CN108112049B - A kind of wireless sensor network efficiency optimization cluster-dividing method based on whale group algorithm - Google Patents
A kind of wireless sensor network efficiency optimization cluster-dividing method based on whale group algorithm Download PDFInfo
- Publication number
- CN108112049B CN108112049B CN201711351966.XA CN201711351966A CN108112049B CN 108112049 B CN108112049 B CN 108112049B CN 201711351966 A CN201711351966 A CN 201711351966A CN 108112049 B CN108112049 B CN 108112049B
- Authority
- CN
- China
- Prior art keywords
- whale
- node
- current
- individual
- cluster
- 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
- H04W40/00—Communication routing or communication path finding
- H04W40/02—Communication route or path selection, e.g. power-based or shortest path routing
- H04W40/04—Communication route or path selection, e.g. power-based or shortest path routing based on wireless node resources
- H04W40/10—Communication route or path selection, e.g. power-based or shortest path routing based on wireless node resources based on available power or energy
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
- H04W52/02—Power saving arrangements
- H04W52/0209—Power saving arrangements in terminal devices
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The wireless sensor network efficiency based on whale group algorithm that the invention discloses a kind of optimizes cluster-dividing method, including, initial configuration information is separately sent to all leader cluster nodes and ordinary node by aggregation node, and carries out convergence acquisition to the network information；According to current network information, optimal routing plan and optimal sub-clustering scheme of all leader cluster nodes to aggregation node are obtained；According to optimal sub-clustering scheme and optimal routing plan, clustering routing configuration is carried out to entire wireless sensor network；Leader cluster node carries out data fusion to place sub-clustering and is sent to aggregation node, completes converging information acquisition.The method of technical solution of the present invention introduces improved whale group algorithm in efficiency optimization sub-clustering problem, using whale group algorithm is improved in the solution ability of multi-modal optimization problem, solves wireless sensor network efficiency and optimizes sub-clustering problem；Meanwhile routing algorithm can energy consumption of the active balance leader cluster node when forwarding data, to be conducive to further extend network lifecycle.
Description
Technical field
The invention belongs to wireless sensor network topology control fields, and in particular to a kind of based on the wireless of whale group algorithm
Sensor network efficiency optimizes cluster-dividing method.
Background technique
With the rapid development of microelectric technique, computing technique, wireless communication technique and distributed information processing, nothing
Line sensor network becomes one of one of key technology in next generation network and 21 century most important emerging technology.
1999, WSNs was classified as one of 21 technologies that 21 century most influences by U.S.'s Business Week, and 2003, Massachusetts science and engineering
WSNs is classified as first of the following ten big new technologies for changing people's life by institute's " technology review ".
In most wireless sensor application scenarios, the use bottleneck of wireless sensor is embodied in its energy to be had very much
Limit, this is because most wireless sensor networks are powered using battery, this greatly limits wireless sensor networks
Working time.Therefore, energy is efficient using for its key effect of the research of wireless sensor network.Wherein topology control
It is to improve wireless sensor network efficiency with routing algorithm, extends two big key technologies of network lifecycle.It is opened up in hierarchical
It flutters in structure, cluster interior nodes (member node) are responsible for data acquisition, and send the data to the leader cluster node of place cluster, cluster head section
It puts responsible data fusion and transmits data to base station, data redundancy can be reduced in this way, and reduce data traffic, thus
Extend Network morals；Good efficiency routing algorithm can guarantee that data are forwarded along the path of optimization, balanced entire net
The energy consumption of network, to extend Network morals.
In many hierarchical application of higher wireless sensor network, leader cluster node is selected from common sensor node
It selects, therefore, these leader cluster nodes can be because of frequent data receiver, data fusion and with base station communication and soon by energy
It is exhausted, to reduce the life cycle of whole network.Currently, Many researchers are made using a kind of special node (gateway)
For cluster head, shown in leader cluster node as shown in figure 1, these nodes have energy more more than general sensor nodes, common to sense
Any one cluster head in its communication range can be added in device node.These gateways Yi Dan arrangement in the scene after, only it
Could be used as cluster head, other general sensor nodes can't be cluster head.If in some cluster general sensor nodes (at
Member's node) excessively, then the energy of leader cluster node is easy to exhaust；If member node is too far from leader cluster node, member node is held
Easily because long distance transmission data and quickly by depleted of energy, both of these case to the life cycle environmental impact of whole network all very
Greatly.Therefore, that is, the energy consumption of leader cluster node is considered it is further contemplated that the Topology Control Algorithm of the energy consumption of member node is raw for whole network
The life period seems particularly critical.
Whale group algorithm (Whale Swarm Algorithm, WSA) (is detailed in bibliography: Bing Zeng, Liang
Gao*,Xinyu Li.Whale Swarm Algorithm for Function Optimization,Intelligent
Computing T heories and Application:13th International Conference,ICIC 2017,
Liverpool,UK,August 7-10,2017,Proceedings,Part I,Springer International
Publishing, Cham, 2017, pp.624-639.) it is a kind of meta-heuristic algorithm.The algorithm simulation whale is looked for food process, is led to
It crosses being in communication with each other between whale and comes search of food source.When whale has found food source, it can make a sound other near notice
Whale about quality of food quality and quantity number information.Therefore, every whale will receive largely from neighbouring whale
Notification information, be then moved to local search of food appropriate according to these information.WSA algorithm frame is simple, easy to accomplish,
It is very suitable to due to solving engineering optimization problem.But whale group algorithm (WSA) be for solving continuous optimization problems, and
Wireless sensor network clustering problem is a discrete optimization problems of device, and whale group algorithm is not directly applicable wireless sensor network
The solution of network sub-clustering problem.
It is of the invention for ease of understanding, it is following first in the embodiment of the present invention there is close term be described collectively and
It explains:
Leader cluster node: node energy is more, and the sensor node with certain computing capability；
General sensor nodes: only carrying out the acquisition of information and the node of information is sent to cluster head, and node energy is smaller, and
Computing capability has larger limitation；
Sub-clustering: several general sensor nodes are assigned to some leader cluster node and form cluster, cluster completes the letter in cluster
Breath convergence, and there is cluster head to carry out information fusion and send it to aggregation node；
Distance: the Euclidean distance between sensor node；
Hop count: current sensor node transmits packets to the minimum sensor node number that aggregation node needs to pass through；
Dump energy: the electricity remaining value of sensor node (including general sensor nodes and leader cluster node)；
Global information: neighbours' leader cluster node set, general sensor nodes and cluster head section including general sensor nodes
The distance of point, the distance between leader cluster node, each sensor node (including general sensor nodes and leader cluster node) it is existing
The hop count of energy and leader cluster node to aggregation node.
In technical solution of the present invention, (pass through known to all wireless sensor network node positions in wireless sensor network
GPS positioning is known or by known to other location algorithms), i.e., known to euclidean distance between node pair.
Summary of the invention
Aiming at the above defects or improvement requirements of the prior art, the present invention provides a kind of based on the wireless of whale group algorithm
Sensor network efficiency optimizes cluster routing method.The method of technical solution of the present invention is difficult to handle for existing whale algorithm
The case where Wireless sensor network clustering problem, carries out wireless sensor network sub-clustering calculating based on improved whale group algorithm, obtains
To entire Wireless sensor network clustering scheme, the standard deviation for both considering leader cluster node life cycle is established, it is further contemplated that commonly
The fitness function of sensor node energy consumption and its standard deviation solves wireless sensor network member node unreasonable distribution,
The problems such as leader cluster node energy consumption is uneven extends net so that the energy consumption of balanced wireless sensor network, effectively improves network energy efficiency
Network life cycle.
To achieve the above object, according to one aspect of the present invention, a kind of wireless biography based on whale group algorithm is provided
Sensor network energy efficiency optimizes cluster-dividing method, which is characterized in that including
Initial configuration information is separately sent to leader cluster node and ordinary node by aggregation node, and to wireless sensor network
Nodal information carry out convergence acquisition, obtain current network information；
According to current network information, the optimal forward-path of each leader cluster node is calculated, to obtain all cluster heads
Optimal routing plan of the node to aggregation node, and the optimal sub-clustering scheme of entire sensor network；
According to optimal sub-clustering scheme and optimal routing plan, clustering routing configuration is carried out to entire wireless sensor network；
Sensor node acquisition and/or processing information, generation includes the number of sensor node self rest energy information
Corresponding leader cluster node is forwarded a packet to according to packet, and by the data；Leader cluster node carries out data fusion to place sub-clustering, according to sub-clustering
Routing configuration transmits data to aggregation node, completes converging information acquisition；
Wherein, the optimal sub-clustering scheme preferably passes through improved whale algorithm calculating and obtains, comprising:
Whale group algorithm parameter is arranged in S1, and initializes to each whale individual in whale group algorithm, obtains just
Beginning whale population calculates the fitness value of each whale individual；The individual that current whale is whale group is set, into step
Rapid S2；
S2 determines current whale with the presence or absence of guidance individual according to the distance between fitness value and whale individual；It is described
Guidance individual is the whale individual nearest from current whale in all whale individuals better than current whale；If it exists, into step
Rapid S3；Otherwise S5 is entered step；
S3 generates the copy whale of current whale, and the guidance individual of current whale guides the copy according to individual movement rule
Whale is moved, and is calculated the fitness value of the copy whale and is compared；If the fitness value of the copy whale is less than current whale
The fitness value of fish then replaces current whale with the copy whale, and sets 0 for the iteration count of current whale, enters
Step S7；Otherwise S4 is entered step；
The iteration count value of the more current whale of S4 and the size of whale group stability threshold；If the iteration of current whale
Counter Value is less than whale group stability threshold, then the iteration count value is added 1, enter step S7；Otherwise it reinitializes
The position of current whale, calculates the fitness value of the whale, enters step S7；
S5 generates the copy of current whale, after executing neighborhood search to the copy, calculates the fitness value of the copy and ratio
Compared with；If the fitness value of the copy is less than the fitness value of current whale, current whale is replaced with into the copy, and will be current
The corresponding whale iteration count value of whale is set as 0, enters step S7；Otherwise S6 is entered step；
If the iteration count of the current whale of S6 is less than whale group stability threshold, 1 is added to the iteration count, is entered
Step S7；Otherwise the size of the fitness value of the fitness value and globally optimal solution of more current whale, if current whale is suitable
Angle value is answered to be less than the fitness value of globally optimal solution, then the fitness value for updating globally optimal solution is the fitness of current whale
Value updates globally optimal solution, initializes the position of current whale, and calculate the fitness value of the whale, enter step S7；If no
Less than then reinitializing the position of current whale, and the fitness value of the whale is calculated, enters step S7；
If the current whale of S7 is not the last one individual of whale group, it is current whale in whale group that current whale, which is arranged,
Next whale individual, enter step S2；If current whale is the last one individual of whale group, judge whether it meets end
Only condition enters step S8 if meeting；Otherwise the individual that current whale is whale group is set, S2 is entered step；
S8 judges whether there is whale individual more better than globally optimal solution in last generation population, there is the then current overall situation of substitution
Optimal solution obtains final globally optimal solution, as optimal sub-clustering scheme.
One as technical solution of the present invention is preferred, it is preferred to use random initializtion mode.
One as technical solution of the present invention is preferred, the calculation formula of fitness are as follows:
F=ω σL+(1-ω)·σD·μD
In formula, ω is weight coefficient, σLFor the standard deviation of each leader cluster node lifetime, μD、σDRespectively indicate ordinary sensors
Mean value and variance of the node to its leader cluster node distance.
In technical solution of the present invention, the encoding scheme of whale individual, first determines following symbol for ease of description
Justice:
Define 1.G: cluster head set, G={ g1,g2,…,gi,…,gm, giIndicate i-th of cluster head, m indicates the number of cluster head
Amount.
Define 2.R: general sensor nodes set, R={ r1,r2,…,ri,…,rn},riIndicate i-th of ordinary sensors
Node, n indicate the quantity of general sensor nodes.
Define 3.RGri: in general sensor nodes riCluster head in communication range numbers set.
The specific design process of individual UVR exposure of whale individual are as follows:
Due to the quantity of general sensor nodes be it is fixed, that most effective fish individual UVR exposure scheme be every
Whale individual represents a kind of complete general sensor nodes to the allocation plan of cluster head.It is as follows:
X=(x1,x2,…,xi,…,xn)
Wherein, X indicates a whale individual, i.e., one solution；xiIndicate general sensor nodes riAffiliated cluster head number,
That is general sensor nodes riDistribute to cluster head xi, xi∈RGri。
Further, fitness function is designed specifically to:
F=ω σL+(1-ω)·σD·μD
Wherein, ω is weight coefficient, and value range is [0,1].
σLIndicate the standard deviation of each leader cluster node lifetime:M is leader cluster node number
Amount, L (gi) indicate leader cluster node giLifetime:μLFor the mean value of all leader cluster node lifetimes,Er(gi) it is leader cluster node giDump energy, Ec(gi) it is leader cluster node giIt is transmitted across in single-wheel data
The energy consumed in journey,niIndicate leader cluster node gi's
Member node quantity, k are to receive data volume size (receiving data bit number), after l indicates that cluster head carries out data fusion, transmission
The size of data packet, EDAIndicate leader cluster node giMerge energy consumed by the data that each general sensor nodes are sent, ETx
Expression leader cluster node transmission data (energy consumed by initial next-hop node of the data package size l) to it,D is the distance between two leader cluster nodes, d0For the door of the distance
Threshold values, k are transmitted data amount size (transmission data bit number), EelecIt indicates to receive in sensor node and transmitting circuit works
When consumed energy, εfsAnd εmpRespectively in free space loss model and multi-path fading channel, in sensor node
Transmit consumed energy when amplifier operation.
σDIndicate all general sensor nodes to its leader cluster node distance mean value and variance:
Wherein, n is the quantity of general sensor nodes, D (ri) it is general sensor nodes riTo affiliated cluster head away from
From μDFor the average value of the distance of each ordinary node to respective cluster head,
The objective function considers the energy consumption balance and general sensor nodes of cluster head and general sensor nodes simultaneously
Wastage in bulk or weight energy.
One as technical solution of the present invention is preferred, and the guidance individual Y of current whale is guided according to individual movement rule
The copy whale X' of current whale carries out mobile process and includes,
S31 setting traversal serial number i=1；
S32 judges i-th of element X' of whale X'iWhether the i-th element Y of whale Y is equal toiIf then entering step
Otherwise S33 enters step 35；
S33 generates a random number P from 0 to 1, if P is less than select probability λ, by X'iIt is assigned a value of Yi, S35 is entered step,
Otherwise S34 is entered step；
S34 randomly chooses a cluster head g ' in the cluster head set in i-th of general sensor nodes communication range, will
X'iIt is assigned a value of g '；
I+1 is assigned to i by S35, if i is not more than | and X'| enters step S32, otherwise enters step S36；
The individual that S36 completes whale X' is mobile.
Preferably as one of technical solution of the present invention, neighborhood search includes,
S41 chooses the best whale individual Ω of fitness value from whale populationb；
S42 calculates whale individual ΩbThe smallest cluster head g of life cyclemin；
S43 is distributing to cluster head gminSensor node set in randomly choose a general sensor nodes ru, general
A cluster head g is randomly choosed in cluster head number set in logical sensor node communication rangev；
S44 is by whale individual ΩbRuDistribute to cluster head gv, complete neighborhood search.
In technical solution of the present invention, first of all for convenient for description neighborhood search strategy, following symbol is defined: definitionTo distribute to cluster head giGeneral sensor nodes number set, i.e., in cluster head giOrdinary sensors in affiliated cluster
The number set of node.The specific design of neighborhood search strategy is as follows:
(131) best individual Ω is found in populationbest；
(132) best individual Ω is calculatedbestThe middle the smallest cluster head g of life cyclemin；
(133) gatheringOne general sensor nodes r of interior random selectionu, from setMiddle random selection one
Cluster head gv；
(134) by best individual ΩbestRuDistribute to cluster head gv, complete neighborhood search.
One as technical solution of the present invention is preferred, it is preferred to use discrete individual coding mode carries out sensor node
Coding.
The whale individual UVR exposure of whale group designs, and using discrete individual UVR exposure mode, i.e. a whale individual indicates one
Paths, X=(s, x2, x3 ..., xi..., d) in formula, X indicates a whale individual, i.e., one solution namely a paths；S table
Show source node, x2Indicate second node in path, xiIndicate that i-th of node in path, d indicate root node.
One as technical solution of the present invention is preferred, and the distance between whale individual preferably uses Hamming distance to be counted
It calculates.Hamming distance is the number of two individual corresponding position different values.I.e. to all elements corresponding position in two individuals into
Row XOR operation, the number that statistical result is 1, the number are the value of Hamming distance.
In general, through the invention it is contemplated above technical scheme is compared with the prior art, have below beneficial to effect
Fruit:
1) method of technical solution of the present invention introduces improved whale group algorithm in efficiency optimization sub-clustering problem, utilizes
Whale group algorithm is improved in the solution ability of multi-modal optimization problem, wireless sensor network efficiency is solved and optimizes sub-clustering problem；Together
When, in the cluster-level routing stage, the routing algorithm of introducing can energy consumption of the active balance leader cluster node when forwarding data, from
And be conducive to further extend network lifecycle.
2) method of technical solution of the present invention encodes whale individual using discretization coding mode, for wireless
Sensor network efficiency optimization sub-clustering problem proposes a kind of discretization coding mode.Make changing for algorithm by improving WSA-IC algorithm
For formula discretization, makes the WSA-IC algorithm for solving continuous optimization problems, be suitable for dispersed problem.
3) method of technical solution of the present invention, constructs suitable fitness function, has both considered the residue of leader cluster node
Energy scale is poor, it is contemplated that the standard deviation of the energy consumption of general sensor nodes and energy consumption.Cluster algorithm passes through this fitness letter
Number enables the energy consumption of the solution acquired balanced leader cluster node and general sensor nodes, while also reducing general sensor nodes
Energy consumption, to extend the life cycle of entire wireless sensor.
4) method of technical solution of the present invention joined Hamming distance algorithm for the coding mode of whale individual to count
The distance between individual is calculated, and neighborhood search strategy is introduced to the best individual of current population in an iterative process, is conducive to reinforce
Algorithm refuses search capability, accelerates the optimal solution for converging to the individual region, improves whale group convergence speed of the algorithm
And the operational efficiency of algorithm.
Detailed description of the invention
Fig. 1 is the sub-clustering schematic diagram of hierarchical wireless sensor network in the prior art；
Fig. 2 is the flow diagram that wireless sensor network efficiency optimizes cluster-dividing method in the embodiment of the present invention；
Fig. 3 is the flow diagram of netinit and network information convergence in the embodiment of the present invention；
Fig. 4 is that efficiency optimizes the flow diagram that sub-clustering calculates in the embodiment of the present invention；
Fig. 5 is simple wireless sensor network schematic diagram in the embodiment of the present invention；
Fig. 6 is the flow diagram that whale individual is mobile in the embodiment of the present invention；
Fig. 7 is the flow diagram of optimal whale individual neighborhood search in the embodiment of the present invention；
Fig. 8 is optimal whale individual neighborhood search rough schematic in the embodiment of the present invention；Wherein, Fig. 8 (a) is this implementation
The best individual schematic diagram of population before progress neighborhood search in example；Fig. 8 (b) is that the individual that population is best after carrying out neighborhood search shows
It is intended to
Fig. 9 (a) is to be generated at random in the embodiment of the present invention including 100 general sensor nodes and 30 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (200,100))；
Fig. 9 (b) is to be generated at random in the embodiment of the present invention including 200 general sensor nodes and 30 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (200,100))；
Fig. 9 (c) is to be generated at random in the embodiment of the present invention including 300 general sensor nodes and 30 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (200,100))；
Fig. 9 (d) is to be generated at random in the embodiment of the present invention including 400 general sensor nodes and 30 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (200,100))；
Fig. 9 (e) is to be generated at random in the embodiment of the present invention including 200 general sensor nodes and 50 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (200,100))；
Fig. 9 (f) is to be generated at random in the embodiment of the present invention including 300 general sensor nodes and 50 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (200,100))；
Fig. 9 (g) is to be generated at random in the embodiment of the present invention including 400 general sensor nodes and 50 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (200,100))；
Fig. 9 (h) is to be generated at random in the embodiment of the present invention including 500 general sensor nodes and 50 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (200,100))；
Fig. 9 (i) is to be generated at random in the embodiment of the present invention including 100 general sensor nodes and 30 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (100,100))；
Fig. 9 (j) is to be generated at random in the embodiment of the present invention including 200 general sensor nodes and 30 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (100,100))；
Fig. 9 (k) is to be generated at random in the embodiment of the present invention including 300 general sensor nodes and 30 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (100,100))；
Fig. 9 (l) is to be generated at random in the embodiment of the present invention including 400 general sensor nodes and 30 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (100,100))；
Fig. 9 (m) is to be generated at random in the embodiment of the present invention including 200 general sensor nodes and 50 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (100,100))；
Fig. 9 (n) is to be generated at random in the embodiment of the present invention including 300 general sensor nodes and 50 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (100,100))；
Fig. 9 (o) is to be generated at random in the embodiment of the present invention including 400 general sensor nodes and 50 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (100,100))；
Fig. 9 (p) is to be generated at random in the embodiment of the present invention including 500 general sensor nodes and 50 leader cluster nodes
The wireless sensor network of (the wherein position that aggregation node is located at coordinate (100,100))；
Figure 10 (a) be the embodiment of the present invention in experiment scene WSN#1 embodiment 3 in, WSA-ICC algorithm of the invention with
The fitness convergent contrast and experiment of DECA, IHSC algorithm；
Figure 10 (b) is WSA-ICC algorithm of the invention in the embodiment of the present invention in the embodiment 11 of experiment scene WSN#2
With the fitness convergent contrast and experiment of DECA, IHSC algorithm；
Figure 10 (c) be the embodiment of the present invention in experiment scene WSN#1 embodiment 8 in, WSA-ICC algorithm of the invention with
The fitness convergent contrast and experiment of DECA, IHSC algorithm；
Figure 10 (d) is WSA-ICC algorithm of the invention in the embodiment of the present invention in the embodiment 16 of experiment scene WSN#2
With the fitness convergent contrast and experiment of DECA, IHSC algorithm；
Figure 11 (a) is in the embodiment of the present invention in the Examples 1 to 4 of experiment scene WSN#1, and WSA-ICC of the invention is calculated
The network lifecycle contrast and experiment of method and EELBCA, DECA, IHSC algorithm；
Figure 11 (b) is in the embodiment of the present invention in the embodiment 5~8 of experiment scene WSN#1, and WSA-ICC of the invention is calculated
The network lifecycle contrast and experiment of method and EELBCA, DECA, IHSC algorithm；
Figure 11 (c) be the embodiment of the present invention in experiment scene WSN#2 embodiment 9~12 in, WSA-ICC of the invention with
The network lifecycle contrast and experiment of EELBCA, DECA, IHSC；
Figure 11 (d) is WSA-ICC of the invention in the embodiment of the present invention in the embodiment 13~16 of experiment scene WSN#2
The network lifecycle contrast and experiment of algorithm and EELBCA, DECA, IHSC algorithm；
Figure 12 (a) is in the embodiment of the present invention in the Examples 1 to 4 of experiment scene WSN#1, and WSA-ICC of the invention is calculated
The RLGD-RFGD contrast and experiment of method and EELBCA, DECA, IHSC algorithm；
Figure 12 (b) is in the embodiment of the present invention in the embodiment 5~8 of experiment scene WSN#1, and WSA-ICC of the invention is calculated
The RLGD-RFGD contrast and experiment of method and EELBCA, DECA, IHSC algorithm；
Figure 12 (c) is in the embodiment of the present invention in the embodiment 9~12 of experiment scene WSN#2, and WSA-ICC of the invention is calculated
The RLGD-RFGD contrast and experiment of method and EELBCA, DECA, IHSC algorithm；
Figure 12 (d) is WSA-ICC of the invention in the embodiment of the present invention in the embodiment 13~16 of experiment scene WSN#2
The RLGD-RFGD contrast and experiment of algorithm and EELBCA, DECA, IHSC algorithm；
Figure 13 (a) be the embodiment of the present invention in experiment scene WSN#1 embodiment 3 in, WSA-ICC algorithm of the invention with
The experiment of energy consumption result of EELBCA, DECA, IHSC algorithm；
Figure 13 (b) is WSA-ICC algorithm of the invention in the embodiment of the present invention in the embodiment 11 of experiment scene WSN#2
With the experiment of energy consumption result of EELBCA, DECA, IHSC algorithm；
Figure 13 (c) be the embodiment of the present invention in experiment scene WSN#1 embodiment 8 in, WSA-ICC algorithm of the invention with
The experiment of energy consumption result of EELBCA, DECA, IHSC algorithm；
Figure 13 (d) is WSA-ICC algorithm of the invention in the embodiment of the present invention in the embodiment 16 of experiment scene WSN#2
With the experiment of energy consumption result of EELBCA, DECA, IHSC algorithm；
Figure 14 (a) be the embodiment of the present invention in experiment scene WSN#1 embodiment 3 in, WSA-ICC algorithm of the invention with
The experiment of energy consumption result of EELBCA, DECA, IHSC algorithm；
Figure 14 (b) is WSA-ICC algorithm of the invention in the embodiment of the present invention in the embodiment 11 of experiment scene WSN#2
With the experiment of energy consumption result of EELBCA, DECA, IHSC algorithm；
Figure 14 (c) be the embodiment of the present invention in experiment scene WSN#1 embodiment 8 in, WSA-ICC algorithm of the invention with
The experiment of energy consumption result of EELBCA, DECA, IHSC algorithm；
Figure 14 (d) is WSA-ICC algorithm of the invention in the embodiment of the present invention in the embodiment 16 of experiment scene WSN#2
With the experiment of energy consumption result of EELBCA, DECA, IHSC algorithm.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.As long as in addition, technical characteristic involved in the various embodiments of the present invention described below
Not constituting a conflict with each other can be combined with each other.The present invention is described in more detail With reference to embodiment.
It is that sub-clustering is optimized based on the wireless sensor network efficiency for improving whale group algorithm in the present embodiment as shown in Figure 2
The flow chart of method for routing, converged including netinit and the network information and etc..Below for the side of the present embodiment
It is further detailed in the specific steps of method.
(1) netinit and the network information converge step, as shown in figure 3, specifically including:
A. unique identification determines: to all biographies in network including aggregation node, leader cluster node and general sensor nodes
Sensor node is numbered, as their unique identification.
B. it calculates netinit configuration information: according to known nodal distance information, using shortest path Dijkstra (enlightening
Jie Sitela) algorithm obtains all nodes to the shortest path of aggregation node, is used as initialization default route using the program；
According to known nodal distance information, the nearest leader cluster node of all ordinary nodes is found, which is the ordinary node
Initialization default cluster head, using the program as initialization default sub-clustering.
C. netinit configure: aggregation node by initial configuration information be sent in sequence to respectively all leader cluster nodes and
General sensor nodes.Leader cluster node and general sensor nodes, after receiving configuration information, the initialization for completing itself is matched
It sets.
D. the network information converges: ordinary node receives the initial configuration information, and it is surplus to send it to its default leader cluster node
Remaining information about power.All information from general sensor nodes and the dump energy information of itself are carried out data by leader cluster node
After fusion, aggregation node is sent it to according to initialization default route, completes the convergence of whole network information.
(2) clustering routing calculates step and carries out net to entire sensor network according to the dump energy information of all nodes
The optimization of network clustering routing calculates and configuration, specifically includes:
A. energy-optimised router-level topology:
Base station, which is used, calculates each leader cluster node based on the wireless sensor network routing method for improving whale group algorithm
Optimal forward-path carry out router-level topology wherein leader cluster node of the invention is considered the sensor node of the above method, obtain
To all leader cluster nodes to the optimal routing plan of aggregation node；
B. efficiency optimization sub-clustering calculates:
Base station calculates wireless sensor network efficiency and optimizes sub-clustering problem, obtain optimal solution using whale group algorithm is improved,
Finally obtain the optimal sub-clustering scheme of entire sensor network；
Efficiency optimizes sub-clustering calculation process as shown in figure 4, including following sub-steps:
Sub-step 1, setting whale group algorithm parameter, including stability threshold Ts=200+n/2, wherein n is currently common
Sensor node quantity, Population Size (i.e. whale individual amount) | Ω |=20 and select probability λ=0.5；
Sub-step 2, each whale individual for initializing whale group algorithm, obtain initial whale population:
Ω={ Ω1, Ω2..., Ωi..., Ω|Ω|}；
Wherein, | Ω | it is whale individual amount, ΩiFor i-th of whale individual；
The wherein individual UVR exposure of whale individual, specific design are as follows:
The encoding scheme of whale individual for ease of description, is defined following symbol:
Define 1.G: cluster head set, G={ g1,g2,…,gi,…,gm, giIndicate i-th of cluster head, m indicates the number of cluster head
Amount.
Define 2.R: general sensor nodes set, R={ r1,r2,…,ri,…,rn},riIndicate i-th of ordinary sensors
Node, n indicate the quantity of general sensor nodes.
Define 3.RGri: in general sensor nodes riCluster head in communication range numbers set.
Due to the quantity of general sensor nodes be it is fixed, most effective whale individual UVR exposure scheme be every whale
Fish individual represents a kind of complete general sensor nodes to the allocation plan of cluster head.It is as follows:
X=(x1,x2..., xi..., xn)
Wherein, X indicates a whale individual, i.e., one solution；xiIndicate general sensor nodes riAffiliated cluster head number,
That is general sensor nodes riDistribute to cluster head xi,
As shown in figure 5, including 5 cluster heads and 12 general sensor nodes, i.e. G={ g in network1,g2,g3,g4,g5, R
={ r1,r2,r3,r4,r5,r6,r7,r8,r9,r10,r11,r12}.Connecting line meaning in figure, between general sensor nodes and cluster head
Taste two sensor nodes can be in communication with each other.As seen from the figure, Assuming that whale individual X=(2,
3,4,2,3,5,1,3, Isosorbide-5-Nitrae, 5,5), it is meant that general sensor nodes r1Distribute to cluster head g2, sensor node r2Distribute to cluster
Head g3, and so on
Wherein present example uses random initializtion strategy to whale group individual；
Sub-step 3 passes through fitness function, calculates the fitness value of each whale individual:
F={ f (Ω1), f (Ω2) ..., f (Ωi) ..., f (Ω|Ω|)}
Wherein, f (Ωi) be i-th of whale individual fitness；
Wherein fitness function is designed specifically to:
F=ω σL+(1-ω)·σD·μD
Wherein, ω is weight coefficient, and value range is [0,1].
σLIndicate the standard deviation of each leader cluster node lifetime:M is leader cluster node number
Amount, L (gi) indicate leader cluster node giLifetime:μLFor the mean value of all leader cluster node lifetimes,Er(gi) it is leader cluster node giDump energy, Ec(gi) it is leader cluster node giIt is transmitted across in single-wheel data
The energy consumed in journey,niIndicate leader cluster node gi's
Member node quantity, k are to receive data volume size (receiving data bit number), after l indicates that cluster head carries out data fusion, transmission
The size of data packet, EDAIndicate leader cluster node giMerge energy consumed by the data that each general sensor nodes are sent, ETx
Expression leader cluster node transmission data (energy consumed by initial next-hop node of the data package size l) to it,D is the distance between two leader cluster nodes, d0For the gate valve of the distance
Value, k are transmitted data amount size (transmission data bit number), EelecWhen indicating to be received in sensor node with transmitting circuit work
Consumed energy, εfsAnd εmpRespectively in free space loss model and multi-path fading channel, passed in sensor node
Consumed energy when sending amplifier operation.
σDIndicate all general sensor nodes to its leader cluster node distance mean value and variance:
Wherein, n is the quantity of general sensor nodes, D (ri) it is general sensor nodes riTo affiliated cluster head away from
From μDFor the average value of the distance of each ordinary node to respective cluster head,
The objective function considers the energy consumption balance and general sensor nodes of cluster head and general sensor nodes simultaneously
Wastage in bulk or weight energy.
Sub-step 4, setting traversal serial number i=1；
Sub-step 5, the distance according to fitness value and whale between individual, find from whale Ωi" more excellent and nearest "
Whale Y, if Y exists, rotor step 6, otherwise rotor step 12；
The wherein distance between whale individual, is calculated using Hamming distance, specifically:
Hamming distance is the number of two individual corresponding position different values.Position is corresponded to all elements in two individuals
Carry out XOR operation, the number that statistical result is 1 are set, which is the value of Hamming distance.
Assuming that two whale individual X1、X2As follows, X1And X2There are 5 elements not identical, so, X1And X2Between the Chinese
Prescribed distance is 5.
Sub-step 6 generates whale ΩiCopy X'；
Sub-step 7, according to individual movement rule, guide whale X' to be moved by whale Y；
Wherein individual movement rule, detailed process is as shown in fig. 6, include the following steps:
(7.1) setting traversal serial number i=1；
(7.2) if X'iNot equal to Yi, then (7.3) are gone to step, (7.5) are otherwise gone to step；
(7.3) one 0 to 1 random number P is generated, if P is less than select probability λ, by Xi' it is assigned a value of Yi, otherwise go to step
(7.4)；
(7.4) gatheringOne cluster head g ' of interior random selection, by Xi' it is assigned a value of g '；
(7.5) i+1 is assigned to i, if i is not more than | X'| goes to step (7.2), otherwise goes to step (7.6)；
(7.6) individual for completing whale X' is mobile；
As follows, whale ΩuCopy X' and its " more excellent and nearest " whale Y, X' and Y first element not phase
Together, therefore, first element of Y is assigned to X''s because of P < λ by the random number P=0.2 generated between one 0 to 1
First element, at this time (2,3,4,2,3,3,1,3,1,2,5,4) X'=；X' is identical as second element of Y, does not need to X'
Carry out any operation；The third element of X' and Y is not identical, therefore, generates the random number P=0.6 between one 0 to 1, because
P > λ, so from setOne value (being assumed to be 1) of interior random selection is assigned to the third element of X', at this time X'=(2,
3,1,2,3,3,1,3,1,2,5,4)；The rest may be inferred, until having judged the last one element, completes corresponding operation, X' is just tied
Movement of the beam under Y guidance.Assuming that the X'=(2,3,4,2,3,5,1,1,1,2,5,5) after mobile, it is seen then that between X' and Y
Distance become 3 by original 6.
Sub-step 8, the fitness value f (X') for calculating X', if f (X') is less than f (Ωi), then rotor step 9, otherwise rotor
Step 10；
Sub-step 9, by i-th of whale Ω in whale groupiIt is replaced by whale copy X', by whale iteration count Ωi.c
It is set as 0, rotor step 18；
If the iteration count Ω of sub-step 10, i-th whalei.cNot equal to Ts, then to iteration count Ωi.cAdd 1,
Rotor step 18, otherwise rotor step 11；
Sub-step 11 reinitializes i-th of whale Ωi, and calculate the fitness value f (Ω of the whalei), rotor step
18；
Sub-step 12 generates i-th of whale ΩiCopy X "；
Sub-step 13 executes neighborhood search to whale copy X ", and calculates the fitness value f of the copy after neighborhood search
(X "), if f (X ") is less than f (Ωi), then rotor step 14, otherwise rotor step 15；
Wherein, field search strategy for ease of description, is defined following symbol:
Define 4.Distribute to cluster head giGeneral sensor nodes number set, i.e., in cluster head giIn affiliated cluster
General sensor nodes number set.
Neighborhood search strategy detailed process is as shown in fig. 7, comprises following steps:
(13.1) best individual Ω is found in populationbest；
(13.2) best individual Ω is calculatedbestThe middle the smallest cluster head g of life cyclemin；
(13.3) gatheringOne general sensor nodes r of interior random selectionu, from setMiddle random selection one
A cluster head gv；
(13.4) by best individual ΩbestRuDistribute to cluster head gv, complete neighborhood search；
If good individual Ω in populationbestAs shown in Fig. 8 (a), as seen from the figure, Lifetime the smallest cluster head is g5, L (g5)=720.Therefore,
From setOne general sensor nodes number of interior random selection, it is assumed that node 8 is selected, then from setIt is interior with
Machine selects a cluster head number, it is assumed that 3 is selected.Finally, by general sensor nodes r8From cluster head g5The cluster at place is deleted, will
r8Distribute to cluster head g3, as shown in Fig. 8 (b), cluster head g5Lifetime increase 790 from 720, cluster head g3Lifetime from 890
820 are reduced to, the standard deviation of each cluster head lifetime is reduced.
Sub-step 14, by i-th of whale Ω in whale groupiIt is replaced by whale copy X ", and by the whale iteration count
Ωi.cIt is set as 0, rotor step 18；
If the iteration count Ω of sub-step 15, i-th whalei.cNot equal to Ts, then to iteration count Ωi.cAdd 1,
Rotor step 18, otherwise rotor step 16；
If the fitness value f (Ω of sub-step 16, i-th whalei) it is less than the fitness value f of globally optimal solutiongbest, then will
fgbestIt is set as f (Ωi), Ω is set by globally optimal solution GBesti；
Sub-step 17 reinitializes i-th of whale ΩiPosition, and calculate the fitness value f (Ω of the whalei)；
I+1 is assigned to i by sub-step 18, if i is less than whale individual amount | Ω |, rotor step 5, otherwise rotor walks
Rapid 19；
Sub-step 19 judges whether termination condition meets, if meeting rotor step 20, otherwise rotor step 4；
Sub-step 20 judges whether there is whale individual more better than GBest in last generation population, has, substitutes GBest,
GBest is current optimal solution, i.e., optimal routing plan.
(3) clustering routing configuration step, specific as follows:
Aggregation node sends optimal routing configuration information and most to all leader cluster nodes using above-mentioned optimal routing plan
Optimal clustering configuration information configures the routing table of transfer routing node in transmission process；Leader cluster node receives sub-clustering configuration information
Afterwards, sub-clustering configuration information is sent after ordinary node stores corresponding sub-clustering configuration information to corresponding general sensor nodes to add
Enter into sub-clustering.
(4) information collection and converging information step, all wireless sensor nodes carry out information collection and processing application letter
Breath is completed to send to the information of aggregation node according to the clustering routing of configuration, specific as follows:
General sensor nodes carry out information collection and processing application message, are processed into data packet, and it is current
Dump energy information is added in data packet, which is forwarded a packet to its corresponding leader cluster node；Leader cluster node carries out information
Acquisition and processing application message, are processed into data packet, and its current remaining information is added in data packet；Leader cluster node
The data fusion of sub-clustering where completing will send aggregation node for all data according to configured routing, and complete information and adopt
Collection and convergence.The step constantly executes periodically, realizes the acquisition information update of wireless sensor network.
This programme, it is preferable that in the information collection and converging information step, after convergence wheel number is more than 20 times, again
It carries out efficiency and optimizes router-level topology, after forming new routing plan, re-execute the steps (3) (4), complete the dynamic of routing plan
It updates.
This programme, it is preferable that in the information collection and converging information step, when some section occurs in wireless sensor network
After point depleted of energy, re-starts efficiency optimization sub-clustering and calculates, after forming new sub-clustering scheme, re-execute the steps (3) (4),
The dynamic for completing sub-clustering scheme updates.
The effect of the present invention program can further be verified by following emulation experiment and relatively.
Experimentation:
In the specific emulation experiment of the present embodiment, preferably by all the sensors inserting knot in 200 × 200m2Region
In.Meanwhile in order to which the effect to emulation experiment is comprehensively compared, the present embodiment devises two kinds of scenes, the first scene is
WS, aggregation node coordinate are (200,100), the i.e. leftmost side of simulating area；Second of scene is WSN#2, aggregation node
Coordinate is (100,100), the i.e. center of simulating area.That is, by 16 specific embodiments in following table 1 to this reality
The wireless sensor network efficiency optimization cluster-dividing method applied in example is further verified.
1 wireless sensor network embodiment scene of table
It is 16 wireless sensor networks generated at random according to above-mentioned network node data as shown in Fig. 9 (a)~(p),
By the cluster-dividing method and energy efficient Load balance clustering method, i.e. EELBCA (Energy Efficient in the present invention
Load-Balanced Clustering Algorithm), the cluster-dividing method based on differential evolution algorithm, i.e. DECA
(Differential Evolution based Clustering Algorithm), based on point for improving harmonic search algorithm
Cluster method IHSC (Improved Harmony Search based Clustering Algorithm) is restrained for fitness
Situation (speed of the quality and algorithmic statement of algorithmic statement to optimal solution to optimal solution), the life cycle (life in the present invention
Period is when there is first cluster head depleted of energy, and network sends the wheel number of data packet), RLGD-RFGD is (i.e. from first cluster head
The wheel number that depleted of energy is passed through to a last cluster head depleted of energy), (every wheel data have been sent surviving node number, present energy
There are no the general sensor nodes quantity exhausted) and energy consumption (until terminating when front-wheel number, the energy of all nodes consumption is gone through
History summation) it is compared.In the present embodiment, Simulation Program language be C++, allocation of computer are as follows: dominant frequency be 3.2GHz and
The intel I5-3470QM processor of 3.6GHz, 4GB memory, 10 64 bit manipulation system of windows.Simultaneously for the ease of reality
It tests and compares and map, the cluster-dividing method in the present invention is denoted as WSA-ICC.
For above-mentioned 16 specific embodiments, network simulation survey is carried out one by one using the cluster-dividing method in the present embodiment
Examination.Each sensor node in 16 wireless sensor networks is respectively with some cycles to the aggregation node of respective place network
Data packet is sent, until meeting the measuring condition of experiment parameter.
The emulation experiment network parameter is as shown in table 2, and wherein ω is the weight coefficient in objective function.
The setting of 2 emulation experiment network parameter of table
| Parameter | Value |
| Communication range | 150m |
| The data package size that general sensor nodes are sent | 256bits |
| Eelec | 50nJ/bit |
| εfs | 0.1nJ/bit/m2 |
| εmp | 0.000013nJ/bit/m4 |
| d0 | 87.0m |
| ω | 0.7 |
| Function evaluates number | 500000 |
| Independent operating number | 10 |
The initialization energy of each example interior joint of emulation experiment is as shown in table 3.
The primary power (unit: J) of each example interior joint of table 3
| Example | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 |
| Cluster head | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 |
| Ordinary node | 0.35 | 0.25 | 0.25 | 0.25 | 0.25 | 0.25 | 0.25 | 0.25 | 0.5 | 0.35 | 0.35 | 0.25 | 0.5 | 0.35 | 0.35 | 0.35 |
The emulation experiment algorithm parameter is as shown in table 4, wherein PopSize is Population Size；Cr is crossover probability；F is
Scale factor；HMS is harmony library size；HMCRmin、HMCRmaxThe respectively minimum value and maximum value of harmony library select probability；|
Ω | it is Population Size；N is the general sensor nodes quantity currently survived.
The setting of 4 algorithm parameter of table
| Algorithm | Parameter |
| DECA | PopSize=100, Cr=0.7, F=0.5 |
| IHSC | HMS=6, HMCRmin=0.2, HMCRmax=0.99 |
| WSA-ICC | | Ω |=20, Ts=200+n/2, λ=0.5 |
Experimental result:
Fitness convergent compares: by the example 3 of WSA-ICC, DECA and IHSC in above-mentioned wireless sensor network
Implementation comparison is carried out with example 11, example 8 and example 16.Figure 10 (a)~Figure 10 (d) is examples detailed above 3, example 11, example 8
With the fitness convergent experimental data comparing result of example 16.It can be seen from the figure that the present invention mentions in all examples
WSA-ICC algorithm out is either in terms of fitness convergence rate, or all than it in terms of the quality of the optimal solution acquired
The performance of its algorithm is much better.This is to have benefited from the outstanding rule of iteration of WSA-IC algorithm: every whale individual its " it is more excellent and
It is moved under the guidance of whale recently ", if it find that better new position, then the whale is moved to the new position；
Otherwise, which keeps as you were.In this way, WSA-IC algorithm can rapidly converge to the extreme point near each individual.Meanwhile
The precision of solution can be improved in the neighborhood search strategy that the present invention is added.Importantly, WSA-IC algorithm can be in iterative process
It is middle to identify and jump out the extreme point found, to improve the ability of searching optimum of algorithm as much as possible, greatly improves and ask
Solve the ability of globally optimal solution.
Network lifecycle compares: by WSA-ICC, EELBCA, DECA and IHSC in above-mentioned wireless sensor network into
Row is implemented, and experiment termination condition is first cluster head depleted of energy.Figure 11 (a)~Figure 11 (d) is respectively Examples 1 to 4, implements
The network lifecycle experimental data comparing result of example 5~8, embodiment 9~12 and embodiment 13~16.It can from experimental result
Out, in most examples, WSA-ICC algorithm proposed by the present invention can obtain longer life cycle than other algorithms；
Especially for 2#WSN, the advantage of WSA-ICC algorithm is more obvious.EELBCA obtains the smallest life in many examples
Period is because in EELBCA, and cluster head is that member node is distributed according to rank and file's number of nodes, rather than according to cluster head
Dump energy distribute rank and file's node, then, dump energy less cluster head and the more cluster head of dump energy possess
The much the same rank and file's number of nodes of quantity exhausts energy so as to cause the less cluster head of dump energy quickly.EELBCA
Much the same life cycle can be obtained with other algorithms in some examples, be because these cluster algorithms all use WSA-
ICR routing algorithm calculates the routing of cluster head, so that the transmission phase in data packet balances the energy consumption of cluster head significantly.WSA-
Why ICC algorithm does very well in terms of life cycle than other comparison algorithms, is to have benefited from the outstanding solution of WSA-ICC algorithm
Performance, it can find better solution than other algorithms, rank and file's node be distributed according to the dump energy of cluster head, so that remaining
The least cluster head of energy possesses rank and file's node of negligible amounts, extends their lifetime as much as possible, to extend net
The life cycle of network.
RLGD-RFGD compares: WSA-ICC, EELBCA, DECA and IHSC being carried out in above-mentioned wireless sensor network real
It applies, experiment termination condition is the last one cluster head depleted of energy.Figure 12 (a)~Figure 12 (d) is respectively Examples 1 to 4, embodiment
5~8, the RLGD-RFGD experimental data comparing result of embodiment 9~12 and embodiment 13~16." RLGD-RFGD " index is to use
Evaluate performance of each cluster-dividing method in terms of balancing cluster head energy consumption, this refers to that target value is smaller, that is, from first cluster head energy
The wheel number that amount is depleted to the last one cluster head depleted of energy process is smaller, it means that cluster-dividing method is in terms of balancing cluster head energy consumption
Performance it is better.From experimental result as can be seen that DECA obtains worst RLGD-RFGD in most examples, especially
It is the example 5~example 8 more for cluster head (search space is bigger).This is because the solution ability of DECA is poor, from DECA
Fitness convergence performance as can be seen that the last solution that acquires of DECA is second-rate, also mean that cannot the sub-clustering stage very
The lifetime of good each cluster head of balance, so the laod unbalance of cluster head；In addition, from the closer cluster head of aggregation node also need for compared with
Remote cluster head forwards data, if cluster head load imbalance, will continue to widen the standard deviation of cluster head lifetime, so as to cause
RLGD-RFGD becomes larger.And due to outstanding solution ability, WSA-ICC algorithm can obtain well in 1~example of example 8
RLGD-RFGD.In each example of 2#WSN, base station is all located at the center of sensing region, and in the communication range of all cluster heads
It is interior, the stage of data packet is sent in cluster head, the energy consumption of each cluster head is more balanced.As shown in Figure 12 (a), due to cluster head quantity
Less, search space is smaller, and the RLGD-RFGD that each algorithm obtains is not much different；And four more for cluster head in Figure 12 (b)
Example, the performance of WSA-ICC algorithm are more prominent.It is outstanding that this shows that WSA-ICC algorithm has in terms of balanced cluster head energy consumption
Performance.
Surviving node number compares: by WSA-ICC, IHSCRA and EELBCA, DECA and IHSC in above-mentioned wireless sensor network
Implemented in network.Wherein test the embodiment 3 and example 8 that embodiment used is scene WSN#1, the implementation of scene WSN#2
Example 11 and example 16.Figure 13 (a)~Figure 13 (d) is respectively the surviving node number of embodiment 3, example 8, embodiment 11 and example 16
Experimental data comparing result.Surviving node number is for evaluating cluster-dividing method in terms of reducing general sensor nodes energy consumption
Performance, this refers to that target value is bigger, it is meant that cluster-dividing method gets over the lifetime that can extend general sensor nodes.It can from experimental result
To find out, the sub-clustering scheme that IHSC algorithm and EELBCA algorithm are calculated can greatly reduce the energy consumption of general sensor nodes.
It is that the nearest general sensor nodes of selection are added to when distributing general sensor nodes to cluster head for EELBCA algorithm
Current cluster head can be reduced greatly common in this way, most of general sensor nodes can be assigned to nearest or closer cluster head
The energy consumption of sensor node, to extend their lifetime；But this method may make the laod unbalance of cluster head,
Cause the cluster head energy consumption of overload too fast, so as to shorten Network morals.Therefore, EELBCA algorithm is in major part
Worst life cycle is all obtained in example.IHSC algorithm all carries out neighborhood to best harmony after the candidate harmony of each generation
Many general sensor nodes can be assigned to nearest or secondary close cluster head by search, the optimal solution finally acquired, can also pole
The big energy consumption for reducing general sensor nodes, to significantly improve the lifetime of general sensor nodes；But it is received from fitness
Holding back situation can be seen that optimal solution that IHSC algorithm acquires in first index (standard deviation of cluster head lifetime) of objective function
On cannot obtain optimal or close excellent value, so, performance ratio WSA-ICC algorithm of the IHSC algorithm in terms of network lifecycle
Difference.It is that cluster head reasonably distributes general sensor nodes quantity that the main object of the present invention, which is by cluster algorithm, to the greatest extent may be used
The energy consumption of the balanced cluster head in energy ground, extends Network morals.Because having the common biography of many redundancies in practical WSN scene
Sensor node, the depleted of energy of certain general sensor nodes, which will not acquire data, causes too much influence；And the energy of cluster head
Amount exhausts the data that will be lost the acquisition of its rank and file's node, and makes its rank and file's node that farther cluster head be added,
To accelerate the energy consumption of these general sensor nodes.Therefore, there is data acquisition in the lifetime of cluster head most important
Effect, need the energy consumption of as much as possible balanced cluster head.So in the objective function that this chapter is proposed, the mark of cluster head lifetime
The weight coefficient of quasi- difference is set as 0.7, based on the energy consumption of the optimal solution symmetrical cluster head found with Expectation Algorithm, while also can
Balance the energy consumption of general sensor nodes.From the performance of fitness convergent as can be seen that WSA-ICC algorithm is with best
Solution ability.The optimal solution that WSA-ICC algorithm acquires can be to balance based on the energy consumption of cluster head, may be by certain ordinary sensors
Node is added to causes the energy consumption of these general sensor nodes to increase apart from the more cluster heads of farther away dump energy, thus,
Influence performance of the WSA-ICC algorithm in terms of surviving node number.
Energy consumption compares: IHSCRA of the present invention and EELBCA, DECA and IHSC are carried out in fact in above-mentioned wireless sensor network
It applies.Wherein test the embodiment 3 and example 8 that embodiment used is scene WSN#1, the embodiment 11 and example of scene WSN#2
16.Figure 14 (a)~Figure 14 (d) is the experimental data comparing result of embodiment 3,16 energy consumption of example 8, embodiment 11 and example.Energy
Consumption index refers to the energy consumption of whole network, the energy consumption including cluster head and general sensor nodes.From experimental result as can be seen that
WSA-ICC method reduce whole network energy consumption in terms of performance ratio EELBCA and IHSC method it is poor, this also because
The optimal solution that WSA-ICC method acquires is mainly to balance the energy consumption of cluster head, may be added to certain general sensor nodes
Apart from the more cluster heads of farther away dump energy, to increase the energy consumption of general sensor nodes.But it will again be seen that
Performance of the WSA-ICC method in terms of energy consumption is better than DECA method.
In conclusion WSA-ICC method of the present invention is in balanced whole network energy consumption, extension network lifecycle and convergence
Speed etc. all has outstanding performance.
As it will be easily appreciated by one skilled in the art that the foregoing is merely illustrative of the preferred embodiments of the present invention, not to
The limitation present invention, any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should all include
Within protection scope of the present invention.
Claims (7)
1. a kind of wireless sensor network efficiency based on whale group algorithm optimizes cluster-dividing method, which is characterized in that including
Initial configuration information is separately sent to leader cluster node and ordinary node by aggregation node, and to the section of wireless sensor network
Point information carries out convergence acquisition, obtains current network information；
According to current network information, the optimal forward-path of each leader cluster node is calculated, to obtain all leader cluster nodes
To the optimal routing plan of aggregation node, and the optimal sub-clustering scheme of entire sensor network；
According to optimal sub-clustering scheme and optimal routing plan, clustering routing configuration is carried out to entire wireless sensor network；
Sensor node acquisition and processing information, generation includes the data packet of sensor node self rest energy information, and
The data are forwarded a packet into corresponding leader cluster node；Leader cluster node carries out data fusion to place sub-clustering, is matched according to clustering routing
It sets and transmits data to aggregation node, complete converging information acquisition；
Wherein, the optimal sub-clustering scheme is calculated by improved whale algorithm and is obtained, comprising:
Whale group algorithm parameter is arranged in S1, and initializes to each whale individual in whale group algorithm, obtains initial whale
Fingerling group calculates the fitness value of each whale individual；The individual that current whale is whale group is set, is entered step
S2；
S2 determines current whale with the presence or absence of guidance individual according to the distance between fitness value and whale individual；The guidance
Individual is the whale individual nearest from current whale in all whale individuals better than current whale；If it exists, it enters step
S3；Otherwise S5 is entered step；
S3 generates the copy whale of current whale, and the guidance individual of current whale guides the copy whale according to individual movement rule
It is moved, calculate the fitness value of the copy whale and is compared；If the fitness value of the copy whale is less than current whale
Fitness value then replaces current whale with the copy whale, and sets 0 for the iteration count of current whale, enters step
S7；Otherwise S4 is entered step；
The iteration count value of the more current whale of S4 and the size of whale group stability threshold；If the iteration count of current whale
Device value is less than whale group stability threshold, then the iteration count value is added 1, enter step S7；Otherwise it reinitializes current
The position of whale calculates the fitness value of the whale, enters step S7；
S5 generates the copy of current whale, after executing neighborhood search to the copy, calculates the fitness value of the copy and compares；If
The fitness value of the copy is less than the fitness value of current whale, then current whale is replaced with the copy, and by current whale
Corresponding whale iteration count value is set as 0, enters step S7；Otherwise S6 is entered step；
If the iteration count of the current whale of S6 is less than whale group stability threshold, 1 is added to the iteration count, is entered step
S7；Otherwise the size of the fitness value of the fitness value and globally optimal solution of more current whale, if the fitness of current whale
Value is less than the fitness value of globally optimal solution, then the fitness value for updating globally optimal solution is the fitness value of current whale, more
New globally optimal solution, initializes the position of current whale, and calculates the fitness value of the whale, enters step S7；If being not less than
The position of current whale is then reinitialized, and calculates the fitness value of the whale, enters step S7；
If the current whale of S7 is not the last one individual of whale group, be arranged current whale be current whale in whale group under
One whale individual, enters step S2；If current whale is the last one individual of whale group, judge whether it meets termination item
Part enters step S8 if meeting；Otherwise the individual that current whale is whale group is set, S2 is entered step；
S8 judges whether there is whale individual more better than globally optimal solution in last generation population, has, substitutes current global optimum
Solution, obtains final globally optimal solution, as optimal sub-clustering scheme.
2. the wireless sensor network efficiency according to claim 1 based on whale group algorithm optimizes cluster-dividing method, wherein
The initialization uses random initializtion mode.
3. the wireless sensor network efficiency according to claim 1 or 2 based on whale group algorithm optimizes cluster-dividing method,
In, the calculation formula of fitness are as follows:
F=ω σL+(1-ω)·σD·μD
In formula, ω is weight coefficient, σLFor the standard deviation of each leader cluster node lifetime, μD、σDRespectively indicate general sensor nodes
To the mean value and variance of its leader cluster node distance.
4. the wireless sensor network efficiency according to claim 1 or 2 based on whale group algorithm optimizes cluster-dividing method,
In, the guidance individual Y of current whale guides the copy whale X' of current whale to carry out mobile process according to individual movement rule
Including,
S31 setting traversal serial number i=1；
S32 judges i-th of element X' of whale X'iWhether the i-th element Y of whale Y is equal toiIt is no if then entering step S33
Then enter step 35；
S33 generates a random number P from 0 to 1, if P is less than select probability λ, by X'iIt is assigned a value of Yi, S35 is entered step, otherwise
Enter step S34；
S34 randomly chooses a cluster head g ' in the cluster head set in i-th of general sensor nodes communication range, by X'iAssignment
For g '；
I+1 is assigned to i by S35, if i is not more than | and X'| enters step S32, otherwise enters step S36；
The individual that S36 completes whale X' is mobile.
5. the wireless sensor network efficiency according to claim 1 or 2 based on whale group algorithm optimizes cluster-dividing method,
In, the neighborhood search includes,
S41 chooses the best whale individual Ω of fitness value from whale populationb；
S42 calculates whale individual ΩbThe smallest cluster head g of life cyclemin；
S43 is distributing to cluster head gminSensor node set in randomly choose a general sensor nodes ru, commonly passing
A cluster head g is randomly choosed in cluster head number set in sensor node communication rangev；
S44 is by whale individual ΩbRuDistribute to cluster head gv, complete neighborhood search.
6. the wireless sensor network efficiency according to claim 1 or 2 based on whale group algorithm optimizes cluster-dividing method,
In, sensor node is encoded using discrete individual coding mode.
7. the wireless sensor network efficiency according to claim 1 or 2 based on whale group algorithm optimizes cluster-dividing method,
In, the distance between whale individual is calculated using Hamming distance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711351966.XA CN108112049B (en) | 2017-12-15 | 2017-12-15 | A kind of wireless sensor network efficiency optimization cluster-dividing method based on whale group algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711351966.XA CN108112049B (en) | 2017-12-15 | 2017-12-15 | A kind of wireless sensor network efficiency optimization cluster-dividing method based on whale group algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108112049A CN108112049A (en) | 2018-06-01 |
| CN108112049B true CN108112049B (en) | 2018-12-07 |
Family
ID=62216343
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711351966.XA Active CN108112049B (en) | 2017-12-15 | 2017-12-15 | A kind of wireless sensor network efficiency optimization cluster-dividing method based on whale group algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108112049B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108934054B (en) * | 2018-06-19 | 2021-07-16 | 昆明理工大学 | Ad-hoc network clustering algorithm based on node data values |
| CN109345005A (en) * | 2018-09-12 | 2019-02-15 | 中国电力科学研究院有限公司 | A kind of integrated energy system multidimensional optimization method based on improvement whale algorithm |
| CN109168502A (en) * | 2018-09-29 | 2019-01-11 | 东莞绿邦智能科技有限公司 | A kind of farmland management system based on technology of Internet of things |
| CN109931903B (en) * | 2019-02-26 | 2021-07-06 | 上海大学 | Cylindricity evaluation method based on improved whale optimization algorithm |
| CN110363214B (en) * | 2019-05-31 | 2021-08-03 | 浙江工业大学 | GWA-SVM-based contact state recognition method for robot assembly |
| CN111542011B (en) * | 2020-04-27 | 2021-09-24 | 中山大学 | Layered wireless sensor network clustering routing method based on particle swarm optimization |
| CN111526555B (en) * | 2020-05-25 | 2022-07-26 | 南京邮电大学 | Multi-hop routing path selection method based on genetic algorithm |
| CN112492661B (en) * | 2020-12-10 | 2022-04-15 | 中南民族大学 | Wireless sensor network clustering routing method based on improved sparrow search algorithm |
| CN112672396B (en) * | 2020-12-10 | 2022-06-14 | 中南民族大学 | Wireless sensor network mobile node clustering method based on improved whale algorithm |
| CN113179544B (en) * | 2021-04-26 | 2022-09-09 | 国网四川省电力公司经济技术研究院 | Wireless sensing clustering routing cluster head selection method applied to power Internet of things |
| CN113891426A (en) * | 2021-09-29 | 2022-01-04 | 光大科技有限公司 | Distributed multi-node networking method and device |
| CN114339785B (en) * | 2021-12-15 | 2024-02-13 | 红云红河烟草(集团)有限责任公司 | Raw smoke maintenance wireless sensor layout optimization method based on group intelligence |
| CN114554505B (en) * | 2022-01-10 | 2024-03-29 | 山东师范大学 | Wireless sensor network node clustering method and system |
| CN114676632A (en) * | 2022-03-29 | 2022-06-28 | 贵州电网有限责任公司 | Method and device for predicting energy consumption of chip special for electric power and computer equipment |
| CN117729518A (en) * | 2023-11-21 | 2024-03-19 | 山东科技大学 | Wireless sensor network clustering method based on improved swan optimization algorithm |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100471153C (en) * | 2007-02-05 | 2009-03-18 | 南京邮电大学 | High-efficiency controlling method of wireless sensor network topology |
| US7969928B2 (en) * | 2007-03-02 | 2011-06-28 | Motorola Solutions, Inc. | Method and apparatus for battery-aware dynamic bandwidth allocation for wireless sensor networks |
| CN103607749B (en) * | 2013-11-13 | 2017-02-08 | 福建工程学院 | A direction perception guided data collection method in an opportunity mobile sensor network |
| CN107016436A (en) * | 2017-03-31 | 2017-08-04 | 浙江大学 | A kind of mixing whale algorithm of bionical policy optimization |
-
2017
- 2017-12-15 CN CN201711351966.XA patent/CN108112049B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108112049A (en) | 2018-06-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108112049B (en) | A kind of wireless sensor network efficiency optimization cluster-dividing method based on whale group algorithm | |
| Wang et al. | A communication strategy of proactive nodes based on loop theorem in wireless sensor networks | |
| Kout et al. | AODVCS, a new bio-inspired routing protocol based on cuckoo search algorithm for mobile ad hoc networks | |
| Rezaeipanah et al. | A Hybrid Approach for Prolonging Lifetime of Wireless Sensor Networks Using Genetic Algorithm and Online Clustering. | |
| Jari et al. | PSO-based sink placement and load-balanced anycast routing in multi-sink WSNs considering compressive sensing theory | |
| Bhardwaj et al. | SecRIP: Secure and reliable intercluster routing protocol for efficient data transmission in flying ad hoc networks | |
| Thiagarajan et al. | RETRACTED ARTICLE: Quality of Service based Ad hoc On-demand Multipath Distance Vector Routing protocol in mobile ad hoc network | |
| CN107846719A (en) | A kind of wireless sensor network routing method based on improvement gam algorithm | |
| Nabavi et al. | Optimal selection of the cluster head in wireless sensor networks by combining the multiobjective genetic algorithm and the gravitational search algorithm | |
| Saminathan et al. | Energy efficient and delay aware clustering in mobile adhoc network: A hybrid fruit fly optimization algorithm and whale optimization algorithm approach | |
| Belkhira et al. | WRE‐OLSR, a new scheme for enhancing the lifetime within ad hoc and wireless sensor networks | |
| Yalçın et al. | A mobile sink path planning for wireless sensor networks based on priority‐ordered dependent nonparametric trees | |
| Yin et al. | Reinforcement Learning‐Based Routing Algorithm in Satellite‐Terrestrial Integrated Networks | |
| Rana et al. | Energy-efficient routing for wireless sensor network using genetic algorithm and particle swarm optimisation techniques | |
| CN107864092A (en) | A kind of cloud content distribution method, device based on multicasting technology | |
| Kumar et al. | An energy efficient clustering using k-means and AODV routing protocol in ad-hoc networks | |
| Nesary Moghadam et al. | Minimum cost load balanced multipath routing protocol for low power and lossy networks | |
| Wang et al. | Routing algorithm of energy efficient wireless sensor network based on partial energy level | |
| Long et al. | Research on applying hierachical clustered based routing technique using artificial intelligence algorithms for quality of service of service based routing | |
| Santos et al. | CGR: Centrality-based green routing for Low-power and Lossy Networks | |
| Zheng et al. | A reliable communication and load balancing scheme for resource-limited networks | |
| Vasanthi | Fractional Gaussian firefly algorithm and darwinian chicken swarm optimization for IoT multipath fault-tolerant routing | |
| Chen et al. | An efficient neural network-based next-hop selection strategy for multi-hop VANETs | |
| Dana et al. | Backup path set selection in ad hoc wireless network using link expiration time | |
| Satyananda et al. | A multi-objective QoS optimization with fuzzy based parameter setting for real-time multicasting |
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 |