CN106508027B - Deployment diagram point army mark automatic avoiding method based on genetic algorithm - Google Patents

Abstract

The invention belongs to military affairs will figure manufacture technology field, and in particular to a kind of deployment diagram point army mark based on genetic algorithm avoids algorithm.Purpose be solve deployment diagram quick Fabrication during there are problems that army's mark gland, suitable position is found for army's mark automatically by computer program, loaded down with trivial details manual work is reduced, is improved the automaticity of deployment diagram making.Including following key step：(1) definition wants icon to paint rule；(2) the plotting Environmental Evaluation Model of design form, quantification；(3) design army mark collision detection algorithm；(4) global optimization is carried out with genetic algorithm.It is an advantage of the invention that the army's of efficiently solving mark gland, collision problem, provide guarantee for deployment diagram quick Fabrication, with good operability, are easy to promote the use of.

Description

Deployment diagram point army mark automatic avoiding method based on genetic algorithm

Technical field

To scheme the deployment diagram in making field the present invention relates to military geographical information system and military affairs and automatically generate Technology, and in particular to the point army mark automatic avoiding method based on genetic algorithm.

Background technology

Deployment diagram is one kind of conventional operation chart, is mainly used in the key elements such as sign troops, mechanism, equipment Deployed position.

At present, deployment map generalization means, which mainly has, mark and draw by hand and computer plotting.It is soft using computer Part carry out deployment diagram mark and draw with high precision, speed is fast, easily change the features such as, but in some application systems , it is necessary to quickly generate large batch of deployment diagram in system, marked and drawed and be still difficult to completely using computer software Its time requirement of foot.Because in deployment diagram generating process, if directly general's mark is as the geography where it Position is marked and drawed, and often produces the phenomenons such as army's mark gland, conflict, that is, army's mark conflict.

And manually mark and draw operation in, mark and draw personnel can according to actual conditions, by way of index wire, For the suitable position of army's mark selection, it is to avoid the generation of army's mark conflict.But, the deployment map generalization time is again It cannot be guaranteed.

Quickly, it is suitable how to be found automatically for army's mark using computer program during batch making deployment diagram Position be a very challenging job, be also one of long-standing problem military mapping personnel important Technical problem.

Because the military symbol system used both at home and abroad is different, therefore external correlation technique enters march to the country Mark avoids research and does not have directive significance.At present, the correlation of the domestic automatic avoidance technology of the not army of finding mark is ground Study carefully army's mark in document, deployment diagram generating process to avoid mainly by manual mode progress.

The content of the invention

It is an object of the invention to provide it is a kind of can quickly, the method for batch making deployment diagram, solve deployment There are problems that during figure quick Fabrication army mark gland, by computer program automatically for army mark Suitable position is found, cumbersome handwork is reduced, the automaticity that deployment diagram makes is improved.

What the present invention was realized in：A kind of deployment diagram point army mark automatic avoiding method based on genetic algorithm, Comprise the following steps：

The first step, first according to point army of army target longitude and latitude, using the clustering algorithm based on grid to point Army's mark is clustered, the army of distinguishing mark compact district and non-dense set area；

Second step, determines coding framework；Deployment diagram point army mark Avoidance is abstracted as competition for space Combinatorial optimization problem, so that coded system uses integer coding；Plotting scheme deployment diagram is whole with one Number vector represents that vectorial each component represents a Ge Dian armies target and marks and draws position；Using with N number of point The vector of amount represents alternative plotting scheme, and vectorial each component represents a Ge Dian armies target and mark and draw position,

The point army mark alternate location model used in the system has 5 position models and 10 position models, for Point army mark in close quarters, its corresponding component value is the integer value in [0,9], for non-dense set area Point army mark in domain, its corresponding component value is the integer value in [0,4]；

3rd step, determines initial population；N number of vector is randomly generated using a kind of method of completely random to make For initial population pop₀；For the point army mark in close quarters, its corresponding chromogene position takes [0,9] Interior random integer value, is marked for the point army in non-dense region, and its corresponding chromogene position takes [0,4] Interior random integer value；

4th step, determines fitness function；Army's mark conflict is detected using the method for grid map, built Vertical Environmental Evaluation Model is to pop₀Evaluated, obtain pop₀Evaluation of estimate；Define fitness function such as Under：

Fit (JB)=E (JB)

Wherein, E (JB) represents plotting scheme JB quality evaluation value, and fit (JB) represents plotting scheme JB correspondences Chromosome fitness value；

Specifically, system chooses 2 independent factors：Conflict, positional priority；Define first single Then the two evaluation functions are combined into single army and marked by army's target conflict, positional priority evaluation function Quality evaluation function, finally by single army's target quality evaluation function it is cumulative obtain view picture will figure quality Evaluation function；

(4.1) conflict evaluation function；

For each army mark, mutually conflict if marked with other armies, the evaluation function value that conflicted is set It is set to -1, is otherwise set as 0；Single army's target conflict evaluation function is：

Wherein L_iRepresent i-th of army's mark, RectIntersect (L_i, L_j) represent i-th and j army mark and collide with each other, E_Conflict(L_i) represent i-th of army's target conflict evaluation of estimate；

(4.2) positional priority evaluation function；

Make Pos_j(L_i) represent L_iJ-th plotting position, Order (Pos_j(L_i)) represent alternate location sequence number, position Sequence number predefined is put, position evaluation function is defined as to the difference of 10 and position number, i.e.,

E_Position(L_i)=10-Order (Pos_j(L_i))；

(4.3) composite evaluation function；

The quality evaluation fonction composition of single factor is taken to the side of weighted sum into total quality evaluation function Single two kinds of evaluations of estimate of army's target, are first added by being multiplied by weight factor and obtain a total quality by formula again Evaluation of estimate, then single army's target quality evaluation value added up entirely will figure plotting quality evaluation Value, i.e.,：

E(L_i)=W_ConflictE_Conflict(L_i)+W_PositionE_Position(L_i)

Wherein W_ConflictRepresent the weight coefficient of the conflict factor, W_PositionRepresent the weight coefficient of the positional priority factor；

5th step, performs selection operation to initial population according to the height of fixed fitness function, obtains To male parent population F_k；

Using the roulette wheel selection with elitism strategy, i.e., the optimum individual in parent is first retained in offspring In colony, other chromosomes are then selected in accordance with the following steps：

1. the fitness value fit (V of individual are calculated_i) (i=1,2 ..., n), wherein V_iFor a dyeing in potential solution colony Body；

2. the accumulative fitness value Accfit (V of individual are calculated_i) (i=1 2 ..., adds up fitness value n) and relatively RelAccfit(V_i) (i=1,2 ..., n), wherein

RelAccfit(V_i)=Accfit (V_i)/Accfit(V_n)

3. the random number r in one [0,1] is generated, if RelAccfit (V_i-1) ＜ r≤RelAccfit (V_i) (i=1,2 ..., n), that Select individual V_i, it is assumed herein that RelAccfit (V₀)=0；

6th step, by male parent population F_kWith probability P_cPopulation C is produced through intersecting_k；

Using random single-point type interleaved mode, i.e., a gene is randomly selected on two father's chromosome strings Position, gene is exchanged with each other using the point as boundary；Before being intersected, first selected chromosome is pressed Packet pairing is carried out according to certain order；It is random to add or move if the total chromosome number chosen is odd number Except a chromosome；

7th step, to population C_kPerform gland conflict component mutation operation and obtain population pop of new generation_k+1；Adopt With gland conflict position variation mode, that is, the point army of all generation glands and conflict phenomenon is chosen to mark corresponding base Because of position, the new plotting of random generation one is position encoded to replace original gene position, i.e., population of new generation； The thus army's of make use of mark gland and the heuristic information of conflict so that improve bad subsolution, got well The possibility of subsolution is bigger；

Army's mark conflict refers to that there is overlapping part in the shared plotting region of Liang Ge armies mark, including index wire intersects； The detection that system is conflicted using the method for grid map；Region will be marked and drawed first according to screen pixel grid Change, then calculate the grid shared by it using army's cursor position and its size, finally by judge Liang Ge armies mark Whether identical grid is occupied come the target conflict of the army of detection；

8th step, judges whether to meet end condition；

The present invention judges that colony has developed maturation using following two condition：, can be with through repeatedly calculating Stably obtain approximate optimum individual；When continuing to develop, optimum individual is without significantly improving；

If meeting end condition, the point military standard plotting scheme optimized；Terminate this method；

If not meeting end condition, the 5th step is returned, continues computing.

A kind of deployment diagram point army mark automatic avoiding method based on genetic algorithm as described above, wherein,

In the third step, the scale N of population is set as 8；

In the 4th step, the weight factor value of each in evaluation function is as follows：W_Conflict=100 W_Position=1；It is single Individual army's mark is marked and drawed quality evaluation function and also reformed into：E(L_i)=100*E_Conflict(L_i)+E_Position(L_i)；

In the 6th step, crossover probability P is taken_c=0.5.

It is an advantage of the invention that solving deployment diagram generating process midpoint army marks automatic Avoidance, deployment diagram Plotting is spatially to place a number of military symbols for meeting ad hoc rules in limited background map, It is substantially also a competition for space optimization problem.Army's mark that will be in figure manufacturing process for long-standing problem Collision problem, deployment diagram point army is marked and avoids the combinatorial optimization problem for being abstracted into competition for space.It is every first Ge Dian armies mark sets the alternative plotting position of limited quantity, is then each alternate location and conflict situations Evaluation function is set up, so as to set up the plotting Environmental Evaluation Model of view picture deployment diagram, heredity calculation is finally utilized Method is iterated optimization to evaluation function value, then obtains satisfactorily marking and drawing result.

Following technique effect can be reached using the present invention：

1st, the use of the technology makes deployment diagram Production Time shorten to averagely less than 2 minutes, hence it is evident that improve Command effectiveness of air defense；All data come from database, improve the degree of accuracy of deployment diagram；Automaticity Height, greatly reduces the cumbersome handwork of commanding and marks on a map, improve simulated training and war preparedness is on duty Level；Flat, war is combined, and facilitates commander and command post commands troops to subordinate unit, is strengthened The fighting capacity and wartime emergency reaction ability of army.

2nd, genetic algorithm is a kind of healthy and strong, easy extension, the algorithm of function admirable, is good at processing complexity high The problem of.It can solve the problem that deployment diagram marks and draws the optimization problem of quality, so as to realize that army's target is kept away automatically Allow.Also, during variation, the method for the invention takes full advantage of gland conflict between army's mark Information, such a form relative at random enter row variation, be more applicable for solve army mark avoid automatically ask Topic, accelerates the process for obtaining optimal solution.

Brief description of the drawings

Fig. 1 is collision detection figure；

Fig. 2 is that army's target alternatively marks and draws position；

Fig. 3 is genetic algorithm for solving flow chart；

Fig. 4 is that chromosome single-point type intersects schematic diagram；

Fig. 5 is the evolutionary process of genetic algorithm.

Embodiment

The present invention is described further with specific embodiment below in conjunction with the accompanying drawings：

The basic step that conventional genetic algorithm solves problem is as follows：

A) solution of problem is expressed as chromosome, each chromosome represents a feasible solution of problem；

B) a number of chromosome is randomly generated as initial population pop₀, the population is exactly that problem is feasible One set of solution；

C) initial chromosome group is placed among " environment " of problem, and provides the chromosome of each in population To the adaptive value of problem context；

D) according to the height of each chromosome fitness to initial population pop₀(or pop_k) perform selection behaviour Make, obtain male parent population F_k；

E) by male parent population F_kWith crossover probability P_cPopulation C is produced through intersecting_k；

F) to population C_kWith mutation probability P_mPerform mutation operation and obtain population pop of new generation_k+1；

G) return to step c.

The present invention in order to utilize above-mentioned genetic algorithm, by solve problem technical scheme it is abstract be 4 steps, Wherein, the committed step for solving the automatic Avoidance of deployment diagram point army mark using genetic algorithm includes：It is determined that Coding framework；Produce initial population；Determine fitness function；The heredity such as design alternative, intersection, variation Operator.The important parameter of algorithm includes population scale, intersection and mutation probability, end condition etc..

Step 1, definition want icon to paint rule；

The rule for wanting icon to paint has：

(1) configuration of two or more command post is marked and drawn corresponding simultaneously at same place on a flagpole Flag face, i.e., same to bit combination, and army's mark size and upper-lower position are distinguished according to attributes such as army's rank, services；

(2) army's target position can not be collided with each other；

Do not allow army's mark conflict situations as shown in Figure 1 occur.

(3) when several labels are located at same place, these labels are placed near, referred to index wire To allocation position.The time efficiency in view of computer disposal marks the abundant generation of alternate location with army simultaneously Table, system sets 5 or 10 alternate locations for each point army mark, as shown in Figure 2.

Environmental Evaluation Model is marked and drawed in step 2, design；

The factor that quality is marked and drawed in influence is a lot, and some of which factor can provide the description of formalization, some The description that factor will provide formalization is extremely difficult.System chooses 2 independent factors：Conflict, position Priority.Single army's target conflict, positional priority evaluation function are defined first, then comment the two Valency fonction composition finally tires out single army's target quality evaluation function into single army's target quality evaluation function Plus obtain view picture will figure quality evaluation function.

(2.1) conflict evaluation function；

When Liang Ge armies cursor position is clashed, the size of location overlap amount is unimportant, so, conflict Evaluation function is only evaluated conflict behavior in itself, and does not measure the size of lap, is not also differentiated between It is that Liang Ge armies mark is clashed or multiple armies mark is while clash.For each army mark, if with Other army's marks mutually conflict, then the evaluation function value that conflicted is set as -1, is otherwise set as 0.Single army Target conflict evaluation function be

Wherein L_iRepresent i-th of army's mark, RectIntersect (L_i, L_j) represent i-th and j army mark and collide with each other, E_Conflict(L_i) represent i-th of army's target conflict evaluation of estimate.

(2.2) positional priority evaluation function；

Make Pos_j(L_i) represent L_iJ-th plotting position, Order (Pos_j(L_i)) represent alternate location sequence number, position Sequence number predefined is put, as shown in fig. 1, position number illustrates the preference to this position in deployment diagram Degree, is defined as sequence number by top-priority position relatively low, position evaluation function is defined as into 10 and position The difference of sequence number, i.e.,

E_Position(L_i)=10-Order (Pos_j(L_i))

(2.3) composite evaluation function；

The quality evaluation fonction composition of single factor is taken to the side of weighted sum into total quality evaluation function Single two kinds of evaluations of estimate of army's target, are first added by being multiplied by weight factor and obtain a total quality by formula again Evaluation of estimate, then single army's target quality evaluation value added up entirely will figure plotting quality evaluation Value, i.e.,

E(L_i)=W_ConflictE_Conflict(L_i)+W_PositionE_Position(L_i)

Wherein W_ConflictRepresent the weight coefficient of the conflict factor, W_PositionThe weight coefficient of the positional priority factor is represented, Learnt from experience, W_Conflict＞ ＞ W_Position, because not conflicting but without being to connect in accordance with the plotting of aesthetic principle Receive, on the contrary even if marking and drawing good again, there is conflict in army cursor position, it is such will figure be also to be difficult to receive 's.Wherein each weight factor value is as follows：W_Conflict=100 W_Position=1

Single army's mark is marked and drawed quality evaluation function and also reformed into：E(L_i)=100*E_Conflict(L_i)+E_Position(L_i)

Step 3, design army mark collision detection method

Army's mark conflict refers to that there is overlapping part in the shared plotting region of Liang Ge armies mark, including index wire intersects. The detection that system is conflicted using the method for grid map.Region will be marked and drawed first according to screen pixel grid Change, then calculate the grid shared by it using army's cursor position and its size, finally by judge Liang Ge armies mark Whether identical grid is occupied come the target conflict of the army of detection.

Step 4, with genetic algorithm carry out global optimization；

Used genetic algorithm each step is embodied in Fig. 3；

(1) coding framework is determined；

Coding framework is the form for the potential solution of problem being expressed as chromosome.The plotting scheme of deployment diagram is with one Integer vectors represent that vectorial each component represents a Ge Dian armies target and marks and draws position.

(2) initial population is produced；

Initial population is the initial population of the potential solution of representative of certain scale.The present invention uses a kind of completely random Method, that is, randomly generate the initial population of certain scale.

(3) fitness function is determined；

Fitness function is the function for evaluating the good and bad degree of candidate solution.Problem is marked and drawed for deployment diagram, using step The quality evaluation function as fitness function set up in rapid 2.

(4) genetic operator is designed；

Genetic operator control changes the mode of offspring's composition, including selection opertor, crossover operator, mutation operator.

1) selection opertor

Using the roulette wheel selection with elitism strategy, i.e., the optimum individual in parent is first retained in offspring group In body, other chromosomes are then selected in accordance with the following steps：

1. the fitness value fit (V of individual are calculated_i) (i=1,2 ..., n), wherein V_iFor a dyeing in potential solution colony Body.

2. the accumulative fitness value Accfit (V of individual are calculated_i) (i=1 2 ..., adds up fitness value n) and relatively RelAccfit(V_i) (i=1,2 .., n), wherein

RelAccfit(V_i)=Accfit (V_i)/Accfit(V_n)

3. the random number r in one [0,1] is generated, if RelAccfit (V_i-1) ＜ r≤RelAccfit (V_i) (i=1,2 ..., n), that Select individual V_i, it is assumed herein that RelAccfit (V₀)=0.

2) crossover operator

The present invention is intersected using random single-point type, i.e., a gene is randomly selected on two father's chromosome strings Position is intersected.Before being intersected, first selected chromosome is grouped according to certain order Pairing.It is random to add or remove a chromosome if the total chromosome number chosen is odd number.

3) mutation operator

For integer vectors coding, conventional variation has point type variation and uniform variation.The former selects a single point Position, the latter is relocated at random according to a stencil-chosen point position to institute's reconnaissance position.Using conflict position Variation mode, that is, choose all conflict armies to mark gene position, at random one new plotting position encoded generation of generation For original gene position.

The process carried out in mutation operation is indicated in Fig. 4.

4) algorithm important parameter is determined

1. population scale

Population scale N influences the validity of genetic algorithm.N is too small, and arithmetic result can be very poor or looked at all not To the solution of problem, because too small population invariable number can not provide enough sampled points；N is too big, can increase Amount of calculation.It is general within tens herein for deployment diagram Shang Dian armies target quantity, the scale of population It is set as 8.

2. crossover probability

Crossover probability p_cControl the frequency of crossover operation, p_cIt is too big, the gene hyte of high fitness value can be made Conjunction is deteriorated quickly, p_cToo small, search can come to a halt, general p_cTake 0.25~0.75.In this algorithm 0.5 is taken, it is so average to there are namely 4 chromosomes of half to be intersected every time.

3. mutation probability

Mutation probability p_mIt is the second factor for increasing population diversity, is that conflict position becomes due to what is taken herein Different mode, avoids the need for being p again_mIf fixed value.

4. end condition

Only when the comparative maturity that colony has developed, it just can guarantee that and obtain optimal or satisfactory solution.This hair It is bright to judge that colony has developed maturation using following two condition：Through repeatedly calculating, can stably it obtain To approximate optimum individual；When continuing to develop, optimum individual is without significantly improving.

According to above-mentioned abstract technical scheme, in specific application, method of the present invention have passed through as Under step：

The first step, first according to point army of army target longitude and latitude, using the clustering algorithm based on grid to point Army's mark is clustered, the army of distinguishing mark compact district and non-dense set area.

Clustering algorithm based on grid this be the art common knowledge.

Second step, determines coding framework.Deployment diagram point army mark Avoidance is abstracted as competition for space Combinatorial optimization problem, so that coded system uses integer coding.Plotting scheme deployment diagram is whole with one Number vector represents that vectorial each component represents a Ge Dian armies target and marks and draws position.Using with N number of point The vector of amount represents alternative plotting scheme, and vectorial each component represents a Ge Dian armies target and mark and draw position, The point army mark alternate location model used in the system has 5 position models and 10 position models, so for Point army mark in close quarters, its corresponding component value is the integer value in [0,9], for non-dense set area Point army mark in domain, its corresponding component value is the integer value in [0,4].

3rd step, determines initial population.N number of vector is randomly generated using a kind of method of completely random to make For initial population pop₀(or pop_k).For the point army mark in close quarters, its corresponding chromosome base Because position takes the random integer value in [0,9], marked for the point army in non-dense region, its corresponding chromosome Gene position takes the random integer value in [0,4].For deployment diagram Shang Dian armies target quantity it is general tens it It is interior, therefore the scale N of population is set as 8.

4th step, determines fitness function.Army's mark conflict is detected using the method for grid map, built Vertical Environmental Evaluation Model is to pop₀Evaluated, obtain pop₀Evaluation of estimate.For deployment diagram point army mark Automatic Avoidance, the quality evaluation function of foundation is a kind of measurement that quality is marked and drawed to deployment diagram, and it is examined Considered the factors such as conflict, gland and positional priority, evaluation of estimate is higher, mark and draw quality it is better, just with The requirement represented fitness is consistent, and adaptive value is bigger, and individual is more excellent.Define fitness function as follows：

Fit (JB)=E (JB)

Wherein, E (JB) represents plotting scheme JB quality evaluation value, and fit (JB) represents plotting scheme JB correspondences Chromosome fitness value.

Specifically, system chooses 2 independent factors：Conflict, positional priority；Define first single Then the two evaluation functions are combined into single army and marked by army's target conflict, positional priority evaluation function Quality evaluation function, finally by single army's target quality evaluation function it is cumulative obtain view picture will figure quality Evaluation function；

(4.1) conflict evaluation function；

For each army mark, mutually conflict if marked with other armies, the evaluation function value that conflicted is set It is set to -1, is otherwise set as 0；Single army's target conflict evaluation function is：

Wherein L_iRepresent i-th of army's mark, RectIntersect (L_i, L_j) represent i-th and j army mark and collide with each other, E_Conflict(L_i) represent i-th of army's target conflict evaluation of estimate；

(4.2) positional priority evaluation function；

Make Pos_j(L_i) represent L_iJ-th plotting position, Order (Pos_j(L_i)) represent alternate location sequence number, position Sequence number predefined is put, position evaluation function is defined as to the difference of 10 and position number, i.e.,

E_Position(L_i)=10-Order (Pos_j(L_i))；

(4.3) composite evaluation function；

The quality evaluation fonction composition of single factor is taken to the side of weighted sum into total quality evaluation function Single two kinds of evaluations of estimate of army's target, are first added by being multiplied by weight factor and obtain a total quality by formula again Evaluation of estimate, then single army's target quality evaluation value added up entirely will figure plotting quality evaluation Value, i.e.,：

E(L_i)=W_ConflictE_Conflict(L_i)+W_PositionE_Position(L_i)

Wherein W_ConflictRepresent the weight coefficient of the conflict factor, W_PositionRepresent the weight coefficient of the positional priority factor；

The weight factor value of each in evaluation function is as follows：W_Conflict=100 W_Position=1；Single army's mark marks and draws matter Amount evaluation function is also reformed into：E(L_i)=100*E_Conflict(L_i)+E_Position(L_i)；

5th step, performs selection operation to initial population according to the height of fixed fitness function, obtains To male parent population F_k.The roulette wheel selection with elitism strategy is used, i.e., first by optimal in parent Body is retained in progeny population, and then genetic operator selects other chromosomes.

Using the roulette wheel selection with elitism strategy, i.e., the optimum individual in parent is first retained in offspring In colony, other chromosomes are then selected in accordance with the following steps：

1. the fitness value fit (V of individual are calculated_i) (i=1,2 ..., n), wherein V_iFor a dyeing in potential solution colony Body；

2. the accumulative fitness value Accfit (V of individual are calculated_i) (i=1 2 ..., adds up fitness value n) and relatively RelAccfit(V_i) (i=1,2 ..., n), wherein

RelAccfit(V_i)=Accfit (V_i)/Accfit(V_n)

3. the random number r in one [0,1] is generated, if RelAccfit (V_i-1) ＜ r≤RelAccfit (V_i) (i=1,2 ..., n), that Select individual V_i, it is assumed herein that RelAccfit (V₀)=0；

6th step, by male parent population F_kPopulation C is produced through intersecting with probability 0.5_k。

Using random single-point type interleaved mode, i.e., a gene is randomly selected on two father's chromosome strings Position, gene is exchanged with each other using the point as boundary.Before being intersected, first selected chromosome is pressed Packet pairing is carried out according to certain order.It is random to add or move if the total chromosome number chosen is odd number Except a chromosome.General crossover probability span is 0.25~0.75.Take crossover probability herein P_c=0.5, so that average to there are namely 4 chromosomes of half to be intersected every time.

7th step, to population C_kPerform gland conflict component mutation operation and obtain population pop of new generation_k+1.Adopt With gland conflict position variation mode, that is, the point army of all generation glands and conflict phenomenon is chosen to mark corresponding base Because of position, the new plotting of random generation one is position encoded to replace original gene position, i.e., population of new generation. The thus army's of make use of mark gland and the heuristic information of conflict so that improve bad subsolution, got well The possibility of subsolution is bigger.

Army's mark conflict refers to that there is overlapping part in the shared plotting region of Liang Ge armies mark, including index wire intersects； The detection that system is conflicted using the method for grid map；Region will be marked and drawed first according to screen pixel grid Change, then calculate the grid shared by it using army's cursor position and its size, finally by judge Liang Ge armies mark Whether identical grid is occupied come the target conflict of the army of detection；

8th step, judges whether to meet end condition；

The present invention judges that colony has developed maturation using following two condition：, can be with through repeatedly calculating Stably obtain approximate optimum individual；When continuing to develop, optimum individual is without significantly improving；

If meeting end condition, the point military standard plotting scheme optimized；Terminate this method；

If not meeting end condition, the 5th step is returned, continues computing.

Only when the comparative maturity that population has developed, the optimal of the problem of obtaining just can guarantee that or satisfied Solution.Judge that population has developed maturation using following two condition：, can be stably through repeatedly calculating Obtain approximate optimum individual；When continuing to develop, optimum individual is without significantly improving.

Claims (2)

1. a kind of deployment diagram point army mark automatic avoiding method based on genetic algorithm, comprises the following steps：

The first step, first according to point army of army target longitude and latitude, using the clustering algorithm based on grid to point Army's mark is clustered, the army of distinguishing mark compact district and non-dense set area；

Second step, determines coding framework；Deployment diagram point army mark Avoidance is abstracted as competition for space Combinatorial optimization problem, so that coded system uses integer coding；Plotting scheme deployment diagram is whole with one Number vector represents that vectorial each component represents a Ge Dian armies target and marks and draws position；Using with N number of point The vector of amount represents alternative plotting scheme, and vectorial each component represents a Ge Dian armies target and mark and draw position,

The point army mark alternate location model used in this method has 5 position models and 10 position models, for Point army mark in close quarters, its corresponding component value is the integer value in [0,9], for non-dense set area Point army mark in domain, its corresponding component value is the integer value in [0,4]；

3rd step, determines initial population；N number of vector is randomly generated using a kind of method of completely random to make For initial population pop₀；For the point army mark in close quarters, its corresponding chromogene position takes [0,9] Interior random integer value, is marked for the point army in non-dense region, and its corresponding chromogene position takes [0,4] Interior random integer value；

4th step, determines fitness function；Army's mark conflict is detected using the method for grid map, built Vertical Environmental Evaluation Model is to pop₀Evaluated, obtain pop₀Evaluation of estimate；Define fitness function such as Under：

Fit (JB)=E (JB)

Wherein, E (JB) represents plotting scheme JB quality evaluation value, and fit (JB) represents plotting scheme JB correspondences Chromosome fitness value；

Specifically, this method chooses 2 independent factors：Conflict, positional priority；Define first single Then the two evaluation functions are combined into single army by individual army's target conflict, positional priority evaluation function Target quality evaluation function, finally by single army's target quality evaluation function it is cumulative obtain view picture will figure matter Measure evaluation function；

(4.1) conflict evaluation function；

For each army mark, mutually conflict if marked with other armies, the evaluation function value that conflicted is set It is set to -1, is otherwise set as 0；Single army's target conflict evaluation function is：

Wherein L_iRepresent i-th of army's mark, RectIntersect (L_i, L_j) represent i-th and j army mark and collide with each other, E_Conflict(L_i) represent i-th of army's target conflict evaluation of estimate；

(4.2) positional priority evaluation function；

Make Pos_j(L_i) represent L_iJ-th plotting position, Order (Pos_j(L_i)) represent alternate location sequence number, it is standby Select position number predefined, by positional priority evaluation function be defined as 10 with alternate location sequence number it Difference, i.e.,

E_Position(L_i)=10-Order (Pos_j(L_i))；

(4.3) composite evaluation function；

The quality evaluation fonction composition of single factor is taken to the side of weighted sum into total quality evaluation function Single two kinds of evaluations of estimate of army's target, are first added by being multiplied by weight factor and obtain a total quality by formula again Evaluation of estimate, then single army's target quality evaluation value added up entirely will figure plotting quality evaluation Value, i.e.,：

E(L_i)=W_ConflictE_Conflict(L_i)+W_PositionE_Position(L_i)

$E\left(L\right)=\underset{i=1}{\overset{N}{\Σ}}E\left(L_i\right)$

Wherein W_ConflictRepresent the weight coefficient of the conflict factor, W_PositionRepresent the weight coefficient of the positional priority factor；

5th step, performs selection operation to initial population according to the height of fixed fitness function, obtains To male parent population F_k；

Using the roulette wheel selection with elitism strategy, i.e., the optimum individual in parent is first retained in offspring In colony, other chromosomes are then selected in accordance with the following steps：

1. the fitness value fit (V of individual are calculated_i) (i=1,2 ..., n), wherein V_iFor a dyeing in potential solution colony Body；

2. the accumulative fitness value Accfit (V of individual are calculated_i) (i=1 2 ..., adds up fitness value n) and relatively RelAccfit(V_i) (i=1,2 ..., n), wherein

$Accfit\left(V_i\right)=\underset{k=1}{\overset{i}{\Σ}}fit\left(V_k\right)$

RelAccfit(V_i)=Accfit (V_i)/Accfit(V_n)

3. the random number r in one [0,1] is generated, if RelAccfit (V_i-1) ＜ r≤RelAccfit (V_i) (i=1,2 ..., n), that Select individual V_i, it is assumed herein that RelAccfit (V₀)=0；

6th step, by male parent population F_kWith crossover probability P_cPopulation C is produced through intersecting_k；

Using random single-point type interleaved mode, i.e., a gene is randomly selected on two father's chromosome strings Position, gene is exchanged with each other using the point as boundary；Before being intersected, first selected chromosome is pressed Packet pairing is carried out according to certain order；It is random to add or move if the total chromosome number chosen is odd number Except a chromosome；

7th step, to population C_kPerform gland conflict component mutation operation and obtain population pop of new generation_k+1；Adopt With gland conflict position variation mode, that is, the point army of all generation glands and conflict phenomenon is chosen to mark corresponding base Because of position, the new plotting of random generation one is position encoded to replace original gene position, i.e., population of new generation； The thus army's of make use of mark gland and the heuristic information of conflict so that improve bad subsolution, got well The possibility of subsolution is bigger；

Army's mark conflict refers to that there is overlapping part in the shared plotting region of Liang Ge armies mark, including index wire intersects； The detection that method is conflicted using the method for grid map；Region will be marked and drawed first according to screen pixel grid Change, then calculate the grid shared by it using army's cursor position and its size, finally by judge Liang Ge armies mark Whether identical grid is occupied come the target conflict of the army of detection；

8th step, judges whether to meet end condition；

This method judges that colony has developed maturation using following two condition：, can be with through repeatedly calculating Stably obtain approximate optimum individual；When continuing to develop, optimum individual is without significantly improving；

If meeting end condition, the point military standard plotting scheme optimized；Terminate this method；

If not meeting end condition, the 5th step is returned, continues computing.

2. a kind of automatic avoidance side of deployment diagram point army mark based on genetic algorithm as claimed in claim 1 Method, it is characterised in that：

In the third step, the scale N of population is set as 8；

In the 4th step, the weight factor value of each in evaluation function is as follows：W_Conflict=100W_Position=1；It is single Individual army's mark is marked and drawed quality evaluation function and also reformed into：E(L_i)=100*E_Conflict(L_i)+E_Position(L_i)；

In the 6th step, crossover probability P is taken_c=0.5.