CN108112049A - A kind of wireless sensor network efficiency optimization cluster-dividing method based on gam algorithm - Google Patents

A kind of wireless sensor network efficiency optimization cluster-dividing method based on gam algorithm Download PDF

Info

Publication number
CN108112049A
CN108112049A CN201711351966.XA CN201711351966A CN108112049A CN 108112049 A CN108112049 A CN 108112049A CN 201711351966 A CN201711351966 A CN 201711351966A CN 108112049 A CN108112049 A CN 108112049A
Authority
CN
China
Prior art keywords
whale
node
current
individual
gam
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201711351966.XA
Other languages
Chinese (zh)
Other versions
CN108112049B (en
Inventor
高亮
曾冰
李新宇
张振东
程璐瑶
董燕
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.)
Huazhong University of Science and Technology
Original Assignee
Huazhong University of Science and Technology
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 Huazhong University of Science and Technology filed Critical Huazhong University of Science and Technology
Priority to CN201711351966.XA priority Critical patent/CN108112049B/en
Publication of CN108112049A publication Critical patent/CN108112049A/en
Application granted granted Critical
Publication of CN108112049B publication Critical patent/CN108112049B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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 invention discloses a kind of wireless sensor network efficiencies based on gam algorithm to optimize 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 gam algorithm in efficiency optimizes sub-clustering problem, using solution ability of the gam algorithm in multi-modal optimization problem is improved, solves wireless sensor network efficiency optimization sub-clustering problem；Meanwhile routing algorithm can energy expenditure of the active balance leader cluster node when forwarding data, so as to be conducive to further extend network lifecycle.

Description

A kind of wireless sensor network efficiency optimization cluster-dividing method based on gam algorithm
Technical field
The invention belongs to wireless sensor network topology control fields, and in particular to a kind of based on the wireless of gam algorithm Sensor network efficiency optimizes cluster-dividing method.
Background technology
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, the U.S.《Business Weekly》WSNs is classified as one of 21 technologies that 21 century most influences, 2003, Massachusetts science and engineering Institute《Technology review》WSNs is classified as first of the following ten big new technologies for changing people's life.
In most wireless senser application scenarios, the use bottleneck of wireless senser is embodied in its energy to be had very much Limit, this is because most wireless sensor networks, using battery powered, this greatly limits wireless sensor networks Working time.Accordingly, it is capable to amount 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, can so reduce data redundancy, and reduce data traffic, so as to Extend Network morals；Good efficiency routing algorithm can ensure that data are forwarded along the path of optimization, balanced entire net The energy consumption of network, so as 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, so as to reduce the life cycle of whole network.At present, Many researchers are made using a kind of special node (gateway) For cluster head, as shown in the leader cluster node in Fig. 1, these nodes have energy more more than general sensor nodes, common to sense Device node can add in any one cluster head in its communication range.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 (into Member's node) excessively, then the energy of leader cluster node easily exhausts；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 given birth to for whole network The life cycle seems particularly critical.
Gam algorithm (Whale Swarm Algorithm, WSA) (refers to 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 is found that food source, it can make a sound other near notice Whale on quality of food quality and quantity number information.Therefore, every whale, which will receive, largely comes from neighbouring whale Notification information, appropriate local search of food is then moved to according to these information.WSA algorithm frames are simple, easy to implement, It is very suitable for due to solving engineering optimization.But gam algorithm (WSA) be for solving continuous optimization problems, and Wireless sensor network clustering problem is a discrete optimization problems of device, and gam 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：The node of information is only sent into the acquisition of row information and to cluster head, 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 be merged into row information and send it to aggregation node；
Distance：Euclidean distance between sensor node；
Hop count：Current sensor node transmits packets to the minimum sensor node number that aggregation node needs 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 Energy and leader cluster node are to the hop count of 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.
The content of the invention
For the disadvantages described above or Improvement requirement of the prior art, the present invention provides a kind of based on the wireless of gam 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 situation of Wireless sensor network clustering problem carries out wireless sensor network sub-clustering calculating based on improved gam 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 so as to the energy consumption of balanced wireless sensor network, effectively improves network energy efficiency, extends net Network life cycle.
To achieve the above object, one side according to the invention provides a kind of wireless biography based on gam algorithm 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, so as to obtain all cluster heads Node is 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 gathers and/or processing information, generation include the number of sensor node self rest energy information Corresponding leader cluster node is forwarded a packet to according to bag, 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 is preferably calculated by improved whale algorithm and obtained, including：
S1 sets gam algorithm parameter, and each whale individual in gam algorithm is initialized, and obtains just Beginning whale population calculates the fitness value of each whale individual；The individual that current whale is gam is set, into step Rapid S2；
S2 determines current whale with the presence or absence of guiding individual according to the distance between fitness value and whale individual；It is described Guiding individual is the whale nearest from current whale individual in all whale individuals better than current whale；If in the presence of into step Rapid S3；Otherwise S5 is entered step；
S3 generates the copy whale of current whale, and the guiding 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 the iteration count of current whale is arranged to 0, enters Step S7；Otherwise S4 is entered step；
The iteration count value of the more current whales of S4 and the size of gam stability threshold；If the iteration of current whale Counter Value is less than gam stability threshold, then the iteration count value is added 1, enter step S7；Otherwise reinitialize The position of current whale, calculates the fitness value of the whale, enters step S7；
S5 generates the copy of current whale, after performing neighborhood search to the copy, calculates the fitness value and ratio of the copy 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 arranged to 0, enters step S7；Otherwise S6 is entered step；
If the iteration count of the current whales of S6 is less than gam 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 calculates the fitness value of the whale, enters step S7；If no Less than then reinitializing the position of current whale, and the fitness value of the whale is calculated, enter step S7；
If the current whales of S7 are not the last one individual of gam, it is current whale in gam to set current whale Next whale individual, enter step S2；If current whale is the last one individual of gam, judge whether it meets end Only condition enters step S8 if meeting；Otherwise the individual that current whale is set to be gam, enters step S2；
S8 judges whether there is whale individual more better than globally optimal solution in last generation population, has and then substitutes the current overall situation Optimal solution obtains final globally optimal solution, is optimal sub-clustering scheme.
One as technical solution of the present invention is preferably, it is preferred to use random initializtion mode.
Preferably, the calculation formula of fitness is one as technical solution of the present invention：
F=ω σL+(1-ω)·σD·μD
In formula, ω is weight coefficient, σLFor the standard deviation of each leader cluster node lifetime, μD、σDOrdinary sensors are represented respectively Node is to the average and variance of 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, giRepresent i-th of cluster head, m represents the number of cluster head Amount.
Define 2.R：General sensor nodes set, R={ r1,r2,…,ri,…,rn},riRepresent i-th of ordinary sensors Node, n represent the quantity of general sensor nodes.
Define 3.RGri：In general sensor nodes riCluster head number set in communication range.
The specific design process of individual UVR exposure of whale individual is：
Since the quantity of general sensor nodes is fixed, that most effective fish individual UVR exposure scheme is every Whale individual represents a kind of complete general sensor nodes to the allocative decision of cluster head.It is as follows：
X=(x1,x2,…,xi,…,xn)
Wherein, X represents a whale individual, i.e., one solution；xiRepresent 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].
σLRepresent the standard deviation of each leader cluster node lifetime：M is leader cluster node number Amount, L (gi) represent leader cluster node giLifetime：μLFor the average of all leader cluster node lifetimes,Er(gi) it is leader cluster node giDump energy, Ec(gi) it is leader cluster node giIn single-wheel data sending mistake The energy consumed in journey,niRepresent leader cluster node gi's Member node quantity, k is receives data volume size (receiving data bit number), after l represents cluster head progress data fusion, transmission The size of data packet, EDARepresent leader cluster node giMerge the energy that the data that each general sensor nodes are sent are consumed, ETx Represent that leader cluster node sends the energy that data (data package size is by l) are consumed to its initial next-hop node,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 represents to receive in sensor node and transmitting circuit works When the energy that is consumed, εfsAnd εmpRespectively in free space loss model and multi-path fading channel, in sensor node Transmit the energy consumed during amplifier operation.
σDRepresent all general sensor nodes to the average and variance of its leader cluster node distance：
Wherein, n be general sensor nodes quantity, D (ri) it is general sensor nodes riTo affiliated cluster head away from From μDFor each ordinary node to the average value of the distance of respective cluster head,
The object function considers the energy consumption balance and general sensor nodes of cluster head and general sensor nodes simultaneously Wastage in bulk or weight energy.
Preferably, the guiding individual Y of current whale is guided one as technical solution of the present invention according to individual movement rule The copy whale X' of current whale, which carries out mobile process, to be included,
S31 sets traversal sequence 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 it then enters 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；
A cluster head g ' is randomly choosed in cluster head set of the S34 in i-th of general sensor nodes communication range, it 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.
One as technical solution of the present invention is preferred, and neighborhood search includes,
S41 chooses the best whale individual Ω of fitness value from whale populationb；
S42 calculates whale individual ΩbThe cluster head g of life cycle minimummin；
S43 is distributing to cluster head gminSensor node set in random selection one 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 describing 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 cluster head g of middle life cycle minimummin；
(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 preferably, it is preferred to use discrete individual coding mode carries out sensor node Coding.
The whale individual UVR exposure design of gam, using discrete individual UVR exposure mode, i.e. a whale individual represents one Paths, X=(s, x2, x3 ..., xi..., d) in formula, X represents a whale individual, i.e., one solves namely a paths；S tables Show source node, x2Represent second node in path, xiRepresent i-th of node in path, d represents root node.
One as technical solution of the present invention preferably, preferably using Hamming distance counted by the distance between whale individual It calculates.Hamming distance is the number of two individual correspondence position different values.I.e. to all elements correspondence position in two individuals into Row XOR operation, statistical result are 1 number, which is the value of Hamming distance.
In general, by the above technical scheme conceived by the present invention compared with prior art, have below beneficial to effect Fruit：
1) method of technical solution of the present invention introduces improved gam algorithm in efficiency optimizes sub-clustering problem, utilizes Solution ability of the gam algorithm in multi-modal optimization problem is improved, solves wireless sensor network efficiency optimization sub-clustering problem；Together When, in the cluster-level routing stage, the routing algorithm of introducing can energy expenditure 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 algorithms For formula discretization, make the WSA-IC algorithms of solution continuous optimization problems, 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 balanced leader cluster node of the solution acquired and general sensor nodes, while also reduces general sensor nodes Energy consumption, so as to extend the life cycle of entire wireless senser.
4) method of technical solution of the present invention for the coding mode of whale individual, adds Hamming distance algorithm 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 strengthen Algorithm refuses search capability, accelerates to converge to the optimal solution of the individual region, improves gam convergence speed of the algorithm And the operational efficiency of algorithm.
Description of the drawings
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 netinit and the flow diagram of 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) shows for the best individual of population after progress neighborhood search 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of 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 (wherein aggregation node is located at the position of coordinate (100,100))；
Figure 10 (a) be the embodiment of the present invention in experiment scene WSN#1 embodiment 3 in, WSA-ICC algorithms of the invention with The fitness convergent contrast and experiment of DECA, IHSC algorithm；
Figure 10 (b) is WSA-ICC algorithms 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 algorithms of the invention with The fitness convergent contrast and experiment of DECA, IHSC algorithm；
Figure 10 (d) is WSA-ICC algorithms 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 Method and the network lifecycle contrast and experiment of 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 Method and the network lifecycle contrast and experiment of 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 Method and the RLGD-RFGD contrast and experiments of 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 Method and the RLGD-RFGD contrast and experiments of 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 Method and the RLGD-RFGD contrast and experiments of 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 experiments 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 algorithms of the invention with The experiment of energy consumption result of EELBCA, DECA, IHSC algorithm；
Figure 13 (b) is WSA-ICC algorithms 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 algorithms of the invention with The experiment of energy consumption result of EELBCA, DECA, IHSC algorithm；
Figure 13 (d) is WSA-ICC algorithms 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 algorithms of the invention with The experiment of energy consumption result of EELBCA, DECA, IHSC algorithm；
Figure 14 (b) is WSA-ICC algorithms 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 algorithms of the invention with The experiment of energy consumption result of EELBCA, DECA, IHSC algorithm；
Figure 14 (d) is WSA-ICC algorithms 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 purpose , technical scheme and advantage of the present invention be clearer, with reference to the accompanying drawings and embodiments, it is 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 Conflict is not formed each other to 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 gam 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 converges step with the network information, as shown in figure 3, specifically including：
A. unique mark determines：Include all biographies of aggregation node, leader cluster node and general sensor nodes to network Sensor node is numbered, as their unique mark.
B. calculating network initial configuration information：According to known nodal distance information, shortest path Dijkstra (enlightening is used Jie Sitela) algorithm obtains all nodes to the shortest path of aggregation node, using the program as initialization default route； According to known nodal distance information, the nearest leader cluster node of all ordinary nodes is found, which is the ordinary node Initialization acquiescence cluster head, using the program as initialization acquiescence sub-clustering.
C. netinit configures：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 somebody with somebody It puts.
D. the network information converges：Ordinary node receives the initial configuration information, and it is surplus to send it to its acquiescence 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) according to the dump energy information of all nodes, net is carried out to entire sensor network for clustering routing calculation procedure The optimization of network clustering routing calculates and configuration, specifically includes：
A. energy-optimised router-level topology：
Base station is used calculates each leader cluster node based on the wireless sensor network routing method for improving gam algorithm Optimal forward-path, wherein the leader cluster node of the present invention is considered the sensor node of the above method, carries out router-level topology, 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 optimization sub-clustering problem, obtains optimal solution using gam 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 sets gam algorithm parameter, including stability threshold Ts=200+n/2, wherein n are 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 gam algorithm, obtain initial whale population：
Ω={ Ω1, Ω2..., Ωi..., Ω|Ω|}；
Wherein, | Ω | for 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, giRepresent i-th of cluster head, m represents the number of cluster head Amount.
Define 2.R：General sensor nodes set, R={ r1,r2,…,ri,…,rn},riRepresent i-th of ordinary sensors Node, n represent the quantity of general sensor nodes.
Define 3.RGri：In general sensor nodes riCluster head number set in communication range.
Since the quantity of general sensor nodes is fixed, most effective whale individual UVR exposure scheme is every whale Fish individual represents a kind of complete general sensor nodes to the allocative decision of cluster head.It is as follows：
X=(x1,x2..., xi..., xn)
Wherein, X represents a whale individual, i.e., one solution；xiRepresent general sensor nodes riAffiliated cluster head number, That is general sensor nodes riDistribute to cluster head xi,
As shown in figure 5,5 cluster heads and 12 general sensor nodes, i.e. G={ g are included in network1,g2,g3,g4,g5, R ={ r1,r2,r3,r4,r5,r6,r7,r8,r9,r10,r11,r12}.In figure, the connecting line meaning between general sensor nodes and cluster head Taste two sensor nodes and 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 gam individual；
Sub-step 3, by fitness function, calculate the fitness value of each whale individual：
F={ f (Ω1), f (Ω2) ..., f (Ωi) ..., f (Ω|Ω|)}
Wherein, f (Ωi) for the fitness of i-th whale individual；
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].
σLRepresent the standard deviation of each leader cluster node lifetime：M is leader cluster node number Amount, L (gi) represent leader cluster node giLifetime：μLFor the average of all leader cluster node lifetimes,Er(gi) it is leader cluster node giDump energy, Ec(gi) it is leader cluster node giIn single-wheel data sending mistake The energy consumed in journey,niRepresent leader cluster node gi's Member node quantity, k is receives data volume size (receiving data bit number), after l represents cluster head progress data fusion, transmission The size of data packet, EDARepresent leader cluster node giMerge the energy that the data that each general sensor nodes are sent are consumed, ETx Represent that leader cluster node sends the energy that data (data package size is by l) are consumed to its initial next-hop node,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 representing to be received in sensor node with transmitting circuit work The energy consumed, εfsAnd εmpRespectively in free space loss model and multi-path fading channel, passed in sensor node Send the energy consumed during amplifier operation.
σDRepresent all general sensor nodes to the average and variance of its leader cluster node distance：
Wherein, n be general sensor nodes quantity, D (ri) it is general sensor nodes riTo affiliated cluster head away from From μDFor each ordinary node to the average value of the distance of respective cluster head,
The object 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 sets traversal sequence number i=1；
Sub-step 5, according to the distance between fitness value and whale individual, find from whale Ωi" more excellent and nearest " Whale Y, if Y exists, rotor step 6, otherwise rotor step 12；
Distance wherein between whale individual, is calculated using Hamming distance, is specially：
Hamming distance is the number of two individual correspondence position different values.Position is corresponded to all elements in two individuals Carry out XOR operation is put, statistical result is 1 number, which is the value of Hamming distance.
Assuming that two whale individual X1、X2It is as follows, X1And X2There are 5 elements to differ, so, X1And X2Between the Chinese Prescribed distance is 5.
Sub-step 6, generation whale ΩiCopy X'；
Sub-step 7, according to individual movement rule, by whale Y whale X' is guided to be moved；
Wherein individual movement rule, idiographic flow is as shown in fig. 6, include the following steps：
(7.1) traversal sequence number i=1 is set；
(7.2) if X'iNot equal to Yi, then (7.3) are gone to step, are otherwise gone to step (7.5)；
(7.3) the random number P of one 0 to 1 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；
It is as follows, whale ΩuCopy X' and its " more excellent and nearest " whale Y, X' and Y first element not phase Together, therefore, the random number P=0.2 between one 0 to 1 is generated, because P<λ, so first element of Y is assigned to X''s First element, at this time X'=(2,3,4,2,3,3,1,3,1,2,5,4)；X' is identical with second element of Y, need not be to X' Carry out any operation；The 3rd element of X' and Y differs, and 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 3rd 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' just terminates Movement under Y guiding.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 becomes 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 gamiWhale copy X' is replaced by, by whale iteration count Ωi.c 0 is arranged to, rotor step 18；
If sub-step 10, the iteration count Ω of 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；
I-th sub-step 12, generation of whale ΩiCopy X "；
Sub-step 13 performs 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 idiographic flow 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 cluster head g of middle life cycle minimummin；
(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, The cluster head of lifetime minimum is g5, L (g5)=720.Therefore, From setInterior one general sensor nodes number of 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 gamiIt is replaced by whale copy X ", and by the whale iteration count Ωi.c0 is arranged to, rotor step 18；
If sub-step 15, the iteration count Ω of i-th whalei.cNot equal to Ts, then to iteration count Ωi.cAdd 1, Rotor step 18, otherwise rotor step 16；
If sub-step 16, the fitness value f (Ω of i-th whalei) it is less than the fitness value f of globally optimal solutiongbest, then will fgbestIt is arranged to f (Ωi), globally optimal solution GBest is arranged to Ωi；
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 be less than whale individual amount | Ω |, rotor step 5, otherwise rotor walk Rapid 19；
Sub-step 19 judges whether end 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 be current optimal solution, i.e., optimal routing plan.
(3) clustering routing configuration step, it is specific as follows：
Aggregation node sends optimal routing configuration information and most using above-mentioned optimal routing plan to all leader cluster nodes 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 to corresponding general sensor nodes, after ordinary node stores corresponding sub-clustering configuration information, added Enter into sub-clustering.
(4) information gathering and converging information step, all wireless sensor nodes carry out information gathering and processing application letter Breath according to the clustering routing of configuration, completes the information transmission to aggregation node, specific as follows：
General sensor nodes carry out information gathering 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 is into row information Acquisition and processing application message, are processed into data packet, and its current remaining information are added in data packet；Leader cluster node The data fusion of sub-clustering where completing by according to the routing configured, by all data sendings to aggregation node, is completed information and is adopted Collection and convergence.The step constantly performs periodically, realizes the acquisition information update of wireless sensor network.
This programme, it is preferable that in described information acquisition and converging information step, after convergence wheel number is more than 20 times, again Efficiency optimization router-level topology is carried out, after forming new routing plan, step (3) (4) is re-executed, completes the dynamic of routing plan Update.
This programme, it is preferable that in described information acquisition and converging information step, when some section occurs in wireless sensor network After point depleted of energy, re-start efficiency optimization sub-clustering and calculate, after forming new sub-clustering scheme, re-execute step (3) (4), Complete the dynamic update of sub-clustering scheme.
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 carry out Integrated comparative to the effect of emulation experiment, the present embodiment devises two kinds of scenes, the first scene is WS, aggregation node coordinate are the leftmost side of (200,100), i.e. simulating area；Second scene is WSN#2, aggregation node Coordinate is the center of (100,100), i.e. simulating area.That is, by 16 specific embodiments in following table 1 to this reality The optimization cluster-dividing method of the wireless sensor network efficiency in example is applied further to be 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 in the present invention and the Load balance clustering method of Energy Efficient, i.e. EELBCA (Energy Efficient 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 When cycle is first cluster head depleted of energy occurs, 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), surviving node number (often take turns data sending it is complete, present energy There are no the general sensor nodes quantity exhausted) and energy consumption (terminate until working as front-wheel number, the energy of all nodes consumption is gone through History summation) it is compared.In the present embodiment, Simulation Program language is C++, and allocation of computer is：Dominant frequency for 3.2GHz and The intel I5-3470QM processors of 3.6GHz, 4GB memories, 10 64 bit manipulation systems 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 object function.
2 emulation experiment network parameter of table is set
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.
Primary power (the unit of 3 each example interior joint of table：J)
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 storehouse size；HMCRmin、HMCRmaxThe respectively minimum value and maximum of harmony storehouse select probability；| Ω | it is Population Size；N is the general sensor nodes quantity currently survived.
4 algorithm parameter of table is set
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 in all examples, the present invention carries The WSA-ICC algorithms gone out are either all than it in terms of fitness convergence rate or 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 algorithms：Every whale individual its " it is more excellent and It is moved under the guiding 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 algorithms can rapidly converge to the extreme point near each individual.Meanwhile The neighborhood search strategy that the present invention adds in can improve the precision of solution.Importantly, WSA-IC algorithms can be in iterative process It is middle to identify and jump out the extreme point found, so as to improve the ability of searching optimum of algorithm as much as possible, drastically increase 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 end 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 Go out, in most examples, WSA-ICC algorithms proposed by the present invention can obtain longer life cycle than other algorithms； Especially for 2#WSN, the advantage of WSA-ICC algorithms is more obvious.EELBCA obtains minimum life in many examples Cycle is because in EELBCA, and cluster head is to distribute member node rather than according to cluster head according to rank and file's number of nodes 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 quickly so as to cause the less cluster head of dump energy.EELBCA Much the same life cycle can be obtained with other algorithms in some examples, be because these cluster algorithms all employ WSA- ICR routing algorithms calculate the routing of cluster head, so as to significantly balance the energy consumption of cluster head in the transmission phase of data packet.WSA- Why ICC algorithms do very well in terms of life cycle than other comparison algorithms, are to have benefited from the outstanding solution of WSA-ICC algorithms Performance, it can find better solution than other algorithms, distribute rank and file's node according to the dump energy of cluster head so that remaining The minimum cluster head of energy possesses rank and file's node of negligible amounts, extends their lifetime as much as possible, so as to extend net The life cycle of network.
RLGD-RFGD compares：WSA-ICC, EELBCA, DECA and IHSC are carried out in fact in above-mentioned wireless sensor network It applies, experiment end 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 results of embodiment 9~12 and embodiment 13~16." RLGD-RFGD " index is to use Evaluate performance of each cluster-dividing method in terms of cluster head energy consumption is balanced, it is smaller that this refers to target value, 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 cluster head energy consumption is balanced Performance it is better.From experimental result as can be seen that DECA obtains worst RLGD-RFGD in most examples, especially It is for the more 5~example of example 8 of cluster head (search space bigger).This is because the solution ability of DECA is poor, from DECA Fitness convergence performance as can be seen that the last solutions that acquire of DECA are 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, the cluster head nearer from aggregation node also need for compared with Remote cluster head forwarding 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 algorithms 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 expenditure 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 for more four of cluster head in Figure 12 (b) Example, the performance of WSA-ICC algorithms are more prominent.It is outstanding that this shows that WSA-ICC algorithms have 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.Embodiment used is wherein tested as the embodiment 3 of scene WSN#1 and example 8, 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 general sensor nodes energy consumption is reduced Performance, it is bigger that this refers to target value, 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 algorithms and EELBCA algorithms 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 algorithms Current cluster head in this way, most of general sensor nodes can be assigned to nearest or nearer cluster head, can be reduced greatly common The energy consumption of sensor node, so as to extend their lifetime；But this method may make the laod unbalance of cluster head, The cluster head energy expenditure for causing overload is too fast, so as to shorten Network morals.Therefore, EELBCA algorithms are in major part Worst life cycle is all obtained in example.IHSC algorithms all carry out neighborhood after candidate's harmony is generated every time to best harmony Many general sensor nodes can be assigned to nearest or secondary near cluster head by search, the optimal solution finally acquired, can also pole The big energy consumption for reducing general sensor nodes, so as to significantly improve the lifetime of general sensor nodes；But it is received from fitness The situation of holding back can be seen that first index (standard deviation of cluster head lifetime) of optimal solution that IHSC algorithms acquire in object function On cannot obtain optimal or near excellent value, so, performance of the IHSC algorithms in terms of network lifecycle is than WSA-ICC algorithm Difference.The main object of the present invention is reasonably to distribute general sensor nodes quantity by cluster algorithm for cluster head, so as to the greatest extent may be used The energy consumption of the balanced cluster head in energy ground, extends Network morals.Because in actual WSN scenes, there is the common biography of many redundancies Sensor node, the depleted of energy of some general sensor nodes will not cause too much influence to data acquisition；And the energy of cluster head Amount exhausts the data that will lose the acquisition of its rank and file's node, and its rank and file's node is caused to add in farther cluster head, So as to accelerate the energy expenditure of these general sensor nodes.Therefore, there is data acquisition in the lifetime of cluster head most important Effect, it is necessary to the energy consumption of cluster head balanced as much as possible.So in the object function proposed in this chapter, the mark of cluster head lifetime The weight coefficient of quasi- difference is arranged to 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 algorithms are with best Solution ability.The optimal solution that WSA-ICC algorithms acquire can be to balance based on the energy consumption of cluster head, may be by some ordinary sensors Node is added to the more cluster heads of distant dump energy, causes the energy consumption of these general sensor nodes to increase, so as to, Influence performance of the WSA-ICC algorithms in terms of surviving node number.
Energy consumption compares：IHSCRA of the present invention carries out real with EELBCA, DECA and IHSC in above-mentioned wireless sensor network It applies.Embodiment used is wherein tested as the embodiment 3 of scene WSN#1 and example 8, the embodiment 11 and example of scene WSN#2 16.Figure 14 (a)~Figure 14 (d) is embodiment 3, the experimental data comparing result of 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 methods reduce whole network energy consumption in terms of performance it is poorer than EELBCA and IHSC methods, this also because The optimal solution that WSA-ICC methods acquire is mainly to balance the energy consumption of cluster head, may be added to some general sensor nodes The more cluster heads of distant dump energy, so as to add the energy consumption of general sensor nodes.But it will again be seen that Performance of the WSA-ICC methods in terms of energy consumption is better than DECA method.
In conclusion WSA-ICC methods of the present invention are 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, all any modification, equivalent and improvement made within the spirit and principles of the invention etc., should all include Within protection scope of the present invention.

Claims (7)

1. a kind of wireless sensor network efficiency optimization cluster-dividing method based on gam algorithm, 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, so as 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 gathers and/or processing information, generation include 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, according to clustering routing Configuration transmits data to aggregation node, completes converging information acquisition；
Wherein, the optimal sub-clustering scheme is preferably calculated by improved whale algorithm and obtained, including：
S1 sets gam algorithm parameter, and each whale individual in gam algorithm is initialized, and obtains initial whale Fingerling group calculates the fitness value of each whale individual；The individual that current whale is set to be gam, enters step S2；
S2 determines current whale with the presence or absence of guiding individual according to the distance between fitness value and whale individual；The guiding Individual is the whale nearest from current whale individual in all whale individuals better than current whale；If in the presence of entering step S3；Otherwise S5 is entered step；
S3 generates the copy whale of current whale, and the guiding 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 the iteration count of current whale is arranged to 0, enters step S7；Otherwise S4 is entered step；
The iteration count value of the more current whales of S4 and the size of gam stability threshold；If the iteration count of current whale Device value is less than gam stability threshold, then the iteration count value is added 1, enter step S7；Otherwise reinitialize current The position of whale calculates the fitness value of the whale, enters step S7；
S5 generates the copy of current whale, after performing 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 arranged to 0, enters step S7；Otherwise S6 is entered step；
If the iteration count of the current whales of S6 is less than gam 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 it is 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 whales of S7 are not the last one individual of gam, set current whale for current whale in gam under One whale individual, enters step S2；If current whale is the last one individual of gam, judge whether it meets termination item Part enters step S8 if meeting；Otherwise the individual that current whale is set to be gam, enters step S2；
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, is optimal sub-clustering scheme.
2. the wireless sensor network efficiency optimization cluster-dividing method according to claim 1 based on gam algorithm, wherein, The initialization preferably uses random initializtion mode.
3. the wireless sensor network efficiency optimization cluster-dividing method according to claim 1 or 2 based on gam algorithm, In, the calculation formula of fitness is：
F=ω σL+(1-ω)·σD·μD
In formula, ω is weight coefficient, σLFor the standard deviation of each leader cluster node lifetime, μD、σDGeneral sensor nodes are represented respectively To the average and variance of its leader cluster node distance.
4. sub-clustering is optimized according to wireless sensor network efficiency of the claims 1 to 3 any one of them based on gam algorithm Method, wherein, the guiding individual Y of current whale guides the copy whale X' of current whale to be moved according to individual movement rule Process include,
S31 sets traversal sequence 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；
A cluster head g ' is randomly choosed in cluster head set of the S34 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. sub-clustering is optimized according to wireless sensor network efficiency of the Claims 1 to 4 any one of them based on gam algorithm Method, wherein, the neighborhood search includes,
S41 chooses the best whale individual Ω of fitness value from whale populationb；
S42 calculates whale individual ΩbThe cluster head g of life cycle minimummin；
S43 is distributing to cluster head gminSensor node set in random selection one 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. sub-clustering is optimized according to wireless sensor network efficiency of the Claims 1 to 5 any one of them based on gam algorithm Method, wherein, it is preferred to use discrete individual coding mode encodes sensor node.
7. sub-clustering is optimized according to wireless sensor network efficiency of claim 1~6 any one of them based on gam algorithm Method, wherein, the distance between whale individual is preferably calculated using Hamming distance.
CN201711351966.XA 2017-12-15 2017-12-15 A kind of wireless sensor network efficiency optimization cluster-dividing method based on whale group algorithm Active CN108112049B (en)

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 true CN108112049A (en) 2018-06-01
CN108112049B 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)

Cited By (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108934054A (en) * 2018-06-19 2018-12-04 昆明理工大学 A kind of Ad-hoc network cluster dividing algorithm based on node data value
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
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
CN109931903A (en) * 2019-02-26 2019-06-25 上海大学 A kind of cylindricity assessment method based on improvement whale optimization algorithm
CN110363214A (en) * 2019-05-31 2019-10-22 浙江工业大学 A kind of contact condition recognition methods of the robotic asssembly based on GWA-SVM
CN111526555A (en) * 2020-05-25 2020-08-11 南京邮电大学 Multi-hop routing path selection method based on genetic algorithm
CN111542011A (en) * 2020-04-27 2020-08-14 中山大学 Layered wireless sensor network clustering routing method based on particle swarm optimization
CN112492661A (en) * 2020-12-10 2021-03-12 中南民族大学 Wireless sensor network clustering routing method based on improved sparrow search algorithm
CN112672396A (en) * 2020-12-10 2021-04-16 中南民族大学 Wireless sensor network mobile node clustering method based on improved whale algorithm
CN113179544A (en) * 2021-04-26 2021-07-27 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
CN114339785A (en) * 2021-12-15 2022-04-12 红云红河烟草(集团)有限责任公司 Raw tobacco maintenance wireless sensor layout optimization algorithm based on group intelligence
CN114554505A (en) * 2022-01-10 2022-05-27 山东师范大学 Wireless sensor network node clustering method and system

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101013987A (en) * 2007-02-05 2007-08-08 南京邮电大学 High-efficiency controlling method of wireless sensor network topology
US20080212557A1 (en) * 2007-03-02 2008-09-04 Motorola, Inc. Method and apparatus for battery-aware dynamic bandwidth allocation for wireless sensor networks
CN103607749A (en) * 2013-11-13 2014-02-26 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

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101013987A (en) * 2007-02-05 2007-08-08 南京邮电大学 High-efficiency controlling method of wireless sensor network topology
US20080212557A1 (en) * 2007-03-02 2008-09-04 Motorola, Inc. Method and apparatus for battery-aware dynamic bandwidth allocation for wireless sensor networks
CN103607749A (en) * 2013-11-13 2014-02-26 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

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
BING ZENG等: "Whale Swarm Algorithm for Function Optimization", 《ICIC 2017》 *
孔召文等: "一种无线传感器网络分簇路由协议优化算法", 《仪表技术》 *

Cited By (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108934054A (en) * 2018-06-19 2018-12-04 昆明理工大学 A kind of Ad-hoc network cluster dividing algorithm based on node data value
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
CN109931903A (en) * 2019-02-26 2019-06-25 上海大学 A kind of cylindricity assessment method based on improvement whale optimization algorithm
CN110363214A (en) * 2019-05-31 2019-10-22 浙江工业大学 A kind of contact condition recognition methods of the robotic asssembly based on GWA-SVM
CN111542011B (en) * 2020-04-27 2021-09-24 中山大学 Layered wireless sensor network clustering routing method based on particle swarm optimization
CN111542011A (en) * 2020-04-27 2020-08-14 中山大学 Layered wireless sensor network clustering routing method based on particle swarm optimization
CN111526555A (en) * 2020-05-25 2020-08-11 南京邮电大学 Multi-hop routing path selection method based on genetic algorithm
CN111526555B (en) * 2020-05-25 2022-07-26 南京邮电大学 Multi-hop routing path selection method based on genetic algorithm
CN112672396A (en) * 2020-12-10 2021-04-16 中南民族大学 Wireless sensor network mobile node clustering method based on improved whale 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
CN112492661A (en) * 2020-12-10 2021-03-12 中南民族大学 Wireless sensor network clustering routing method based on improved sparrow search algorithm
CN113179544A (en) * 2021-04-26 2021-07-27 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
CN114339785A (en) * 2021-12-15 2022-04-12 红云红河烟草(集团)有限责任公司 Raw tobacco maintenance wireless sensor layout optimization algorithm based on group intelligence
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
CN114554505A (en) * 2022-01-10 2022-05-27 山东师范大学 Wireless sensor network node clustering method and system
CN114554505B (en) * 2022-01-10 2024-03-29 山东师范大学 Wireless sensor network node clustering method and system

Also Published As

Publication number Publication date
CN108112049B (en) 2018-12-07

Similar Documents

Publication Publication Date Title
CN108112049B (en) A kind of wireless sensor network efficiency optimization cluster-dividing method based on whale group algorithm
Qiu et al. TOSG: A topology optimization scheme with global small world for industrial heterogeneous Internet of Things
Junhai et al. A survey of multicast routing protocols for mobile ad-hoc networks
Misra et al. An ant swarm-inspired energy-aware routing protocol for wireless ad-hoc networks
Rezaeipanah et al. A Hybrid Approach for Prolonging Lifetime of Wireless Sensor Networks Using Genetic Algorithm and Online Clustering.
Verma et al. Integrated routing protocol for opportunistic networks
CN107846719B (en) A kind of wireless sensor network routing method based on improvement gam algorithm
Bhardwaj et al. SecRIP: Secure and reliable intercluster routing protocol for efficient data transmission in flying ad hoc networks
Yao et al. A learning-based approach to intra-domain QoS routing
Wang et al. An improved routing algorithm based on social link awareness in delay tolerant networks
You et al. Virtual position based geographic routing for wireless sensor networks
Nesary Moghadam et al. Minimum cost load balanced multipath routing protocol for low power and lossy networks
Rana et al. Energy-efficient routing for wireless sensor network using genetic algorithm and particle swarm optimisation techniques
Long et al. Research on applying hierachical clustered based routing technique using artificial intelligence algorithms for quality of service of service based routing
Kumar et al. An energy efficient clustering using k-means and AODV routing protocol in ad-hoc networks
Maniscalco et al. Binary and m-ary encoding in applications of tree-based genetic algorithms for QoS routing
Ganesh et al. A multicast transmission routing protocol for low power lossy network based IoT ecosystem
Al-Nahari et al. Receiver-based ad hoc on demand multipath routing protocol for mobile ad hoc networks
El-Sayed Shortest paths routing problem in MANETs
Vimala et al. Strongest persistent multicast routing protocol for reliable transmission in both ad-hoc and mobile ad-hoc networks
Murthy et al. Network lifetime analytical model for node-disjoint multipath routing in wireless sensor networks
Vasanthi Fractional Gaussian firefly algorithm and darwinian chicken swarm optimization for IoT multipath fault-tolerant routing
Dong et al. Topology control mechanism based on link available probability in aeronautical ad hoc network
Boyd et al. A collaborative wireless sensor network routing scheme for reducing energy wastage
Singh et al. Optimization of Routing Protocol in MANET using GA

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