CN105101333A - Wireless sensor network cluster-based routing method adopting living space evolution (LSE) intelligent algorithm - Google Patents

Abstract

The invention discloses a wireless sensor network cluster-based routing method adopting a living space evolution (LSE) intelligent algorithm. The LSE algorithm mainly comprises: I, living space evolution for the offspring of life to gradually transfer to a rich or high-quality living space under the guidance of living condition information; and II, multi-offspring reproduction to match with the reproduction phenomenon of real life in nature. In the method, one life is introduced, and a plurality of offspring can be reproduced within one generation. According to the method disclosed by the invention, a cluster-based routing structure is optimized through the LSE intelligent algorithm, so that the data transmission efficiency of a wireless sensor network is increased, and the service life of the wireless sensor network is prolonged greatly.

Description

A kind of wireless sensor network clustering routing adopting living space evolution intelligent algorithm

Technical field

The present invention relates to a kind of wireless sensor network clustering routing, particularly relate to a kind of wireless sensor network clustering routing adopting living space evolution intelligent algorithm.

Background technology

Bio-inspired computing (Bio-inspiredcomputation) opens intelligent computation (Computationalintelligence) new way, and its basic thought takes out intelligent computation method according to the behavioural characteristic of biosystem or working method.Under the guidance of this basic thought, a series of intelligent computation method is arisen at the historic moment, and comprises artificial neural net, evolutionary computation, colony intelligence technology, artificial immune system etc.The neuromechanism of these method mimic biologies or various actions, countless application their validity and reliability verified.

At present, bio-inspired computing develops into today, has mainly imitated the related characteristics of nature biotechnology system from the following aspects.

(1) neuromechanism of mimic biology: artificial neural net (Artificialneuralnetwork)

Artificial neural net (ANN) is the Mathematical Modeling of a multi input, multi output, from the imitation to biological nervous system structure.ANN is made up of multiple neuron, and each neuron is with corresponding weights, and these weights can dynamic conditioning.These can the weights of dynamic conditioning just, constitute neuronic memory capability.Neuronic connection can have different structures, which reflects the dissimilar of neural net.Verified in theory, neural net can approach multi input, the multi output nonlinear function of any type.

(2) evolutionary process of mimic biology: evolutionary computation (Evolutionarycomputation)

Evolutionary computation (EC) originates from the emulation to biological evolution, and its basic skills has natural selection and gene genetic etc.Evolutionary computation method the most famous is genetic algorithm (GA), its basic skills is: a given initial population, calculate their fitness, the excellent individuality of fitness is extracted by selection opertor, exchange individual gene by crossover operator by certain probability and produce new individuality, by mutation operator, some individuals is made a variation.The feature of GA algorithm is the individuality keeping some, by the continuous search problem domain space of the method for iteration, thus reaches the object of optimizing.In addition, the mutation of other GA has Evolutionary Programming (EP), evolution strategy (ES), gene programming (GP) etc.

(3) group behavior of mimic biology: colony intelligence (Swarmintelligence)

Colony intelligence (SI) is mainly subject to the social life style of biocenose and the inspiration of behavior, such as, and flock of birds, the shoal of fish, ant group, bee colony etc.These colonies with centralized control or Ad hoc mode, with the mode of hand-in-glove search of food in certain spatial dimension.Swarm intelligence algorithm the most famous has particle cluster algorithm (PSO), and its operation principle is: have the population of fixed size as colony, and the behavior of flock of birds predation is imitated in group behavior.In the process of predation, the individuality of particle moves certain distance according to certain algorithm to current local and global optimum.Algorithm carries out in an iterative manner, thus approaches the globally optimal solution of search volume.Other similar swarm intelligence algorithm has ant group algorithm (ACO), artificial honeybee algorithm (ABC) etc.

(4) immune system of mimic biology: artificial immune system (ArtificialImmuneSystem)

The dynamic immune system model that artificial immune system (AIS) is the immune working mechanism of mimic biology and produces.It has imitated antibody in Immune System and antibody, course of reaction between antibody and antigen, and adopts the mode of complementary matched character string to achieve this process.From the angle of information processing, immune system possesses powerful identification, the ability of learning and memory and distributed, self-organizing and diversity characteristic, these significant characteristics constantly attract researcher from immune system, extract useful metaphoric mechanism, develop corresponding AIS model and algorithm for information processing and problem solving.

The primary biological system more than summarizing four aspects of bio-inspired computing inspires achievement.Therefrom can see, up to the present, bio-inspired computing is mainly limited to special sexual enlightenments such as biosystem behavior and structures, and they have a common defect, and that is exactly the life condition that have ignored biological its own existence space.Make a general survey of the biological phenomena of occurring in nature, life cycle and survival of the fittest rule are the eternal laws of living individual and colony two of must observe.Under the guidance of these two rules, the life generation procreation generation in the Nature, life and growth in nature, endless.But do not have certain life condition, life cannot exist.Meanwhile, do not go to find more superior life condition, life also just loses life.In addition, colony is also often confined to certain quantity, and the true biological reproductive patterns of this and occurring in nature also misfits.

In Wireless Sensor Network Routing Protocol, LEACH is a famous clustering route protocol, its objective is the energy imbalance problem for solving between node, and its basic means is sub-clustering again in each is taken turns, so that rotation race head between the individual nodes.LEACH Stochastic choice part node is as race's head node, and race's head node is collected neighbor information thus formed a race.Each node collects the information of institute's perception, sends race's head node to.Again by race's head node fused data, and send aggregation node to.LEACH considers the energy balance between data fusion and different node.But it does not consider the distance of different node to aggregation node.In fact, the difference of distance also can cause the imbalance of energy.In order to address this problem, CPSOCH agreement not only considers the energy balance of different node, and have also contemplated that the distance of different node to aggregation node.But CPSOCH agreement needs a process optimized.

Summary of the invention

The present invention proposes a kind of wireless sensor network clustering routing adopting living space evolution intelligent algorithm, is the application of a kind of new living space evolution (LivingSpaceEvolution is called for short LSE) intelligent algorithm.For the deficiency that have ignored biological its own existence condition and fixed group quantity etc. in bio-inspired computing method in the past, in this intelligent algorithm LSE, propose two new thoughts.One is that living space is evolved: under the guide of life condition information, and the offspring of life is gradually to living space transfer that is rich or high-quality.Two is many offsprings breedings: in order to match with the reproductive patterns of occurring in nature real life, introduce a life in the method and can breed multiple offspring in a generation.

Step of the present invention is:

1st step: initialization

{

1-1 walks: the nodes arranging sensor network;

1-2 walks: the energy parameter arranging following (1), (2), (3) formula:

Transmission k bit data to d rice distance needed for energy ezpenditure, E _tx(k, d), its size is:

E _Tx(k,d)＝kE _elec+kε _fsd ²(d＜d ₀)

E _Tx(k,d)＝kE _elec+kε _mpd ⁴(d≥d ₀)(1)

Receive the energy that these information consume, E _rx(k), its size is:

E _Rx(k)＝kE _elec(2)

Process the energy required for these information, E _da-fus(k), its size is:

E _da-fus(k)＝kE _da(3)

Wherein, E _electranstation mission circuit and the energy ezpenditure accepting circuit; E _dait is the energy ezpenditure of deal with data; ε _fsand ε _mpthe power loss of free space model and multichannel attenuation model respectively.D ₀it is transmission range threshold value.

1-3 walks: arrange fitness function parameter according to following (4) formula:

fitness＝α ₁f ₁+α ₂f ₂(4)

$f_1=\underset{k=1}{\overset{K}{\Σ}}{q}_k/\underset{i=1}{\overset{N}{\Σ}}{q}_i---\left(5\right)$

$f_2=\underset{i=1}{\overset{N}{\Σ}}{l}_i/\underset{k=1}{\overset{K}{\Σ}}{l}_k---\left(6\right)$

Wherein, α ₁and α ₂be weights coefficients, and meet α ₁+ α ₂=1; q _iit is the dump energy of i-th node in network; q _kthe dump energy of ShikGe race head node; f ₁it is the evaluation factor of ability; l _ithe distance of i-th node to aggregation node; l _kshikGe race head node is to the distance of aggregation node; f ₂it is the evaluation factor of distance.

1-4 walks: the maximum number arranging transmission data.

}

2nd step: the method provided according to LEACH, selects race's head, sets up group, creates dispatch list, then transmit data.

3rd step: judge whether the maximum reaching transmission data.If reach, then jump to the 7th step.

4th step: select to be optimized to race's head by LSE intelligent algorithm

{

4-1 walks: to each node serial number of sensitizing range;

4-2 walks: arrange race's head number;

4-3 walks: according to the calculated capacity of each node, arrange maximum iteration time;

4-4 walks: judge whether to reach maximum iteration time.If reached, then jump to the 5th step;

4-5 walks: the living space creating life, and its method is as follows:

The living space of life=candidate 1, candidate 2 ..., each candidate of candidate K| can get some sequence numbers 1 to N}.

4-6 walks: optimize discrete living space, its method is as follows:

{

1, go to replace continuous print coordinate figure with the discrete coordinate figure of each dimension coordinate axle.

2, adaptive value calculates according to following (4) formula:

fitness＝α ₁f ₁+α ₂f ₂(4)

$f_1=\underset{k=1}{\overset{K}{\Σ}}{q}_k/\underset{i=1}{\overset{N}{\Σ}}{q}_i---\left(5\right)$

$f_2=\underset{i=1}{\overset{N}{\Σ}}{l}_i/\underset{k=1}{\overset{K}{\Σ}}{l}_k---\left(6\right)$

3, the optimizing process of LSE is as follows:

{

1st step: initialization

Living individual number in initial life group is set;

Initial living space scope is set;

The number of the offspring that each existence living individual can be bred is set;

Nutrient distribution function f () is set;

Maximum reproductive order of generation is set.

2nd step: be each the living individual uniform distribution living space in initial population, briefly, distribute living space equably for each individuality exactly;

3rd step: each living individual obtains nutrition in its living space, the size of nutrition is exactly the value of distribution function f () at its living space coordinate mid point.

4th step: to each living individual, does following circulation

{

4-1 walks: calculated threshold nutrition, and its value equals the mean value that current all living individuals obtain nutrition;

4-2 walks: the nutrition of comparing living individual, is more than or equal to threshold value nutrition person existence, otherwise dead;

4-3 walks: the living individual of all existence, in their living space, breed multiple offspring;

4-4 walks: these offsprings inherit and the living space sharing their older generations equably, also, equably for offspring distributes subspace of surviving;

4-5 walks: the former generation's death raised up seed;

4-6 walks: each offspring draws nourishment from its living space, and the size of nutrition is exactly the value of distribution function f () at its living space coordinate mid point.

}

5th step: judge whether to reach maximum reproductive order of generation, if reached, then jumps to the 8th step.

6th step: to each offspring's living individual, do following circulation

{

6-1 walks: calculated threshold nutrition, and its value equals the mean value that current all offspring's living individuals obtain nutrition;

6-2 walks: the nutrition of comparing offspring's living individual, is more than or equal to threshold value nutrition person existence, otherwise dead;

6-3 walks: the living individual of all offspring's existence, breeds multiple new offspring in their living space;

6-4 walks: these new offsprings inherit and the living space sharing their older generations equably, also, equably for new offsprings distribute subspace of surviving;

6-5 walks: the former generation's death raised up seed;

6-6 walks: each new offspring draws nourishment from its living space, and the size of nutrition is exactly the value of distribution function f () at its living space coordinate mid point.

}

7th step: jump to the 5th step.

8th step: terminate.

}

4-7 walks: issue race's header;

4-8 walks: jump to 4-4 step.

}

5th step: the method provided according to LEACH, sets up group, creates dispatch list, then transmits data.

6th step: jump to the 3rd step.

7th step: terminate.

Beneficial effect of the present invention: use the inventive method, adopts living space evolution intelligent algorithm to optimize clustering routing structure, improves wireless sensor network data efficiency of transmission, greatly improve the useful life of wireless sensor network.

Accompanying drawing explanation

Fig. 1 is network architecture schematic diagram of the present invention.

Wherein: Candidate1 and Candidate2 is candidate 1 and candidate 2; SinkNode is aggregation node.

Fig. 2 is two-dimentional living space schematic diagram of the present invention.

Wherein: Candidate1 and Candidate2 is candidate 1 and candidate 2.

Fig. 3 a ~ 3i is verification the verifying results figure of the present invention.

Wherein: (a) Sphere; (b) WeightedSphere; (c) Schwefel ' s; (d) Rosenbrock; (e) Rastrigin; (f) Griewank; (g) Schwefei; (h) Ackley; (i) Schaffer ' sf6; Generation is reproductive order of generation, and Minimum is minimum value.

Table 1 is the Mathematical Modeling about test function of the present invention and relevant parameters.

In table: Name representative function name, Formula expression, Range represents independent variable span, and Optimalvalue represents optimal value.

Table 2 is that the relevant parameters that LSE of the present invention tests is arranged.

In table: Function representative function name, Dimension representation dimension, Initialization represents initialization value.

Embodiment

Describe the present invention below in conjunction with accompanying drawing.

1 problem describes

1.1 networks and capability model

Suppose that network model is as follows:

(1) have N number of sensor node to be distributed in sensitizing range, each node has unique identification number.

(2) all nodes are homogeneities, that is, have identical primary power and transmission range.

(3) all nodes have power adjustments ability, can communicate with aggregation node.

(4) aggregation node is positioned at suitable position, the periphery of sensitizing range, and has enough energy.

(5) aggregation node and sensor node position are fixed, and the positional information of mutual energy perception the other side.

Energy model hypothesis is as follows:

Transmission k bit data to d rice distance needed for energy ezpenditure, E _tx(k, d), its size is:

E _Tx(k,d)＝kE _elec+kε _fsd ²(d＜d ₀)

E _Tx(k,d)＝kE _elec+kε _mpd ⁴(d≥d ₀)(1)

Receive the energy that these information consume, E _rx(k), its size is:

E _Rx(k)＝kE _elec(2)

Process the energy required for these information, E _da-fus(k), its size is:

E _da-fus(k)＝kE _da(3)

Wherein, E _electranstation mission circuit and the energy ezpenditure accepting circuit; E _dait is the energy ezpenditure of deal with data; ε _fsand ε _mpthe power loss of free space model and multichannel attenuation model respectively.D ₀it is transmission range threshold value.

The target of 1.2 clustering routings

The target of clustering routing is from N number of sensor node, select K race's head, thus forms K race.But following requirement must be met: remaining energy more and from aggregation node more close to, then the possibility becoming race's head node is larger.Its fitness function is designed to:

fitness＝α ₁f ₁+α ₂f ₂(4)

$f_1=\underset{k=1}{\overset{K}{\Σ}}{q}_k/\underset{i=1}{\overset{N}{\Σ}}{q}_i---\left(5\right)$

$f_2=\underset{i=1}{\overset{N}{\Σ}}{l}_i/\underset{k=1}{\overset{K}{\Σ}}{l}_k---\left(6\right)$

Wherein, α ₁and α ₂be weights coefficients, and meet α ₁+ α ₂=1; q _iit is the dump energy of i-th node in network; q _kthe dump energy of ShikGe race head node; f ₁it is the evaluation factor of ability; l _ithe distance of i-th node to aggregation node; l _kshikGe race head node is to the distance of aggregation node; f ₂it is the evaluation factor of distance.

From equation (4), select K race's head from N number of sensor node, its essence is that one is selected K race's head to make its fitness function value become optimum, this is a typical optimization problem.

The living space of 2 establishment life

Different from continuous function, the problem selecting K race's head from N number of node is a discrete event.How to apply LSE to discrete event problem, its key is the living space how creating life.

First might as well analyze, what does is life in this problem? in continuous function problem, life is exactly a point in the variable space.But now, what is variable in clustering route protocol? in fact, this K candidate is exactly variable.Because each candidate in K candidate can value be that numbering 1 arrives numbering N, that is, each candidate is exactly a variable.By this method, can create the living space of life, wherein, each candidate is exactly the coordinate in this space, as follows:

The living space of life=candidate 1, candidate 2 ..., each candidate of candidate K| can get some sequence numbers 1 to N}.

Such as, from 100 nodes be randomly dispersed in a sensitizing range, 2 race's head nodes are selected.First, each node serial number should be given according to certain order, such as from numbering 1 to numbering 100, as shown in Figure 1.So, living space just can create out, as shown in Figure 2.Can see, this living space is exactly a discrete two-dimensional space.

In Fig. 2, the probable value of candidate 1 and candidate 2 is all that numbering 1 arrives numbering 100, and what numbering wherein represented is this node having this numbering.Clearly, these two candidates cannot get some identical numberings simultaneously, and therefore, the position of any life can not be thought in the position being labeled as stain in figure.But be labeled as the position of white point in figure, then constitute the living space of life, this is a two-dimentional living space.In this living space, any one is labeled as the position of white point, and represent two corresponding candidates, namely abscissa represents the numbering of candidate 1, and ordinate represents the numbering of candidate 2.That is, the numbering of these two candidates is exactly the x of life at living space and the coordinate figure of y-axis, and this position being labeled as white point just represents a life.

Be more than select 2 race's head nodes to be example, can obtain following conclusion from analysis above: if select K race's head from N number of node, so, the living space of life is exactly the space of a K dimension.

3LSE optimizes discrete living space

Does is next problem exactly how to use LSE to optimize this living space after founding discrete living space? in fact, the using method of this and continuous space is identical, only goes to replace continuous print coordinate figure with the discrete coordinate figure of x and y-axis.

Such as, in the discrete living space of the two dimension shown in Fig. 2, the coordinate figure of x and y can only be got 1,2 ..., each discrete value of 100}.In addition, remaining process is identical with the optimization method of continuous function.

4 clustering route protocols optimized based on LSE

The clustering route protocol step optimized based on LSE is as follows:

1st step: initialization

The nodes of sensor network is set;

According to (1), (2), (3) formula, energy parameter is set;

According to (4) formula, fitness function parameter is set;

The maximum number of transmission data is set.

2nd step: the method provided according to LEACH, selects race's head, sets up group, creates dispatch list, then transmit data.

3rd step: judge whether the maximum reaching transmission data.If reach, then jump to the 7th step.

4th step: select to be optimized to race's head by LSE

{

4-1 walks: to each node serial number of sensitizing range;

4-2 walks: arrange race's head number;

4-3 walks: according to the calculated capacity of each node, arrange maximum iteration time;

4-4 walks: judge whether to reach maximum iteration time.If reached, then jump to the 5th step;

4-5 walks: be optimized living space according to the method that the living space of the establishment life of the 2nd trifle and the LSE of the 3rd trifle are optimized discrete living space and provided, wherein adaptive value calculates according to (4) formula;

4-6 walks: issue race's header;

4-7 walks: jump to 4-4 step.

}

5th step: the method provided according to LEACH, sets up group, creates dispatch list, then transmits data.

6th step: jump to the 3rd step.

7th step: terminate.

Operation principle of the present invention and the course of work are described:

1, operation principle of the present invention and the course of work

Operation principle of the present invention and the course of work can be expressed as follows by a flow process:

1st step: initialization

{

1-1 walks: the nodes arranging sensor network;

1-2 walks: the energy parameter arranging following (1), (2), (3) formula:

Transmission k bit data to d rice distance needed for energy ezpenditure, E _tx(k, d), its size is:

E _Tx(k,d)＝kE _elec+kε _fsd ²(d＜d ₀)

E _Tx(k,d)＝kE _elec+kε _mpd ⁴(d≥d ₀)(1)

Receive the energy that these information consume, E _rx(k), its size is:

E _Rx(k)＝kE _elec(2)

Process the energy required for these information, E _da-fus(k), its size is:

E _da-fus(k)＝kE _da(3)

Wherein, E _electranstation mission circuit and the energy ezpenditure accepting circuit; E _dait is the energy ezpenditure of deal with data; ε _fsand ε _mpthe power loss of free space model and multichannel attenuation model respectively.D ₀it is transmission range threshold value.

1-3 walks: arrange fitness function parameter according to following (4) formula:

fitness＝α ₁f ₁+α ₂f ₂(4)

$f_1=\underset{k=1}{\overset{K}{\Σ}}{q}_k/\underset{i=1}{\overset{N}{\Σ}}{q}_i---\left(5\right)$

$f_2=\underset{i=1}{\overset{N}{\Σ}}{l}_i/\underset{k=1}{\overset{K}{\Σ}}{l}_k---\left(6\right)$

Wherein, α ₁and α ₂be weights coefficients, and meet α ₁+ α ₂=1; q _iit is the dump energy of i-th node in network; q _kthe dump energy of ShikGe race head node; f ₁it is the evaluation factor of ability; l _ithe distance of i-th node to aggregation node; l _kshikGe race head node is to the distance of aggregation node; f ₂it is the evaluation factor of distance.

1-4 walks: the maximum number arranging transmission data.

}

2nd step: the method provided according to LEACH, selects race's head, sets up group, creates dispatch list, then transmit data.

3rd step: judge whether the maximum reaching transmission data.If reach, then jump to the 7th step.

4th step: select to be optimized to race's head by LSE intelligent algorithm

{

4-1 walks: to each node serial number of sensitizing range;

4-2 walks: arrange race's head number;

4-3 walks: according to the calculated capacity of each node, arrange maximum iteration time;

4-4 walks: judge whether to reach maximum iteration time.If reached, then jump to the 5th step;

4-5 walks: the living space creating life, and its method is as follows:

The living space of life=candidate 1, candidate 2 ..., each candidate of candidate K| can get some sequence numbers 1 to N}.

4-6 walks: optimize discrete living space, its method is as follows:

{

1, go to replace continuous print coordinate figure with the discrete coordinate figure of each dimension coordinate axle.

2, adaptive value calculates according to following (4) formula:

fitness＝α ₁f ₁+α ₂f ₂(4)

$f_1=\underset{k=1}{\overset{K}{\Σ}}{q}_k/\underset{i=1}{\overset{N}{\Σ}}{q}_i---\left(5\right)$

$f_2=\underset{i=1}{\overset{N}{\Σ}}{l}_i/\underset{k=1}{\overset{K}{\Σ}}{l}_k---\left(6\right)$

3, the optimizing process of LSE is as follows:

{

1st step: initialization

Living individual number in initial life group is set;

Initial living space scope is set;

The number of the offspring that each existence living individual can be bred is set;

Nutrient distribution function f () is set;

Maximum reproductive order of generation is set.

2nd step: be each the living individual uniform distribution living space in initial population, briefly, distribute living space equably for each individuality exactly;

3rd step: each living individual obtains nutrition in its living space, the size of nutrition is exactly the value of distribution function f () at its living space coordinate mid point.

4th step: to each living individual, does following circulation

{

4-1 walks: calculated threshold nutrition, and its value equals the mean value that current all living individuals obtain nutrition;

4-2 walks: the nutrition of comparing living individual, is more than or equal to threshold value nutrition person existence, otherwise dead;

4-3 walks: the living individual of all existence, in their living space, breed multiple offspring;

4-4 walks: these offsprings inherit and the living space sharing their older generations equably, also, equably for offspring distributes subspace of surviving;

4-5 walks: the former generation's death raised up seed;

4-6 walks: each offspring draws nourishment from its living space, and the size of nutrition is exactly the value of distribution function f () at its living space coordinate mid point.

}

5th step: judge whether to reach maximum reproductive order of generation, if reached, then jumps to the 8th step.

6th step: to each offspring's living individual, do following circulation

{

6-1 walks: calculated threshold nutrition, and its value equals the mean value that current all offspring's living individuals obtain nutrition;

6-2 walks: the nutrition of comparing offspring's living individual, is more than or equal to threshold value nutrition person existence, otherwise dead;

6-3 walks: the living individual of all offspring's existence, breeds multiple new offspring in their living space;

6-4 walks: these new offsprings inherit and the living space sharing their older generations equably, also, equably for new offsprings distribute subspace of surviving;

6-5 walks: the former generation's death raised up seed;

6-6 walks: each new offspring draws nourishment from its living space, and the size of nutrition is exactly the value of distribution function f () at its living space coordinate mid point.

}

7th step: jump to the 5th step.

8th step: terminate.

}

4-7 walks: issue race's header;

4-8 walks: jump to 4-4 step.

}

5th step: the method provided according to LEACH, sets up group, creates dispatch list, then transmits data.

6th step: jump to the 3rd step.

7th step: terminate.

A brief description is done to the LSE optimizing process in said method below.

1st step completes Initialize installation, comprise initial life colony is set individual number, living space scope, each existence individual can raise up seed number, breeding maximum algebraically, and nutrient distribution function etc.2nd step distributes living space equably for each living individual.3rd step, each living individual obtains nutrition in its living space, and its computational methods get the value of nutrition function at its living space mid point.

4th step, for each living individual, enters into a circulation.1st step of circulation, calculates nutrition threshold value, and its value equals the mean value that current all individualities obtain nutrition; In the 2nd step of circulation, compare, the living individual that acquisition nutritive value is more than or equal to nutrition threshold value survives, other individual death; 3rd step of circulation, the individuality survived breeds multiple offspring in its living space; 4th step of circulation, the offspring bred inherits the living space of its former generation equably, that is, for offspring divides gamete living space equably; 5th step of circulation, the former generation raised up seed is dead at once; 6th step of circulation, each offspring's living individual obtains nutrition in its existence subspace, and its computational methods get the value of nutrition function at its existence subspace mid point.

5th step, judges whether the maximum reaching reproductive order of generation.If reached, then jump to the 8th step, thus terminate whole computational process.

6th step, for each offspring's living individual, enter into a circulation, its method of operation is identical with computational process and the 4th step.

7th step, jumps to the 5th step.

8th step, terminates whole computational process.

Breed it is important to note that the 4th step and the 6th step complete living space evolution respectively with many offsprings.On the one hand in the evolutionary process of living space, a part of living individual disappears together with their living space.On the other hand, many offspring's breedings continuously for biological population injects fresh living individual, so just can ensure that the life in life colony can not gradually reduce.Therefore, many offspring's breedings are the preconditions and guarantees realizing living space evolution.

2, validation verification of the present invention

In order to test to validity of the present invention, standard test functions (BenchmarkFunctions) is adopted to test the core content LSE intelligent algorithm in this method.9 standard test functions comprise 3 unimodal functions and 6 Solving Multimodal Functions and are used for test, and its target calculates these functional minimum value.Table 1 is Mathematical Modeling and the relevant parameters of these test functions.The relevant parameters of LSE test arranges as shown in table 2.

Optimum results as shown in Figure 3.As seen from Figure 3, for all functions, the minimum value that offspring searches is all less than former generation, and this just illustrates that life is constantly evolved to high-quality living space.Can see simultaneously, for Rosenbrock, directly obtain its minimum value, and for other functions, along with the increase of reproductive order of generation, the minimum value searched be more and more close to functional minimum value.Therefore, LSE intelligent algorithm is effective for the optimization problem of these standard test functions.

Table 1

Table 2

Claims (1)

1. adopt a wireless sensor network clustering routing for living space evolution intelligent algorithm, its feature comprises the following steps:

1st step: initialization

1-1 walks: the nodes arranging sensor network;

1-2 walks: the energy parameter arranging following (1), (2), (3) formula:

Transmission k bit data to d rice distance needed for energy ezpenditure, E _tx(k, d), its size is:

E _Tx(k,d)＝kE _elec+kε _fsd ²(d＜d ₀)

E _Tx(k,d)＝kE _elec+kε _mpd ⁴(d≥d ₀)(1)

Receive the energy that these information consume, E _rx(k), its size is:

E _Rx(k)＝kE _elec(2)

Process the energy required for these information, E _da-fus(k), its size is:

E _da-fus(k)＝kE _da(3)

Wherein, E _electranstation mission circuit and the energy ezpenditure accepting circuit; E _dait is the energy ezpenditure of deal with data; ε _fsand ε _mpthe power loss of free space model and multichannel attenuation model respectively, d ₀it is transmission range threshold value;

1-3 walks: arrange fitness function parameter according to following (4) formula:

fitness＝α ₁f ₁+α ₂f ₂(4)

Wherein, α ₁and α ₂be weights coefficients, and meet α ₁+ α ₂=1; q _iit is the dump energy of i-th node in network; q _kthe dump energy of ShikGe race head node; f ₁it is the evaluation factor of ability; l _ithe distance of i-th node to aggregation node; l _kshikGe race head node is to the distance of aggregation node; f ₂it is the evaluation factor of distance;

1-4 walks: the maximum number arranging transmission data;

2nd step: the method provided according to LEACH, selects race's head, sets up group, creates dispatch list, then transmit data;

3rd step: judge whether the maximum reaching transmission data, if reach, then jump to the 7th step;

4th step: select to be optimized to race's head by LSE intelligent algorithm

4-1 walks: to each node serial number of sensitizing range;

4-2 walks: arrange race's head number;

4-3 walks: according to the calculated capacity of each node, arrange maximum iteration time;

4-4 walks: judge whether to reach maximum iteration time, if reached, then jump to the 5th step;

4-5 walks: the living space creating life, and its method is as follows:

The living space of life=candidate 1, candidate 2 ..., each candidate of candidate K| can get some sequence numbers 1 to N};

4-6 walks: optimize discrete living space, its method is as follows:

4-6-1 walks: go to replace continuous print coordinate figure with the discrete coordinate figure of each dimension coordinate axle;

4-6-2 walks: adaptive value calculates according to following (4) formula:

fitness＝α ₁f ₁+α ₂f ₂(4)

4-6-3 walks: the optimizing process of LSE is as follows:

4-6-3-1 walks: initialization:

Living individual number in initial life group is set;

Initial living space scope is set;

The number of the offspring that each existence living individual can be bred is set;

Nutrient distribution function f () is set;

Maximum reproductive order of generation is set;

4-6-3-2 walks: be each the living individual uniform distribution living space in initial population, briefly, distribute living space equably exactly for each individuality;

4-6-3-3 walks: each living individual obtains nutrition in its living space, and the size of nutrition is exactly the value of distribution function f () at its living space coordinate mid point;

4-6-3-4 walks: to each living individual, do following circulation:

4-6-3-4-1 walks: calculated threshold nutrition, and its value equals the mean value that current all living individuals obtain nutrition;

4-6-3-4-2 walks: the nutrition of comparing living individual, is more than or equal to threshold value nutrition person existence, otherwise dead;

4-6-3-4-3 walks: the living individual of all existence, in their living space, breed multiple offspring;

4-6-3-4-4 walks: these offsprings inherit and the living space sharing their older generations equably, also, equably for offspring distributes subspace of surviving;

4-6-3-4-5 walks: the former generation's death raised up seed;

4-6-3-4-6 walks: each offspring draws nourishment from its living space, and the size of nutrition is exactly the value of distribution function f () at its living space coordinate mid point;

4-6-3-5 walks: judge whether to reach maximum reproductive order of generation, if reached, then jump to the 8th step;

4-6-3-6 walks: to each offspring's living individual, do following circulation:

4-6-3-6-1 walks: calculated threshold nutrition, and its value equals the mean value that current all offspring's living individuals obtain nutrition;

4-6-3-6-2 walks: the nutrition of comparing offspring's living individual, is more than or equal to threshold value nutrition person existence, otherwise dead;

4-6-3-6-3 walks: the living individual of all offspring's existence, breeds multiple new offspring in their living space;

4-6-3-6-4 walks: these new offsprings inherit and the living space sharing their older generations equably, also, equably for new offsprings distribute subspace of surviving;

4-6-3-6-5 walks: the former generation's death raised up seed;

4-6-3-6-6 walks: each new offspring draws nourishment from its living space, and the size of nutrition is exactly the value of distribution function f () at its living space coordinate mid point;

4-6-3-7 walks: jump to the 5th step;

4-6-3-8 walks: terminate;

4-7 walks: issue race's header;

4-8 walks: jump to 4-4 step;

5th step: the method provided according to LEACH, sets up group, creates dispatch list, then transmits data;

6th step: jump to the 3rd step;

7th step: terminate.