CN108966241B - Optimization method for self-adaptively improving fish swarm algorithm - Google Patents

Optimization method for self-adaptively improving fish swarm algorithm Download PDF

Info

Publication number
CN108966241B
CN108966241B CN201810685277.0A CN201810685277A CN108966241B CN 108966241 B CN108966241 B CN 108966241B CN 201810685277 A CN201810685277 A CN 201810685277A CN 108966241 B CN108966241 B CN 108966241B
Authority
CN
China
Prior art keywords
optimization
fish
behavior
visual
pixel point
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
Application number
CN201810685277.0A
Other languages
Chinese (zh)
Other versions
CN108966241A (en
Inventor
秦宁宁
许健
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangnan University
Original Assignee
Jiangnan University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Jiangnan University filed Critical Jiangnan University
Priority to CN201810685277.0A priority Critical patent/CN108966241B/en
Publication of CN108966241A publication Critical patent/CN108966241A/en
Application granted granted Critical
Publication of CN108966241B publication Critical patent/CN108966241B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

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 an optimization method for a self-adaptive improved fish swarm algorithm, and belongs to the field of sensor network coverage. By the aid of the increased jumping behaviors, the speed and efficiency of jumping out of the local optimum of the fish school are improved; a new food concentration judgment method is set by combining the K neighborhood mean thought, so that the probability of falling into local optimum is reduced; the wall collision behavior is added, so that the critical processing of the node coverage condition at the network boundary is perfected; the step length and the visual field of the fish are adjusted in a self-adaptive mode through the attenuation factor theta, and the stability of convergence is guaranteed. On the premise of ensuring the network coverage rate, the invention solves the problem that the fish swarm algorithm has slow response to the severe local optimum, reduces the probability of falling into the local optimum, perfects the critical processing of the node coverage condition at the network boundary and ensures the stability of convergence.

Description

Optimization method for self-adaptively improving fish swarm algorithm
Technical Field
The invention relates to an optimization method for a self-adaptive improved fish swarm algorithm, and belongs to the field of sensor network coverage.
Background
In the wireless sensor network performance optimization problem, network coverage has become an important research direction. A large number of sensor nodes are placed in a monitoring environment in a scattering mode. Due to the high node density and the random deployment, a large number of redundant nodes usually exist, and the network coverage efficiency is low. Usually, an intelligent algorithm is adopted, and the effective coverage area is increased and the coverage performance is improved in a mode of moving the sensor nodes.
In the process of node optimization by adopting an intelligent algorithm, the node optimization often falls into the adverse situation of local optimization (node large-scale accumulation), the performance of the algorithm is influenced, and the optimization efficiency is further reduced. In order to improve coverage performance and solve local optimization problem, improvement or combination of multiple algorithms is usually adopted to deal with the problem. At present, an algorithm (VFDPSO) combining Virtual Force and Particle Swarm Optimization is available, and high coverage rate is obtained through overall Optimization in a whole network range. VFDPSO achieves automatic Optimization and network coverage improvement, but the stability of coverage performance is not high due to failure to overcome the problem that a typical Particle Swarm Optimization (PSO) is prone to fall into a locally optimal aeipathia. Aiming at the problem, the wireless sensor network coverage optimization based on the artificial fish swarm algorithm (published source: computer application research, 2013,30(02): 554) 556) introduces the fish swarm algorithm into the research of the sensor network coverage problem, and promotes the objective function to jump out of local optimum by simulating the behaviors of herd clustering, rear-end collision and the like. Although AFSA solves the local optimization problem, the algorithm is time consuming. In order to solve the problem that the timeliness of the algorithm is improved, a fish swarm algorithm fusing virtual force is proposed in wireless sensor network coverage efficiency optimization simulation (published source: computer simulation, 2017,34(08):297 + 301) and the speed of jumping out a local optimal value is improved by introducing the fusion virtual force, but the algorithm still has the defects of slow response to severe local optimal value, unobvious coverage rate improvement in the later period and the like.
The above has proven that the fish swarm algorithm can be applied to solve the locally optimal trap, but the speed of jumping out of the trap, the reaction facing the local optimization, and the processing of the coverage performance improvement and boundary problem of the algorithm are still to be improved.
Disclosure of Invention
In order to solve the problem that the response of the fish swarm algorithm to the severe local optimal is slow at present, the invention provides an optimization method for adaptively improving the fish swarm algorithm, and the technical scheme is as follows:
suppose that N isomorphic sensor nodes S ═ S are scattered randomly in the detected region I1(x1,y1),S2(x2,y2),..Sk(xk,yk),...,SN(xN,yN) Where k is 1,2, 3.., N; the nodes all adopt a Boolean sensing model with radius r;
for convenience of description, the region I is discretized into l × l pixel point sets M ═ M simultaneously(1,1),M(1,2),...,M(x,y),...,M(l,l)N, and all nodes which are set to be randomly deployed fall on a certain pixel point in the I;
step 1: each node SkRespectively executing 5 behaviors of foraging, clustering, repelling, jumping and wall collision, and recording the position of behavior optimization in a Billboard;
step 2: calculating the network coverage rate Y after the current round of optimization;
and step 3: if the network coverage rate Y obtained after the current round of optimization is improved relative to the network coverage rate Y obtained before optimization, the node set S moves according to the position points recorded by the Billboard; otherwise, according to whether reaching local optimizing upper limitCthJudging whether the visual field needs to be attenuated or not and judging the step length step;
wherein step represents a step size for the fish to perform one movement, and step satisfies step ═ step0× rand representing a random number between (0,1), step0A distance base representing a movement of a fish performing a behavior;
and 4, step 4: judging Y is more than YthOr visual<VthIf yes, exiting the algorithm; if not, returning to the step 2; wherein Y isthAnd VthIndicating the ending thresholds for visual and Y.
Optionally, the jump behavior is defined as follows:
at node SkIn the visual, any point of pixel point M is selected(x,y)；
If node SkSatisfies the following conditions: densityk＞QthAnd p isk≥PthThen S iskExecuting jumping action, jumping to ArgMax (T) pixel point(x,y))；
Wherein, pixel point M(x,y)Satisfy the requirement of
Figure BDA0001711529030000021
If | ArgMax (T)(x,y)) If | is greater than 1, then SkRandom jump to ArgMax (T)(x,y)) Any one of the pixel points; ArgMax (T)(x,y)) Represents T(x,y)Obtaining the corresponding (x, y) coordinate when the maximum value is obtained;
defining a node SkThe sum of the coverage rates of the pixel points in the sensing range r is the node SkDegree of congestion ofk：
Figure BDA0001711529030000022
M(e,f)Satisfy the requirement of
Figure BDA0001711529030000023
Wherein
Figure BDA0001711529030000025
Presentation pairr is rounded down;
food concentration T(x,y)Is defined as: pixel point M(x,y)And the parameter K, then the pixel point M(x,y)Food concentration T of(x,y)And M(x,y)The coverage rate of the pixel points in the K neighborhood is related:
Figure BDA0001711529030000024
applying K neighborhood averaging to T(x,y)In the calculation, M is realized(x,y)The coverage of the pixel points in the surrounding K neighborhood is brought into T(x,y)The calculation category of (1); by setting different weight coefficients for each layer neighborhood
Figure BDA0001711529030000031
Characterization M(x,y)Differences in influence between layers within a neighborhood;
k neighborhood, representing M(x,y)The surrounding K layers are neighbors, optionally K3.
Optionally, the wall-touching behavior is defined as follows:
if node SkAnd the vertical distance d between the boundary of the region I
Figure BDA0001711529030000032
Then the node SkMoving in the opposite direction to the boundary for a distance rand × (r-d);
update visual and step: given CthAfter the optimization of the round, if the convergence increment of Y is insufficient, namely Y does not meet the increment index after the multi-round optimization, updating the visual sum by using the attenuation factor theta, and reducing the visual field and the step length; the update formula is as follows:
visual＝visual×θ (3)
step＝step×θ (4)
wherein the attenuation factor theta is 0.8
Optionally, the network coverage Y is defined as:
Figure BDA0001711529030000033
wherein, when the pixel point M(x,y)And any one of SkWhen the distance between the two is less than r, M is considered(x,y)Can be covered by S, i.e. network S at point M(x,y)The coverage rate is:
Figure BDA0001711529030000034
wherein, when the pixel point M(x,y)And a sensor node SkWhen the distance between the two is less than r, the S is consideredkAt M(x,y)Is 1, otherwise is 0, i.e.:
Figure BDA0001711529030000035
optionally, the foraging, clustering, and repelling behaviors are respectively defined as follows:
foraging behavior: at node SkIn the visual, any point of pixel point M is selected(x,y)If M is present(x,y)Satisfy the requirement of
Figure BDA0001711529030000036
Then consider point M(x,y)The food concentration of (A) is higher than SkConcentration of food at the location, SkTo M(x,y)Moving the step; if given the selection constraint, point M cannot be found(x,y)Then the current fish SkStarting random walk;
wherein D (·) represents the euclidean distance between two elements;
clustering behavior: if node SkSatisfy the requirement of
Figure BDA0001711529030000037
When S is presentkMoving step to the current nearest fish direction;
rejection behavior: if node SkSatisfy the requirement of
Figure BDA0001711529030000041
When S is presentkWhen facingMoving step in the opposite direction of the former nearest fish;
wherein, α and β are α values and are the distance between the node when the perception scope boundary of two nodes is tangent, namely α is 2r is 16m, exclusion threshold β carries out differentiation setting according to the experimental demand, sets up
Figure BDA0001711529030000042
Optionally, before step 1, the method further includes:
various parameters are initialized.
Optionally, the method is applied to coverage performance optimization of a wireless sensor network.
Optionally, the method is applied to the optimization of the coverage performance of the wireless sensor network, and particularly applied to the detection of blue-green algae in lakes.
Optionally, the application to the coverage performance optimization of the wireless sensor network is specifically applied to emergency search of an aged people's home.
The invention has the beneficial effects that:
by the aid of the increased jumping behaviors, the speed and efficiency of jumping out of the local optimum of the fish school are improved; a new food concentration judgment method is set by combining the K neighborhood mean thought, so that the probability of falling into local optimum is reduced; the wall collision behavior is added, so that the critical processing of the node coverage condition at the network boundary is perfected; the step length and the visual field of the fish are adjusted in a self-adaptive mode through the attenuation factor theta, and the stability of convergence is guaranteed. On the premise of ensuring the network coverage rate, the invention solves the problem that the fish swarm algorithm has slow response to the severe local optimum, reduces the probability of falling into the local optimum, perfects the critical processing of the node coverage condition at the network boundary and ensures the stability of convergence.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 is a graph comparing the effect of network size on network coverage in different algorithms;
fig. 2 is a graph of coverage versus number of moves for locally optimal analysis of different algorithms, where (a) Test-50, (b) Test-40, (c) Test-30, and (d) Test-20;
FIG. 3 is a graph of the AIFS coverage acceleration analysis results for different local optima.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
The first embodiment is as follows:
the embodiment provides an optimization method for adaptively improving a fish swarm algorithm, and the agreed experimental scene is as follows: in a given detected area I, randomly scattering N isomorphic sensor nodes S ═ S1(x1,y1),S2(x2,y2),..Sk(xk,yk),...,SN(xN,yN) For convenience of description, the region I is discretized into l × pixel point sets M ═ M ·(1,1),M(1,2),...,M(x,y),...,M(l,l)And f, setting x, y to be 1,2, 3.
The invention provides an optimization method for a self-adaptive improved fish swarm algorithm, which comprises the following steps:
the method comprises the following steps: and initializing each parameter.
Step two: each node SkThe 5 actions of foraging, bunching, repelling, jumping and bumping are performed separately, and the position after action optimization is recorded in the Billboard.
Step three: and calculating the optimized Y of the current wheel.
Step four: if the network coverage rate Y obtained by the current round of optimization is improved compared with the network coverage rate Y obtained before optimization, the node setS, moving the position points recorded according to the Billboard; otherwise, according to whether the local optimization upper limit C is reachedthAnd judging whether the visual field and the step length need to be attenuated or not so as to reduce the visual field and the step length and enter a new round of optimization.
step denotes a step size of a movement performed by the fish, and step satisfies step ═ step0× rand, rand representing a random number between (0,1)0The distance base representing the movement of a fish performing a behavior.
Step five: judging Y is more than YthOr visual<VthAnd if so, exiting the algorithm. If not, returning to the step 2. Wherein Y isthAnd VthIndicating the ending thresholds for visual and Y.
Behavioral interpretation
Before explaining the behavior in detail, a few key definitions in the present invention are agreed as follows:
definition 1 (pixel perception ratio): when pixel point M(x,y)And a sensor node SkWhen the distance between the two is less than r, the S is consideredkAt M(x,y)Is 1, otherwise is 0, i.e.:
Figure BDA0001711529030000051
definition 2 (pixel dot coverage): when pixel point M(x,y)And any one of SkWhen the distance between the two is less than r, M is considered(x,y)Can be covered by S, i.e. network S at point M(x,y)The coverage rate is:
Figure BDA0001711529030000052
definition 3 (network coverage): defining the ratio of the sum of the coverage rates of all pixel points in the network to the total number of the pixel points as the network coverage rate Y, namely:
Figure BDA0001711529030000061
from the formulae (6) and(5) it can be seen that P (M) is satisfied(x,y)And the more the number of the pixels of which the S) ═ 1 is, the more effectively the coverage rate Y of the network to the detection area I can be improved. Therefore, the goal of optimizing the coverage performance of the sensor network is to move the nodes with redundant coverage and increase the coincidence P (M) as much as possible(x,y)And S) 1 pixel point.
In step two, performing actions in the basic fish swarm algorithm, including foraging, herding, and repelling; the three behaviors are as follows:
behavior 1 (foraging behavior): at a current certain node SkIn the visual, any point of pixel point M is selected(x,y)If M is present(x,y)Satisfy the requirement of
Figure BDA0001711529030000062
Then consider point M(x,y)The food concentration of (A) is higher than SkConcentration of food at the location, SkTo M(x,y)Step is moved. If given the selection constraint, point M cannot be found(x,y)Then the current fish SkA random walk is initiated. Where D (-) represents the euclidean distance between two elements.
Behavior 2 (clustering behavior): if it is present SkSatisfy the requirement of
Figure BDA0001711529030000063
When S is presentkStep is moved to the direction of the current nearest fish.
Behavior 3 (repulsive behavior): if it is present SkSatisfy the requirement of
Figure BDA0001711529030000064
When S is presentkStep is moved in the opposite direction to the current nearest fish.
Wherein, α and β are α values and are the distance between the node when the perception scope boundary of two nodes is tangent, namely α is 2r is 16m, exclusion threshold β carries out differentiation setting according to the experimental demand, sets up
Figure BDA0001711529030000065
In addition to the basic fish school behavior, two behaviors are added, namely a jumping behavior and a wall hitting behavior.
Behavior 4 (jump behavior): if node SkSatisfies the following conditions: densityk＞QthAnd p isk≥PthThen S iskExecuting jumping action, jumping to ArgMax (T) pixel point(x,y)) In which satisfy
Figure BDA0001711529030000066
If | ArgMax (T)(x,y)) If | is greater than 1, then SkRandom jump to ArgMax (T)(x,y)) Any one of the pixel points. Among them, ArgMax (T)(x,y)) Represents T(x,y)The (x, y) coordinate corresponding to the maximum value is obtained.
Definition 4 (crowdedness): defining a certain node SkThe sum of the coverage rates of the pixel points in the sensing range r is the node SkDegree of congestion ofk
Figure BDA0001711529030000067
M(e,f)Should satisfy
Figure BDA0001711529030000068
Wherein
Figure BDA0001711529030000069
Indicating rounding down r. DensitykEmbody SkThe degree of congestion with the neighbor node is a measure of SkAnd sensing whether important parameters of redundant nodes exist in the range.
Definition 5 (food concentration T)(x,y)): given pixel point M(x,y)And the parameter K, then the pixel point M(x,y)Food concentration T of(x,y)And M(x,y)The coverage rate of the pixel points in the K neighborhood is defined as follows:
Figure BDA0001711529030000071
applying K neighborhood averaging to T(x,y)In the calculation, realizeWill M(x,y)The coverage of the pixel points in the surrounding K neighborhood is brought into T(x,y)The category of calculation of (1). By setting different weight coefficients for each layer neighborhood
Figure BDA0001711529030000072
Characterization M(x,y)Differences in influence between layers within a neighborhood.
Behavior 5 (wall-strike behavior): if node SkAnd the vertical distance d between the boundary of the region I
Figure BDA0001711529030000073
The sensor node moves in the opposite direction toward the boundary and the distance traveled is rand × (r-d).
In order to further enhance the convergence of the algorithm, when the convergence increment of the algorithm is insufficient, the overall accuracy of the algorithm is improved by adjusting the step size and the visual field.
visual and step updates: given CthAfter the round of optimization, if the convergence increment of Y is insufficient, namely Y does not meet the increment index after the round of optimization, updating the visual sum by using the attenuation factor theta, and reducing the visual field and the step length. The update formula is as follows:
visual＝visual×θ (3)
step＝step×θ (4)
the values of visual and step are reduced, the purpose is to improve the optimizing precision, improve the possibility of increasing the network coverage rate and help the algorithm to continuously and effectively run.
To illustrate the optimization effect of the optimization method of the adaptive fish swarm algorithm provided by the invention in detail, the following description is given by means of experimental data:
in the experiment, the network scale is adjusted by adjusting the number N of nodes so as to investigate the degree of influence of the network scale N on the network coverage rate Y of the AIFS, a Particle Swarm Optimization (PSO) and a coverage Optimization algorithm (Coverageoptimization algorithm based on Sampling for Homogeneous WSNs, COSH) based on Sampling information. And setting N epsilon [30,65] in the experiment, and carrying out 20 rounds of optimization at the growth interval of 5 node numbers.
The results of the comparative experiment are shown in FIG. 1. Wherein origin represents Y measured in the original environment. As can be seen from the graph, the optimization effect of each algorithm pair is improved with the increase of N, and when the number of nodes N is greater than 40, the improvement effect is more remarkable. Although Y after optimization of each algorithm is increased, the promotion of Y by the intelligent algorithms PSO and AIFS is obviously superior to that by the non-intelligent algorithm COSH, and the optimization effect of AIFS is superior to that of PSO. In particular, as the network scale is further increased (N > 50), the gain of network coverage improvement of AIFS is more prominent than that of PSO, because as N increases, the degree of congestion between nodes also increases to different degrees, and AIFS with hopping and repulsion behavior exhibits better convergence performance.
To verify the ability of the AIFS to jump out of local optima, i.e., the ability of the sensor node to quickly discover clusters trapped in the node and quickly go off-cluster. And simulating the local optimum condition at the central position of the area I, and randomly arranging all the node sets S in a square with the side length of Test at the central position of the area I, wherein the Test is 50,40,30 and 20. Consider the changes in network coverage rate Y during each round of node movement in AIFS, Fish Swarm algorithm afsa (artificial Fish Swarm algorithm) and Particle Swarm virtual force in combination with vfdpso (virtual forces mining Particle Swarm optimization). The results of the experiment are shown in FIG. 2.
As shown in fig. 2, when the node falls into the local optimum, after moving for many times, the three algorithms of AIFS, AFSA and VFDPSO can increase Y to different degrees, so that the sensor network jumps out of the local optimum trap. However, compared with the local irregularity of the VSDPSO coverage rate Y improvement, the Y rising curves of the AIFS and the AFSA are smoother, so that the AIFS and the AFSA both have the capability of stably jumping out of the local optimum when the local optimum is responded. Compared with the rising speed of the curve, the AIFS is obviously superior to the AFSA in jumping out of the local optimum, and the advantage is more obvious along with the deterioration of the local optimum (Test ↓).
In order to analyze the capability of the AIFS jumping out of the local optimum in more detail, time is used for representing the number of node movement, DY is used for representing the acceleration of Y relative to the previous movement, DY is extracted when the AIFS is in four local optimum states, namely 5 time points (time is 2,4,6,8 and 10), and statistical data are shown in FIG. 3; after the node movement time is 10 times, the coverage incremental value DY of VFDPSO and AFSA gradually decreases as the Test decreases. It is worth to be noted that, under any given local optimum condition, the AIFS can obtain DY being 7-14% coverage gain at an earlier time point (time being 2), which means that when the worst local optimum is met, the AIFS can accelerate to jump out of the local optimum at the initial stage of the algorithm, and stronger adaptability is embodied.
The method is applied to lake blue-green algae detection, such as lake Taihu blue-green algae monitoring:
if the blue algae is to be monitored, the sensors need to be scattered on the water surface, but after a period of water surface fluctuation, the positions of the sensors are changed randomly, and at the moment, the sensors are moved and deployed again by adopting an AIFS algorithm, so that the effect of covering more water surfaces can be achieved.
The method is applied to the emergency search of the nursing home,
such as: the method is characterized in that the old people need to be urgently searched at a certain moment in a nursing home, a sensor capable of sensing a signal transmitter on the old people within a certain distance is placed on a nursing worker under the condition that a traditional indoor positioning algorithm cannot guarantee an accurate searching rate, when an emergency request is executed, the AIFS algorithm can give an instruction to guide the nursing worker with random positions to move, the nursing home area is covered as far as possible, and therefore the region where the target old people are located is searched.
According to the experimental process, the speed and the efficiency of the fish shoal jumping out of the local optimum are improved through the increased jumping behavior; a new food concentration judgment method is set by combining the K neighborhood mean thought, so that the probability of falling into local optimum is reduced; the wall collision behavior is added, so that the critical processing of the node coverage condition at the network boundary is perfected; the step length and the visual field of the fish are adjusted in a self-adaptive mode through the attenuation factor theta, and the stability of convergence is guaranteed. On the premise of ensuring the network coverage rate, the invention solves the problem that the fish swarm algorithm has slow response to the severe local optimum, reduces the probability of falling into the local optimum, perfects the critical processing of the node coverage condition at the network boundary and ensures the stability of convergence.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

Claims (7)

1. An optimization method for adaptively improving a fish swarm algorithm, the method comprising:
suppose that N isomorphic sensor nodes S forming network S are scattered randomly within detected region I as { S }1(x1,y1),S2(x2,y2),..Sk(xk,yk),...,SN(xN,yN) Where k is 1,2, 3.., N; the sensor nodes all adopt a Boolean sensing model with radius r;
for convenience of description, the region I is discretized into l × l pixel point sets M ═ M simultaneously(1,1),M(1,2),...,M(x,y),...,M(l,l)N, and setting randomly deployed sensor nodes to fall on a certain pixel point in the I;
step 1: each sensor node SkRespectively executing 5 actions of foraging, clustering, repelling, jumping and wall collision, and recording the position of the action after optimization;
step 2: calculating the network coverage rate Y after the current round of optimization;
and step 3: if the network coverage rate Y obtained after the current round of optimization is improved relative to the network coverage rate Y obtained before the current round of optimization, the sensor node set S moves according to the recorded position points; otherwise, according to whether the local optimization upper limit C is reachedthJudging whether the visual field needs to be attenuated or not and judging the step length step;
wherein step represents a step size for the fish to perform one movement, and step satisfies step ═ step0× rand representing a random number between (0,1), step0A distance base representing a movement of a fish performing a behavior;
and 4, step 4: judging Y is more than YthOr visual < VthIf yes, exiting the algorithm; if not, returning to the step 2; wherein Y isthAnd VthAn end threshold representing Y and visual;
the jump behavior is defined as follows:
at sensor node SkIn the visual, any point of pixel point M is selected(x,y)；
If sensor node SkSatisfies the following conditions: densityk＞QthAnd p isk≥PthThen S iskExecuting jumping action, jumping to ArgMax (T) pixel point(x,y))；
Wherein, pixel point M(x,y)Satisfy the requirement of
Figure FDA0002599782960000011
If | ArgMax (T)(x,y)) If | is greater than 1, then SkRandom jump to ArgMax (T)(x,y)) Any one of the pixel points; ArgMax (T)(x,y)) Represents T(x,y)Obtaining the corresponding (x, y) coordinate when the maximum value is obtained;
defining sensor nodes SkThe sum of the coverage rates of the network S at each pixel point in the sensing range r is the sensor node SkDegree of congestion ofk：
Figure FDA0002599782960000021
M(e,f)Satisfy the requirement of
Figure FDA0002599782960000022
Wherein
Figure FDA0002599782960000023
Represents rounding down r;
food concentration T(x,y)Is defined as: pixel point M(x,y)And the parameter K, then the pixel point M(x,y)Food concentration T of(x,y)And M(x,y)Coverage phase of pixel points in K neighborhoodTurning off:
Figure FDA0002599782960000024
applying K neighborhood averaging to T(x,y)In the calculation, M is realized(x,y)The coverage of the pixel points in the surrounding K neighborhood is brought into T(x,y)The calculation category of (1); by setting different weight coefficients for each layer neighborhood
Figure FDA0002599782960000025
Characterization M(x,y)Differences in influence between layers within a neighborhood;
the wall-strike behavior is defined as follows:
if sensor node SkAnd the vertical distance d between the boundary of the region I
Figure FDA0002599782960000026
Then the sensor node SkMoving in the opposite direction to the boundary for a distance rand × (r-d);
update visual and step: given CthAfter the optimization of the round, if the convergence increment of Y is insufficient, namely Y does not meet the increment index after the multi-round optimization, updating visual and step length by using an attenuation factor theta; the update formula is as follows:
visual＝visual×θ (3)
step × θ (4), where the attenuation factor θ is 0.8.
2. The method of claim 1, wherein the network coverage Y is defined as:
Figure FDA0002599782960000027
wherein, when the pixel point M(x,y)And any one of SkWhen the distance between the two is less than r, M is considered(x,y)Can be covered by S, i.e. network S at point M(x,y)The coverage rate is:
Figure FDA0002599782960000028
wherein, when the pixel point M(x,y)And a sensor node SkWhen the distance between the two is less than r, the S is consideredkAt M(x,y)Is 1, otherwise is 0, i.e.:
Figure FDA0002599782960000031
3. the method of claim 1, wherein the foraging, clumping, and repelling behaviors are each defined as follows:
foraging behavior: at sensor node SkIn the visual, any point of pixel point M is selected(x,y)If M is present(x,y)Satisfy the requirement of
Figure FDA0002599782960000032
Then consider point M(x,y)The food concentration of (A) is higher than SkConcentration of food at the location, SkTo M(x,y)Moving the step; if given the selection constraint, point M cannot be found(x,y)Then the current fish SkStarting random walk;
wherein D (·) represents the euclidean distance between two elements;
clustering behavior: if sensor node SkSatisfy the requirement of
Figure FDA0002599782960000033
When S is presentkMoving step to the current nearest fish direction;
rejection behavior: if sensor node SkSatisfy the requirement of
Figure FDA0002599782960000034
When S is presentkMoving step in the opposite direction of the current nearest fish;
wherein alpha is the distance between the sensor nodes when the sensing range boundaries of the two sensor nodes in the clustering behavior are tangent, and beta is the distance between the sensor nodes when the sensing range boundaries of the two sensor nodes in the repulsion behavior are tangent.
4. The method of claim 1, wherein step 1 is preceded by:
various parameters are initialized.
5. The method according to any one of claims 1 to 4, wherein the method is applied to wireless sensor network coverage performance optimization.
6. The method according to claim 5, wherein the method is applied to optimization of wireless sensor network coverage performance, in particular to detection of blue-green algae in lakes.
7. The method according to claim 5, wherein the method is applied to wireless sensor network coverage performance optimization, in particular to a senior citizens home emergency search.
CN201810685277.0A 2018-06-28 2018-06-28 Optimization method for self-adaptively improving fish swarm algorithm Active CN108966241B (en)

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 CN108966241A (en) 2018-12-07
CN108966241B true 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)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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
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

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106845623B (en) * 2016-12-13 2019-09-13 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

Also Published As

Publication number Publication date
CN108966241A (en) 2018-12-07

Similar Documents

Publication Publication Date Title
CN110856134B (en) Large-scale wireless sensor network data collection method based on unmanned aerial vehicle
Yang et al. Machine learning techniques and a case study for intelligent wireless networks
Tang et al. Multi-strategy adaptive particle swarm optimization for numerical optimization
CN110062390B (en) Wireless sensor network node optimization deployment method based on improved wolf colony algorithm
CN108966241B (en) Optimization method for self-adaptively improving fish swarm algorithm
CN103401922B (en) Distributed localization apparatus and method based on game method in wireless sensor network
CN106792750B (en) A kind of node deployment method based on guiding particle swarm algorithm
CN103747537B (en) Wireless sensor network outlier data self-adaption detecting method based on entropy measurement
CN110381442B (en) Swarm robot target searching method based on implicit information interaction mode
CN113552899B (en) Multi-unmanned-aerial-vehicle autonomous collaborative search method under communication link limitation
CN112084914B (en) Multi-target tracking method integrating space motion and apparent feature learning
CN114415663A (en) Path planning method and system based on deep reinforcement learning
CN102014344A (en) Clustering control method of intelligent wireless sensor network based on DPSO (Discrete Particle Swarm Optimization)
Kong et al. An improved method of WSN coverage based on enhanced PSO algorithm
CN105050095B (en) A kind of topological construction method of the heterogeneous wireless sensor net based on energy predicting
CN112469050B (en) WSN three-dimensional coverage enhancement method based on improved wolf optimizer
Sinclair et al. A kernel methods approach to reducing handover occurrences within LTE
CN114827191B (en) Dynamic task unloading method for fusing NOMA in vehicle-road cooperative system
CN109242157A (en) A kind of improved power transmission engineering cost intelligent Forecasting
Dziwiński et al. Fully controllable ant colony system for text data clustering
CN107148026A (en) A kind of source of radio frequency energy Optimization deployment method energized for body network node
Shahzad et al. Weighted clustering using comprehensive learning particle swarm optimization for mobile ad hoc networks
Li et al. Motion adaptive vertical handoff in cellular/WLAN heterogeneous wireless network
CN114980134B (en) Mobile multi-target dynamic coverage method based on virtual force
Chen et al. Accuracy-based learning classifier systems for multistep reinforcement learning: a fuzzy logic approach to handling continuous inputs and learning continuous actions

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