireless sensor network corresponding with each intelligent body according to the following formula:
Wherein, SkIndicate the area coverage of p-th of wireless sensor network corresponding with k-th of intelligent body, the number of k and p
It is worth equal, N indicates population of the uniformly dispersing in working region, and q indicates to fall in the work covered by wireless sensor network
Population in region, A indicate the area of the working region of wireless sensor network；
Second step, according to the following formula, calculating has overlapping in p-th of wireless sensor network corresponding with k-th of intelligent body
Penalty value between all wireless sensors of covering:
φk(C)=λ × ∑ S'(Ci,Cj)
Wherein, φk(C) indicate that there is overlapping covering in p-th of wireless sensor network corresponding with k-th of intelligent body
Penalty value between all wireless sensors, λ indicate that the penalty factor that numerical value is greater than zero, ∑ indicate sum operation, S'(Ci,Cj) table
Show i-th of wireless sensor C in wireless sensor networkiWith j-th of wireless sensor CjOverlapping coverage, work as wireless sensing
Device CiWith CjWhen overlapping covering, S'(Ci,Cj) value 1, as wireless sensor CiWith CjWhen non-overlapping covering, S'(Ci,Cj) value 0；
Third step, according to the following formula, the energy of each intelligent body in computational intelligence volume mesh:
Energyk=Sk+φk(C)
Wherein, EnergykIndicate the energy of k-th of intelligent body in Agent Grid, SkIt indicates corresponding with k-th of intelligent body
P-th of wireless sensor network area coverage, φk(C) p-th of wireless sensor corresponding with k-th of intelligent body is indicated
There is the penalty value between all wireless sensors of overlapping covering in network.
The Agent Grid that (8d) is 5 × 5 to size executes neighborhood contention operation.
Specific step is as follows for the neighborhood contention operation:
The first step, an optional intelligent body are found out in four neighborhoods from four neighborhoods up and down of selected intelligent body
The maximum intelligent body of energy.
The energy of energy maximum intelligent body is compared by second step with the energy of selected intelligent body, if energy most pansophy
When the energy of energy body is greater than the energy of selected intelligent body, selected intelligent body is updated with the maximum intelligent body of energy, after obtaining update
Intelligent body.
The Agent Grid that (8e) is 5 × 5 to size executes mutation operation.
The mutation operation refers to, the random perturbation that one meets Gaussian Profile is added on updated intelligent body,
Intelligent body after being made a variation.
(8f) calculates the energy of each intelligent body in the Agent Grid after making a variation using the Monte Carlo method with punishment
Amount, finds out the maximum intelligent body of energy as 5 × local optimum intelligent body.
Specific step is as follows for the Monte Carlo method with punishment:
The first step calculates the area coverage of each wireless sensor network corresponding with each intelligent body according to the following formula:
Wherein, SkIndicate the area coverage of p-th of wireless sensor network corresponding with k-th of intelligent body, the number of k and p
It is worth equal, N indicates population of the uniformly dispersing in working region, and q indicates to fall in the work covered by wireless sensor network
Population in region, A indicate the area of the working region of wireless sensor network；
Second step, according to the following formula, calculating has overlapping in p-th of wireless sensor network corresponding with k-th of intelligent body
Penalty value between all wireless sensors of covering:
φk(C)=λ × ∑ S'(Ci,Cj)
Wherein, φk(C) indicate that there is overlapping covering in p-th of wireless sensor network corresponding with k-th of intelligent body
Penalty value between all wireless sensors, λ indicate that the penalty factor that numerical value is greater than zero, ∑ indicate sum operation, S'(Ci,Cj) table
Show i-th of wireless sensor C in wireless sensor networkiWith j-th of wireless sensor CjOverlapping coverage, work as wireless sensing
Device CiWith CjWhen overlapping covering, S'(Ci,Cj) value 1, as wireless sensor CiWith CjWhen non-overlapping covering, S'(Ci,Cj) value 0；
Third step, according to the following formula, the energy of each intelligent body in computational intelligence volume mesh:
Energyk=Sk+φk(C)
Wherein, EnergykIndicate the energy of k-th of intelligent body in Agent Grid, SkIt indicates corresponding with k-th of intelligent body
P-th of wireless sensor network area coverage, φk(C) p-th of wireless sensor corresponding with k-th of intelligent body is indicated
There is the penalty value between all wireless sensors of overlapping covering in network.
The energy of 9 × 9 local optimum intelligent bodies is compared by (8g) with the energy of 5 × 5 local optimum intelligent bodies, if 5
When the energy of × 5 local optimum intelligent bodies is greater than the energy of 9 × 9 local optimum intelligent bodies, more with 5 × 5 local optimum intelligent bodies
New 9 × 9 local optimum intelligent body.
(8h) is using updated 9 × 9 local optimum intelligent body as optimal wireless sensor network.
(8i) judges whether the cyclic algebra of current self study operation reaches maximum number of iterations, if so, thening follow the steps
9, otherwise, after the cyclic algebra that self study operates is added 1, execute step (8d)；
Step 9, fictitious force operates.
(9a) arbitrarily chooses a wireless sensor from optimal wireless sensor network.
(9b) according to the following formula, it is wireless with each in addition to selected wireless sensor respectively to calculate selected wireless sensor
Distance between sensor:
Wherein, dmIndicate m-th of wireless sensor in addition to m-th of wireless sensor n-th of wireless sensor it
Between distance, xmAnd ymRespectively indicate the cross, ordinate position of m-th of wireless sensor, xnAnd ynIt respectively indicates in addition to m-th
The cross of n-th of wireless sensor, ordinate position except wireless sensor；
(9c) according to the following formula, it is wireless with each in addition to selected wireless sensor respectively to calculate selected wireless sensor
The sum of covering radius of sensor:
Lm=Rm+Rn
Wherein, LmIndicate m-th of wireless sensor and n-th of wireless sensor in addition to m-th of wireless sensor
The sum of covering radius, RmIndicate the covering radius of m wireless sensor, RnIndicate n-th in addition to m-th of wireless sensor
The covering radius of wireless sensor；
(9d) judges the distance d between wireless sensormWhether covering radius the sum of L is less thanm, if so, to m-th of wireless biography
Sensor applies repulsive force, otherwise, applies attraction to m-th of wireless sensor.
Specific step is as follows for the application repulsive force:
The first step calculates the repulsive force applied to wireless sensor according to the following formula:
Wherein,It indicates to m-th of wireless sensor CmThe repulsive force of application, RmIndicate m-th of wireless sensor CmCover
Lid radius, RnIndicate n-th of wireless sensor CnCovering radius, d (Cm,Cn) indicate m-th of wireless sensor CmWith n-th of nothing
Line sensor CnDistance；
Second step, by the size and Orientation of repulsive force, the wireless sensor of mobile repulsive force to be applied.
Specific step is as follows for the application attraction:
The first step calculates the attraction applied to wireless sensor according to the following formula:
Wherein,It indicates to m-th of wireless sensor CmThe attraction of application, RmIndicate m-th of wireless sensor CmCover
Lid radius, RnIndicate n-th of wireless sensor CnCovering radius, d (Cm,Cn) indicate m-th of wireless sensor CmWith n-th of nothing
Line sensor CnDistance；
Second step, by the size and Orientation of attraction, the wireless sensor of mobile attraction to be applied.
(9e) judges whether each wireless sensor has been selected in optimal wireless sensor network, if so, executing
Step 10, otherwise, step (9a) is executed；
Step 10, judge whether the cyclic algebra of current multi-Agent evolutionary Algorithm reaches maximum number of iterations, if so,
Step 11 is executed, otherwise, after the cyclic algebra of multi-Agent evolutionary Algorithm is added 1, executes step 3；
Step 11, wireless sensor network distribution results are exported.
Effect of the invention is further described below with reference to analogous diagram.
1. simulated conditions:
Emulation experiment of the invention is to be configured to Intel Core (TM) i5-3210M CPU@in computer hardware
2.50GHz, the hardware environment of 4.00GB RAM and computer software are configured under the software environment of Visual Studio 2010
It carries out.
2. emulation content:
Table 1 is the data set that uses of the present invention, and r1, r2, r3 respectively indicate the covering radius of three kinds of wireless sensors, n1,
N2, n3 respectively indicate the number of three kinds of wireless sensors, and N indicates the sum of three kinds of wireless sensors.
Emulation experiment of the invention is that 15 groups of data are respectively adopted with method optimization wireless sensor network of the invention.?
On the basis of optimizing wireless sensor network 30 times to every group of data using method of the invention, then calculate separately every group of data 30
Mean value, standard deviation and the average simulation time of wireless sensor network area coverage after secondary emulation.
Fig. 2 is analogous diagram of the invention, and wherein Fig. 2 (a) is the knot emulated using S1-0.7 group data in data set
Fruit figure, Fig. 2 (b) are the result figures emulated using S2-0.7 group data in data set, and Fig. 2 (c) is using S3- in data set
The result figure that 0.7 group of data is emulated, Fig. 2 (d) are the result figure emulated using S4-0.7 group data in data set, figure
2 (e) be the result figure emulated using S5-0.7 group data in data set, and Fig. 2 (f) is using S1-0.8 group number in data set
According to the result figure emulated, Fig. 2 (g) is the result figure emulated using S2-0.8 group data in data set, and Fig. 2 (h) is
Using the result figure that S3-0.8 group data are emulated in data set, Fig. 2 (i) is carried out using S4-0.8 group data in data set
The result figure of emulation, Fig. 2 (j) are the result figures emulated using S5-0.8 group data in data set, and Fig. 2 (k) is using number
According to the result figure for concentrating S1-0.9 group data to be emulated, Fig. 2 (l) is emulated using S2-0.9 group data in data set
Result figure, Fig. 2 (m) are the result figures emulated using S3-0.9 group data in data set, and Fig. 2 (n) is using in data set
The result figure that S4-0.9 group data are emulated, Fig. 2 (o) are the results emulated using S5-0.9 group data in data set
Figure.
The catalog data that 1 present invention emulation of table uses
3. analysis of simulation result:
Table 2 is the present invention using 15 groups of data, the wireless sensor network that every group is calculated after data simulation 30 times
Mean value, standard deviation and the average simulation time of area coverage.
2 simulation result list of table
By table 2 it can be seen that, the mean value of every group of calculated wireless sensor network area coverage after data simulation 30 times
It is close with work area, illustrate that method of the invention keeps wireless sensor network larger to the coverage area of working region,
Monitoring of the wireless sensor network to working region is realized well.
By table 2 it can be seen that, the standard of every group of calculated wireless sensor network area coverage after data simulation 30 times
Difference is smaller, illustrates that the method for the invention stability when optimizing wireless sensor network is high, optimization performance is good.
By table 2 it can be seen that, the average time of every group of calculated optimization wireless sensor network after data simulation 30 times
It is short, illustrate that method of the invention improves the speed of optimization wireless sensor network, to the large-scale wireless sensor network of optimization
Network is high-efficient.For the wireless sensor network of different scales, method of the invention can quickly and effectively realize optimization nothing
The requirement of line sensor network.
Claims (5)
1. a kind of wireless sensor network optimizing method based on multi-Agent evolutionary Algorithm, comprising the following steps:
(1) wireless sensor network is constructed:
The working region range of (1a) input wireless sensor network；
(1b) inputs the covering radius of three kinds of wireless sensors；
Three kinds of wireless sensors are randomly dispersed among the working region of wireless sensor network by (1c), complete a wireless biography
The building of sensor network,
(2) maximum number of iterations is set:
In the range of [0,500], according to the optimum results of multi-Agent evolutionary Algorithm, multi-Agent evolutionary Algorithm is set most
Big the number of iterations；
(3) Agent Grid is constructed:
Using a wireless sensor network as an intelligent body, being built into size with 81 intelligent bodies is 9 × 9 intelligent body nets
Lattice；
(4) using the Monte Carlo method with punishment, the energy of each intelligent body in the Agent Grid that size is 9 × 9 is calculated
Amount；
Specific step is as follows for the Monte Carlo method with punishment:
Step 1 calculates the area coverage of each wireless sensor network corresponding with each intelligent body according to the following formula:
Wherein, SkIndicate the area coverage of p-th of wireless sensor network corresponding with k-th of intelligent body, the numerical value phase of k and p
Population of the uniformly dispersing in working region is indicated Deng, N, and q indicates to fall in the working region covered by wireless sensor network
In population, A indicate wireless sensor network working region area；
Step 2, according to the following formula, calculating has overlapping covering in p-th of wireless sensor network corresponding with k-th of intelligent body
Penalty value between all wireless sensors:
φk(C)=λ × ∑ S'(Ci,Cj)
Wherein, φk(C) indicate that there are all of overlapping covering in p-th of wireless sensor network corresponding with k-th of intelligent body
Penalty value between wireless sensor, λ indicate that the penalty factor that numerical value is greater than zero, ∑ indicate sum operation, S'(Ci,Cj) indicate without
I-th of wireless sensor C in line sensor networkiWith j-th of wireless sensor CjOverlapping coverage, as wireless sensor Ci
With CjWhen overlapping covering, S'(Ci,Cj) value 1, as wireless sensor CiWith CjWhen non-overlapping covering, S'(Ci,Cj) value 0；
Step 3, according to the following formula, the energy of each intelligent body in computational intelligence volume mesh:
Energyk=Sk+φk(C)
Wherein, EnergykIndicate the energy of k-th of intelligent body in Agent Grid, SkIndicate pth corresponding with k-th of intelligent body
The area coverage of a wireless sensor network, φk(C) it indicates in p-th of wireless sensor network corresponding with k-th of intelligent body
Penalty value between all wireless sensors with overlapping covering；
(5) Agent Grid for being 9 × 9 to size executes neighborhood contention operation；
Specific step is as follows for the neighborhood contention operation:
Step 1, an optional intelligent body find out energy in four neighborhoods from four neighborhoods up and down of selected intelligent body
Maximum intelligent body；
The energy of energy maximum intelligent body is compared by step 2 with the energy of selected intelligent body, if energy maximum intelligent body
When energy is greater than the energy of selected intelligent body, selected intelligent body is updated with the maximum intelligent body of energy, obtains updated intelligence
Body；
(6) Agent Grid for being 9 × 9 to size executes mutation operation；
The mutation operation refers to, the random perturbation that one meets Gaussian Profile is added on updated intelligent body, is obtained
Intelligent body after variation；
(7) determine that size is the local optimum intelligent body of 9 × 9 Agent Grid:
The Monte Carlo method punished using band identical with step (4) calculates the Agent Grid that size is 9 × 9 after making a variation
In each intelligent body energy, using the maximum intelligent body of energy as 9 × 9 local optimum intelligent bodies；
(8) self study operates:
According to the optimum results of multi-Agent evolutionary Algorithm, the maximum that self study operates is arranged in the range of [0,100] in (8a)
The number of iterations；
(8b) using each wireless sensor network as an intelligent body, constructing a size with 25 intelligent bodies is 5 × 5
Agent Grid；
(8c) is calculated in the Agent Grid that size is 5 × 5 using the Monte Carlo method of band punishment identical with step (4)
The energy of each intelligent body；
The Agent Grid that (8d) is 5 × 5 to size executes and step (5) identical neighborhood contention operation；
The Agent Grid that (8e) is 5 × 5 to size executes and step (6) identical mutation operation；
(8f) is calculated each in the Agent Grid after making a variation using the Monte Carlo method of band punishment identical with step (4)
The energy of intelligent body finds out the maximum intelligent body of energy as 5 × 5 local optimum intelligent bodies；
The energy of 9 × 9 local optimum intelligent bodies is compared by (8g) with the energy of 5 × 5 local optimum intelligent bodies, if 5 × 5 innings
When the energy of the optimal intelligent body in portion is greater than the energy of 9 × 9 local optimum intelligent bodies, 9 × 9 are updated with 5 × 5 local optimum intelligent bodies
Local optimum intelligent body；
(8h) is using updated 9 × 9 local optimum intelligent body as optimal wireless sensor network；
(8i) judges whether the cyclic algebra of current self study operation reaches maximum number of iterations, if so, (9) are thened follow the steps,
Otherwise, it after the cyclic algebra that self study operates being added 1, executes step (8d)；
(9) fictitious force operates:
(9a) arbitrarily chooses a wireless sensor from optimal wireless sensor network；
(9b) according to the following formula, calculate selected wireless sensor respectively with each wireless sensing in addition to selected wireless sensor
Distance between device:
Wherein, dmIndicate between m-th of wireless sensor and n-th of wireless sensor in addition to m-th of wireless sensor away from
From xmAnd ymRespectively indicate the cross, ordinate position of m-th of wireless sensor, xnAnd ynIt respectively indicates in addition to m-th of wireless biography
The cross of n-th of wireless sensor, ordinate position except sensor；
(9c) according to the following formula, calculate selected wireless sensor respectively with each wireless sensing in addition to selected wireless sensor
The sum of covering radius of device:
Lm=Rm+Rn
Wherein, LmIndicate covering half of m-th of wireless sensor with n-th of wireless sensor in addition to m-th of wireless sensor
The sum of diameter, RmIndicate the covering radius of m wireless sensor, RnIndicate n-th of wireless biography in addition to m-th of wireless sensor
The covering radius of sensor；
(9d) judges the distance d between wireless sensormWhether covering radius the sum of L is less thanm, if so, to m-th of wireless sensor
Apply repulsive force, otherwise, attraction is applied to m-th of wireless sensor；
(9e) judges whether each wireless sensor has been selected in optimal wireless sensor network, if so, thening follow the steps
(10), otherwise, step (9a) is executed；
(10) judge whether the cyclic algebra of current multi-Agent evolutionary Algorithm reaches maximum number of iterations, if so, executing step
Suddenly (11) otherwise after the cyclic algebra of multi-Agent evolutionary Algorithm is added 1, execute step (3)；
(11) wireless sensor network distribution results are exported.
2. the wireless sensor network optimizing method according to claim 1 based on multi-Agent evolutionary Algorithm, feature
It is, the working region range of wireless sensor network described in step (1a) is 100 × 100 meters.
3. the wireless sensor network optimizing method according to claim 1 based on multi-Agent evolutionary Algorithm, feature
It is, the covering radius of three kinds of wireless sensors described in step (1b) is respectively r1, r2, r3, wherein r3=0.8 × r2
Rice, r2=0.8 × r1 meters.
4. the wireless sensor network optimizing method according to claim 1 based on multi-Agent evolutionary Algorithm, feature
It is, specific step is as follows for application repulsive force described in step (9d):
Step 1 calculates the repulsive force applied to wireless sensor according to the following formula:
Wherein,It indicates to m-th of wireless sensor CmThe repulsive force of application, RmIndicate m-th of wireless sensor CmCovering half
Diameter, RnIndicate n-th of wireless sensor CnCovering radius, d (Cm,Cn) indicate m-th of wireless sensor CmWith n-th of wireless biography
Sensor CnDistance；
Step 2, by the size and Orientation of repulsive force, the wireless sensor of mobile repulsive force to be applied.
5. the wireless sensor network optimizing method according to claim 1 based on multi-Agent evolutionary Algorithm, feature
It is, specific step is as follows for step (9d) the application attraction:
Step 1 calculates the attraction applied to wireless sensor according to the following formula:
Wherein,It indicates to m-th of wireless sensor CmThe attraction of application, RmIndicate m-th of wireless sensor CmCovering half
Diameter, RnIndicate n-th of wireless sensor CnCovering radius, d (Cm,Cn) indicate m-th of wireless sensor CmWith n-th of wireless biography
Sensor CnDistance；
Step 2, by the size and Orientation of attraction, the wireless sensor of mobile attraction to be applied.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610817374.1A CN106488482B (en) | 2016-09-12 | 2016-09-12 | Wireless sensor network optimizing method based on multi-Agent evolutionary Algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610817374.1A CN106488482B (en) | 2016-09-12 | 2016-09-12 | Wireless sensor network optimizing method based on multi-Agent evolutionary Algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106488482A CN106488482A (en) | 2017-03-08 |
| CN106488482B true CN106488482B (en) | 2019-05-24 |
Family
ID=58273772
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610817374.1A Active CN106488482B (en) | 2016-09-12 | 2016-09-12 | Wireless sensor network optimizing method based on multi-Agent evolutionary Algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106488482B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108430105A (en) * | 2017-12-28 | 2018-08-21 | 衢州学院 | Distributed sensor networks cooperate with target state estimator and interference source passive location method |
| CN108462608B (en) * | 2018-03-26 | 2020-10-23 | 西安电子科技大学 | Wireless sensor network life optimization method based on multi-agent evolutionary algorithm |
| CN108770001B (en) * | 2018-04-20 | 2021-02-12 | 西安电子科技大学 | Wireless chargeable sensor network optimization method based on memetic algorithm |
| CN114764251B (en) * | 2022-05-13 | 2023-10-10 | 电子科技大学 | Multi-agent collaborative search energy-saving method based on energy consumption model |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101286911A (en) * | 2008-03-05 | 2008-10-15 | 中科院嘉兴中心微***所分中心 | Networking method of wireless sensor network based on combination of cluster and Mesh |
| CN102663438A (en) * | 2012-05-08 | 2012-09-12 | 哈尔滨工程大学 | Monte carlo characteristics dimension reduction method for small-sample hyperspectral image |
| CN105764084A (en) * | 2016-02-05 | 2016-07-13 | 南京邮电大学 | Wireless sensor network multipath routing protocol realization method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8369242B2 (en) * | 2009-03-31 | 2013-02-05 | Empire Technology Development Llc | Efficient location discovery |
-
2016
- 2016-09-12 CN CN201610817374.1A patent/CN106488482B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101286911A (en) * | 2008-03-05 | 2008-10-15 | 中科院嘉兴中心微***所分中心 | Networking method of wireless sensor network based on combination of cluster and Mesh |
| CN102663438A (en) * | 2012-05-08 | 2012-09-12 | 哈尔滨工程大学 | Monte carlo characteristics dimension reduction method for small-sample hyperspectral image |
| CN105764084A (en) * | 2016-02-05 | 2016-07-13 | 南京邮电大学 | Wireless sensor network multipath routing protocol realization method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106488482A (en) | 2017-03-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106488482B (en) | Wireless sensor network optimizing method based on multi-Agent evolutionary Algorithm | |
| CN104408518B (en) | Based on the neural network learning optimization method of particle swarm optimization algorithm | |
| CN103971160B (en) | particle swarm optimization method based on complex network | |
| Zhu et al. | Wind speed behaviors feather analysis and its utilization on wind speed prediction using 3D-CNN | |
| CN103489038A (en) | Photovoltaic ultra-short-term power prediction method based on LM-BP neural network | |
| CN103455612B (en) | Based on two-stage policy non-overlapped with overlapping network community detection method | |
| CN110222883A (en) | Load Prediction In Power Systems method based on wind Drive Optimization BP neural network | |
| CN103455610A (en) | Network community detecting method based on multi-objective memetic computation | |
| CN103870642A (en) | Complex supply network robust performance analyzing method based on topological structure | |
| CN104462657B (en) | System of selection based on Petri network complication system multi-design methods | |
| CN106803135A (en) | The Forecasting Methodology and device of a kind of photovoltaic power generation system output power | |
| CN106570250A (en) | Power big data oriented microgrid short-period load prediction method | |
| Cao et al. | An algorithm for classification over uncertain data based on extreme learning machine | |
| CN109543914B (en) | Method for predicting RBF building operation energy consumption based on AP clustering algorithm | |
| NavnathDattatraya et al. | Maximising network lifetime and energy efficiency of wireless sensor network using group search Ant lion with Levy flight | |
| Samsudin et al. | A hybrid least squares support vector machines and GMDH approach for river flow forecasting | |
| Du et al. | Novel machine learning approach for shape-finding design of tree-like structures | |
| Li et al. | Coverage Enhancement Strategy for WSNs Based on Multi-Objective Ant Lion Optimizer | |
| CN104182854A (en) | Mixed energy consumption measuring method for enterprise energy management system | |
| Sha et al. | Prediction of commercial building lighting energy consumption based on EPSO-BP | |
| CN104703207A (en) | Intelligent sensing equipment-based traversal search method | |
| CN109086952A (en) | It is a kind of based on genetic algorithm-neural network heat load prediction method | |
| CN109034275A (en) | Prediction technique, system, medium and the equipment of polycrystalline reduction process energy consumption value | |
| CN103440540B (en) | A kind of parallel method of land utilization space layout artificial immunity Optimized model | |
| CN114531665A (en) | Wireless sensor network node clustering method and system based on Laiwei flight |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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 |