CN107168365B - Unmanned-manned aircraft formation information dynamic distribution processing method - Google Patents

Abstract

The invention relates to an unmanned-man-machine formation information dynamic distribution processing method, which includes calling a pre-planning model for a task pool that does not include mandatory task information; initializing the distribution and transmission attributes of the task information to be distributed, and obtaining the first an initial solution; solve the pre-planning model to obtain a pre-planning scheme for distributing and transmitting the task information to be distributed; if the task information to be distributed in the task pool is distributed and transmitted according to the pre-planning scheme In the process, the task pool receives the mandatory task information, then stops the distribution and transmission of the task information to be distributed, and calls the interruption model; Sexual task information is distributed and transmitted. The present invention can take into account the particularity of mandatory task information in the process of task distribution and arrangement, so that the task information can be reasonably arranged, and an optimal information distribution and transmission sequence can be formed.

Description

Unmanned-manned aircraft formation information dynamic distribution processing method

technical field

The invention relates to the technical field of unmanned aerial vehicle-manned aircraft information processing, in particular to a method for dynamic distribution and processing of unmanned aerial vehicle formation information.

Background technique

In the process of unmanned-man-machine collaborative task execution, there are certain differences in the type of information to be processed, the amount of information required, and the importance of the information itself at different stages. Therefore, in the effective distribution of task information During the transmission process, it is necessary to consider that task information may have different priority levels, such as mandatory task information, important task information, general task information and low priority task information. in:

Mandatory task information refers to the task information that plays a key role in the entire process of cooperative task execution due to its own timing requirements, importance, and needs to be immediately distributed and processed in the unmanned-man-machine system, which is the highest priority. task information.

Important-level task information refers to the task information that has an important impact on the entire collaborative process and the benefits of task completion are significantly higher than that of general-level and low-priority tasks, such as reconnaissance tasks. In modern warfare, battlefield reconnaissance determines the trend of war. Accurate and timely battlefield information can determine the success or failure of wars. Because UAVs perform reconnaissance missions with no risk of casualties, flexible deployment, and timely response, they have attracted the attention of various countries. Missions have become one of the most important mission modes for UAVs at present.

General-level mission information refers to regular mission instructions issued by demand forecasting and command and control centers, such as airborne early warning missions. Deploy the drone in the sky close to the enemy in advance, and then transmit the information obtained by the drone to the manned aircraft parked in the safe area through the communication link, and then the manned aircraft will timely transmit the information to the control center for interception tasks. .

Low-priority tasks refer to the completion time and whether to perform tasks that have little effect on the efficiency of the collaborative process, such as daily cruise tasks.

At present, in the process of unmanned-man-machine formation cooperative task execution, there is no method that takes into account the particularity of mandatory task information and dynamically distributes the task information according to the type of task information to be distributed in the unmanned-man-machine formation system task pool. and transmission to form an optimal information distribution and transmission sequence.

SUMMARY OF THE INVENTION

(1) Technical problems solved

The present invention provides an unmanned-manned aircraft formation information dynamic distribution processing method, which can take into account the particularity of mandatory task information in the process of distributing and transmitting the task information to be distributed in the task pool, so that the tasks in the task pool can be The information is reasonably arranged to form the optimal information distribution and transmission sequence.

(2) Technical solutions

The unmanned-manned aircraft formation information dynamic distribution processing method provided by the present invention includes:

For a task pool that does not include mandatory task information, a pre-established pre-planning model is invoked, and the optimization goal of the pre-planning model is to maximize the total weight of the task information to be distributed in the task pool under the first preset constraint condition value;

Using the first encoding method to initialize the distribution and transfer attributes of the task information to be distributed, to obtain a first initial solution;

Based on the first initial solution, a first genetic algorithm is used to solve the pre-planning model to obtain a pre-planning scheme for distributing and transmitting the task information to be distributed;

If the task pool receives mandatory task information during the process of distributing and transmitting the task information to be distributed in the task pool according to the pre-planning scheme, the distribution and transmission of the to-be-distributed task information will be stopped, and Invoke a pre-built interrupt model;

Use the second encoding method to initialize the distribution and transfer attributes of the mandatory task information to obtain a second initial solution;

setting the current optimization objective of the outage model to maximize the distribution quantity of mandatory task information under a second preset constraint; and based on the second initial solution, using a second genetic algorithm to solve the current outage model , to obtain the first scheme for distributing and transmitting the mandatory task information;

Determine whether the distribution quantity of the mandatory task information in the first solution is equal to the total quantity of the mandatory task information in the task pool;

If so, the current optimization objective of the interruption model is set to minimize the total completion time of distributing the mandatory task information under the second preset constraint condition, and the third genetic algorithm is used to solve the current interruption model, and obtain the correct value. The second scheme for distributing and transmitting the mandatory task information, and distributing and transmitting the mandatory task information in the task pool according to the second scheme.

(3) Beneficial effects

The method for dynamically distributing unmanned-manned aircraft formation information provided by the present invention adopts a pre-planning model for the task pool that does not include mandatory task information, and then uses the pre-planning scheme obtained by solving the pre-planning model to analyze the task information in the task pool. distribution and delivery. However, if the mandatory task information is received in the task pool, the interrupt model is called. The model first takes maximizing the distribution quantity of mandatory task information as the optimization goal. When the goal is achieved, it will minimize the distribution of mandatory task information. The total completion time is the goal, to achieve immediate distribution and transmission on the basis of all mandatory task information distribution and transmission, to ensure the timeliness of task information, and to avoid delays in the coordinated operation of drones and manned aircraft. It can be seen that the present invention can take into account the particularity of mandatory task information in the process of distributing and arranging the task information to be distributed in the task pool, so that the task information in the task pool can be reasonably arranged to form optimal information distribution and transmission. sequence.

Description of drawings

In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only These are some embodiments of the present disclosure. For those of ordinary skill in the art, other drawings can also be obtained from these drawings without creative efforts.

FIG. 1 shows a partial flowchart of a mandatory task-oriented information distribution processing method in an embodiment of the present invention;

FIG. 2 shows a schematic diagram of a chromosome composed of 5 genes in an embodiment of the present invention.

Detailed ways

The technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, but not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present disclosure.

In a first aspect, the present invention provides an unmanned-manned aircraft formation information dynamic distribution processing method, the method comprising:

A1. For a task pool that does not include mandatory task information, a pre-established pre-planning model is invoked, and the optimization objective of the pre-planning model is to maximize the amount of task information to be distributed in the task pool under the first preset constraint condition. total weight value;

It is understandable that the so-called mandatory task information is the task information that plays a key role in the entire collaborative process and needs to be executed immediately. For example: attack missions, reconnaissance-attack integration missions, and firepower assessment missions. When performing attack missions and reconnaissance-attack integration tasks, the UAV will immediately transmit the information of the target of the attack and the captured images to the manned aircraft, and then the manned aircraft will issue an order to direct the UAV to carry out the attack. Fire assessment tasks are divided into fire guidance and fire assessment and fire strike assessment. Fire guidance and calibration is the use of drones to enter the fire strike target area, shoot image information of the corresponding area, and transmit it to the commander to help observe the impact point, correct the shooting deviation, improve the fire strike accuracy, and reduce ammunition consumption. Fire strike assessment means that after the early strike is over, the drone enters the fire strike target area to help observe the fire strike effect and provide an important basis for the next action.

A2, using the first encoding method to initialize the distribution and transfer attributes of the task information to be distributed to obtain a first initial solution;

A3. Based on the first initial solution, use the first genetic algorithm to solve the pre-planning model to obtain a pre-planning scheme for distributing and transmitting the task information to be distributed;

A4. If the task pool receives mandatory task information during the process of distributing and transmitting the task information to be distributed in the task pool according to the pre-planning scheme, the distribution and transmission of the to-be-distributed task information will be stopped. , and call the pre-established interrupt model;

A5, using the second encoding method to initialize the distribution and transmission attributes of the mandatory task information to obtain a second initial solution;

A6. Set the current optimization objective of the outage model to maximize the distribution quantity of mandatory task information under the second preset constraint condition; and based on the second initial solution, adopt the second genetic algorithm to analyze the current outage model Solve to obtain the first scheme for distributing and transmitting the mandatory task information;

It can be understood that the goal of the current outage model is to maximize the distribution of mandatory task information under preset constraints.

A7. Determine whether the distribution quantity of the mandatory task information in the first solution is equal to the total quantity of the mandatory task information in the task pool;

A8, if yes, then set the current optimization objective of the interruption model to minimize the total completion time of distributing the mandatory task information under the second preset constraint condition, and adopt the third genetic algorithm to solve the current interruption model, A second scheme for distributing and transmitting the mandatory task information is obtained, and the mandatory task information in the task pool is distributed and transmitted according to the second scheme.

It is understandable that, if the distribution quantity of mandatory task information in the first solution is equal to the total quantity of mandatory task information in the task pool, it means that the goal of maximizing the distribution quantity of mandatory task information has been achieved, and at this time, it will be interrupted. The optimization goal of the model is set to minimize the total completion time of distributing mandatory task information, that is, on the basis of maximizing the distribution quantity of mandatory task information, strive to minimize the total completion time of distributing mandatory task information.

The method for dynamically distributing unmanned-manned aircraft formation information provided by the present invention adopts a pre-planning model for the task pool that does not include mandatory task information, and then uses the pre-planning scheme obtained by solving the pre-planning model to analyze the task information in the task pool. distribution and delivery. However, if the mandatory task information is received in the task pool, the interrupt model is called. The model first takes maximizing the distribution quantity of mandatory task information as the optimization goal. When the goal is achieved, it will minimize the distribution of mandatory task information. The total completion time is the goal, to achieve immediate distribution and transmission on the basis of all mandatory task information distribution and transmission, to ensure the timeliness of task information, and to avoid delays in the coordinated operation of drones and manned aircraft. It can be seen that the present invention can take into account the particularity of mandatory task information in the process of distributing and arranging the task information to be distributed in the task pool, so that the task information in the task pool can be reasonably arranged to form optimal information distribution and transmission. sequence.

For the sake of clarity, the formula parameters involved in each formula are described below:

In this paper, a directed graph G(V, E, W) is used to represent all available communication network topologies between UAVs/man-machines, and UAVs/man-machines are described as nodes in the communication network topology. The specific model parameters are as follows :

V={1,2,...,m} represents the set of nodes in the communication network topology, and m represents the total number of nodes in the communication network topology.

E={<i,j>|i,j∈V,i≠j} represents the set of directed edges, where <i,j> represents the directed edge from node i to node j in the communication network topology;

W={w _ij |i,j∈V} represents the weight set of each directed edge in the graph, where w _i j represents the Euclidean distance between node i and node j.

B _v represents the maximum amount of data that node v can provide, where v represents any node in the communication network topology, v∈V;

T represents the set of information to be distributed, n represents the number of elements in the set, t represents any information to be distributed, t∈T; where T ^a represents mandatory task information, T ^b represents important level information, and T ^c represents general level information , T ^d represents low priority information;

[et ,l _t ] indicates that the information to be distributed _{t needs to arrive at the information sink within this time window, e t indicates the earliest arrival time, and l t indicates the} latest arrival time;

ST _t represents the time when the information t to be distributed actually starts to be distributed from the information source, and ET _t represents the time when the information t to be distributed actually arrives at the information sink;

SN _t represents the actual information source of the information t to be distributed, and EN _t represents the information sink that needs to receive the information t to be distributed;

表示待分发信息t从节点i传递到节点j发生的传输时延；

表示待分发信息t从节点i传递到节点j发生的传播时延；

Represents the transmission delay that the information t to be distributed is transmitted from node i to node j;

Represents the propagation delay that the information to be distributed t is transmitted from node i to node j;

D represents the maximum acceptable delay in the communication network topology;

TW _t represents the bandwidth required for the information t to be distributed;

NW _ij represents the maximum bandwidth that the directed edge <i,j> in the communication network topology can bear;

P _t represents the priority of the information t to be distributed, P _t =1 represents a low priority task, P _t =2 represents a general level task, P _t =3 represents an important level task, and P _t =4 represents an interrupt level task;

H _t represents the income that can be obtained after completing the task of distributing information t;

G _t represents the weight value of the information t to be distributed;

C _t represents the possible disturbance cost of the information t to be distributed;

取1或0，其中，取1表示待分发信息t从节点i发送到节点j，取0表示待分发信息t不从节点i发送到节点j。Decision variables

Take 1 or 0, where 1 indicates that the information t to be distributed is sent from node i to node j, and taking 0 indicates that the information t to be distributed is not sent from node i to node j.

In specific implementation, the objective function of the above pre-planning model can be:

In the above formula, F1 is the total weight value of the task information to be distributed.

In specific implementation, the objective function of the interruption model to maximize the distribution quantity of mandatory task information under preset constraints as the current optimization goal may be:

In the above formula, F2 is the distribution quantity of mandatory task information.

In a specific implementation, the objective function of the interruption model that minimizes the total completion time of distributing mandatory task information under the preset constraints as the current optimization goal may be:

In the formula, F3 is the total completion time of distributing mandatory task information.

At different stages, different objective functions are used to achieve different optimization goals.

During specific implementation, the above-mentioned first or second preset constraints can be set as required, such as time window constraints, delay constraints, bandwidth constraints, information source constraints, access uniqueness constraints, etc., wherein the so-called time window constraints are mandatory The mandatory task information needs to be distributed and delivered within a preset time window, the delay constraint is that the transmission delay and propagation delay of the mandatory task information do not exceed the maximum delay of the communication network topology, and the bandwidth constraint is that in the communication link At the same time, the sum of the amount of mandatory task information data that can be transmitted does not exceed the maximum bandwidth that the communication network topology can bear. There is only one information source for each mandatory task information, only one information sink for each mandatory task information, and the number of times that any node forwards the same mandatory task information is less than or equal to 1.

Because different models target different types of task information, the applicable scope of preset conditions is slightly different. For example, the first preset constraint condition can be expressed by the following formula:

ET _t ≤l _t ,t∈T

ET _t ≥e _t ,t∈T

ET _t -ST _t ≤D,t∈T

It is understandable that T in the above constraints does not include mandatory task information.

For another example, the second preset constraint condition can be expressed by the following formula:

ET _t ≤l _t ,t∈T ^a

ET _t ≥e _t ,t∈T ^a

ET _t -ST _t ≤D,t∈T ^a

During specific implementation, the specific process of obtaining the first initial solution by initializing the distribution and transfer attributes of the task information to be distributed by using the first encoding method may include:

Using the first encoding method to decode the pre-planning model into chromosomes, the chromosomes include genes that correspond one-to-one with the task information to be distributed in the task pool;

It is understandable that the number of task information to be distributed is the same as the number of genes on the chromosome, and one gene corresponds to one piece of task information to be distributed.

For example, the number n of the information to be distributed is the number of genes in the chromosome, the genes are encoded in a tuple, and m is the total number of nodes in the communication network topology. The basic encoding method is as follows:

Gene=(Flag,Node ₁ ,Node ₂ ,...,Node _m ,Time ₁ ,Time ₂ ,...,Time _m ,Weight)

Among them, Flag indicates whether the information to be distributed can be distributed, Node ₁ , Node ₂ , . . . Node _m indicates the nodes through which the information to be distributed is forwarded, Node ₁ indicates the information source of the information to be distributed, and Node _m indicates the source of the information to be distributed. Information sink, Time ₁ , Time ₂ ,..., Time _m represents the forwarding time of the information to be distributed at the corresponding node, Time ₁ represents the time when the information to be distributed starts to be distributed from the information source, Time _m represents the time when the information to be distributed reaches the information sink, Weight is the weight value of the task information to be distributed.

The first identifier of each gene on the chromosome is set to 1, and the first identifier set to 1 indicates that the task information to be distributed corresponding to the gene can be distributed and transmitted;

It is understandable that the first identifier here is the above-mentioned Flag, and the mandatory task information corresponding to the identifier can be assigned and transmitted by setting the first identifier to 1.

Obtain the sink node and weight value of each task information to be distributed, and randomly generate a source node different from the sink node for each task information to be distributed;

Understandably, since the sink node is different from the source node, Node ₁ ≠Node _m .

Determine whether each task information to be distributed needs to be forwarded; for the task information to be distributed that needs to be forwarded, multiple different forwarding nodes are randomly generated to form a forwarding path; for the task information to be distributed that does not need to be forwarded, the forwarding node is set to -1 ;

It is understandable that, for the task information to be distributed that does not need to be forwarded, let Node ₂ =Node ₃ =...=Node _m-1 =-1.

It is understandable that, for the task information to be distributed that needs to be forwarded, the number of random forwarding is c<=m-2, and the numbers of c forwarding nodes generated randomly are recorded in Node ₂ ..., Node _m-1 , and it is guaranteed that Node1≠Node ₂ ≠…≠Node _m .

Read the time window of each task information to be distributed; for each task information to be distributed, randomly generate a time point in the time window, and use this time point as the time when the task information to be distributed reaches the destination node; For the task information to be distributed that needs to be forwarded, calculate the time when the task information to be distributed reaches each forwarding node and the time it is sent from the source node according to the forwarding path; The time when the node sends out, and the forwarding time of each forwarding node is set to -1;

The first identification of each task information to be distributed, the source node, the forwarding node, the sink node, the time of sending from the source node, the time of arriving at each forwarding node, the time of arriving at the sink node, and the weight value are used as the task to be distributed. The distribution and transmission attributes of information, and the distribution and transmission attributes of each task information to be distributed form the first initial solution.

Similar to the above method, the specific process of obtaining the second initial solution by initializing the distribution and transfer attributes of the mandatory task information using the second encoding method may include:

Using the second encoding method to decode the interruption model into chromosomes, the chromosomes include genes that correspond one-to-one with the mandatory task information to be distributed in the task pool;

Understandably, the number of mandatory task information is the same as the number of genes on the chromosome, and one gene corresponds to one mandatory task information.

For example, the number n of the information to be distributed is the number of genes in the chromosome, the genes are encoded in a tuple, and m is the total number of nodes in the communication network topology. The basic encoding method is as follows:

Gene=(Flag,Node ₁ ,Node ₂ ,...,Node _m ,Time ₁ ,Time ₂ ,...,Time _m )

Among them, Flag indicates whether the information to be distributed can be distributed, Node ₁ , Node ₂ , . . . Node _m indicates the nodes through which the information to be distributed is forwarded, Node ₁ indicates the information source of the information to be distributed, and Node _m indicates the source of the information to be distributed. Information sink, Time ₁ , Time ₂ ,..., Time _m represents the forwarding time of the information to be distributed at the corresponding node, Time ₁ represents the time when the information to be distributed starts to be distributed from the information source, and Time _m represents the time when the information to be distributed arrives at the information sink.

The first identification of each gene on the chromosome is set to 1, and the first identification set to 1 indicates that the mandatory task information corresponding to the gene can be distributed and transmitted;

It is understandable that the first identifier here is the above-mentioned Flag, and the mandatory task information corresponding to the identifier can be assigned and transmitted by setting the first identifier to 1.

Obtain the sink node of each mandatory task information, and randomly generate a source node different from its sink node for each mandatory task information;

Understandably, since the sink node is different from the source node, Node ₁ ≠Node _m .

Determine whether each mandatory task information needs to be forwarded; for mandatory task information that needs to be forwarded, multiple different forwarding nodes are randomly generated to form a forwarding path; for mandatory task information that does not need to be forwarded, its forwarding node is set to -1 ;

It is understandable that, for the mandatory task information that does not need to be forwarded, let Node ₂ =Node ₃ =...=Node _m-1 =-1.

It is understandable that, for the mandatory task information that needs to be forwarded, the number of random forwarding is c<=m-2, and the numbers of c forwarding nodes generated randomly are recorded to Node ₂ ..., Node _m-1 , and it is guaranteed that Node1≠Node ₂ ≠…≠Node _m .

Read the time window of each mandatory task information; for each mandatory task information, randomly generate a time point in the time window, and use this time point as the time when the mandatory task information reaches the destination node; For the mandatory task information that needs to be forwarded, the time when the mandatory task information arrives at each forwarding node and the time when it is sent from the source node is calculated according to the forwarding path; for the mandatory task information that does not need to be forwarded, the mandatory task information is calculated from the source The time when the node sends out, and the forwarding time of each forwarding node is set to -1;

The first identification of each mandatory task information, the source node, the forwarding node, the sink node, the time of sending from the source node, the time of reaching each forwarding node, and the time of arriving at the sink node are used as the distribution of the mandatory task information With transfer properties, the distribution and transfer properties of each mandatory task information form a second initial solution.

For example, as shown in Figure 2, a chromosome is formed by 5 genes, taking the first gene as an example, (1,1,-1,2,9.5,-1,12.5) represents the first information to be distributed The information is sent from the information source numbered 1 to the information sink numbered 2 without forwarding. The sending time is 9.5 seconds and the arrival time is 10.5 seconds.

During specific implementation, the process of using the first genetic algorithm to solve the pre-planning model may include:

S1. Set the initial value of the iteration number k to 1;

S2, taking the objective function of the pre-planning model as a fitness function, and calculating the fitness function value of the chromosome in the initial population;

S3. The roulette selection method is used to select the preset number of chromosomes with the highest fitness function value from the parent population and inherit them into the offspring population;

Understandably, the basic idea of the so-called roulette selection method is that the probability of each chromosome being selected is proportional to the value of its fitness function. Calculate the fitness function value of the chromosome according to the fitness function, and calculate the proportion of the total fitness of the chromosome individuals in the population relativefitness=fitness/sum(fitness), which is the probability of being selected and inherited to the next generation, the ratio The larger it is, the greater the probability of being selected for inheritance to the next generation.

S4. Perform a pairwise single-point crossover operation on the chromosomes in the population;

It is understandable that the single-point crossover method is adopted, that is, a crossover point is randomly generated, and the parts of two adjacent chromosomes located behind this point in the population are exchanged with each other in turn to generate two new chromosomes.

S5. Perform reset mutation processing on the chromosome obtained by the crossover operation;

S6. Perform a first update operation on the chromosomes obtained by the reset mutation process, specifically combining the first preset number of chromosomes with the lowest fitness in the progeny population and the second preset number of chromosomes with the lowest fitness in the progeny population , forming a new population;

For example, the mutated progeny population is arranged in ascending order of fitness value, the first SonNum chromosomes are taken out, the parent population is arranged in descending order of fitness value, and the FatherNum chromosomes are taken out to form a new population.

S7, determine whether the current number of iterations reaches the preset maximum number of iterations km _max ;

If so, take the solution corresponding to the new population obtained in the last iteration process as the pre-planning scheme;

Otherwise, take the new population as the initial population, increase the number of iterations by 1, and return to S2.

Here, by performing operations on chromosomes such as selection, crossover, and mutation, the obtained chromosomes are used as a pre-planning scheme.

In the specific implementation, before the first update operation, it is also possible to check whether the distribution and transmission attributes corresponding to the genes on the chromosome after the reset mutation process satisfy the preset constraint conditions;

If so, execute the first update operation;

Otherwise, the first update operation is performed after adjusting the fitness function value of the chromosome after the reset mutation process.

Considering that the information to be distributed needs to satisfy constraints such as bandwidth, delay, time window, and information source of the communication network topology, a constraint check is also performed on chromosomes here. For chromosomes that fail to pass the constraint check, add or subtract a penalty factor on their fitness function value as needed to make their fitness function value smaller or larger, and in the selection operation to remove those that do not meet the given constraints chromosome.

During specific implementation, the second genetic algorithm may also be used to solve the current interruption model, and to determine whether the distribution quantity of mandatory task information in the first solution is equal to the total quantity of mandatory task information in the task pool. and the operation of judging whether the number of cycles reaches the preset number of times, until the number of mandatory task information distributed in the first solution after the solving operation is equal to the total number of mandatory task information in the task pool or the number of cycles reaches the preset number of times ;

If the number of cycles reaches the preset number of times, the mandatory task information in the task pool is distributed and transmitted according to the first solution after the last solving operation.

Here, the chromosome with the highest fitness is finally searched as the optimal solution by means of loop or iteration.

During specific implementation, the process of using the second genetic algorithm to solve the current interruption model may include:

Taking the objective function of the interruption model that maximizes the distribution quantity of mandatory task information as the current optimization objective as the current fitness function, calculate the fitness function value of the chromosomes in the population;

It can be understood that the fitness function fitness=Z*Count+(1-Z)*Time, where Z is a variable between 0 and 1. When Z=1, the interruption model takes maximizing the distribution quantity of mandatory task information as the current optimization goal; when Z=0, the interruption model takes minimizing the total completion time of distributing mandatory task information as the optimization goal. In S3, Z=1. Count is the total number of mandatory task information in the task pool, and Time is the sum of the total time to complete the distribution and transfer the mandatory task information.

The roulette selection method is used to select the preset number of chromosomes with the highest fitness function value from the parent population and inherit them into the offspring population;

Understandably, the basic idea of the so-called roulette selection method is that the probability of each chromosome being selected is proportional to the value of its fitness function. Calculate the fitness function value of the chromosome according to the fitness function, and calculate the proportion of the total fitness of the chromosome individuals in the population relativefitness=fitness/sum(fitness), which is the probability of being selected and inherited to the next generation, the ratio The larger the probability, the greater the probability of being selected for inheritance to the next generation.

Perform pairwise single-point crossover operations on chromosomes in the population;

It is understandable that the single-point crossover method is adopted, that is, a crossover point is randomly generated, and the parts of two adjacent chromosomes located behind this point in the population are exchanged with each other in turn to generate two new chromosomes.

Perform reset mutation processing on chromosomes obtained by crossover operation;

Perform a first update operation on the chromosomes obtained by the reset mutation process, specifically combining the first preset number of chromosomes with the lowest fitness in the progeny population and the second preset number of chromosomes with the lowest fitness in the progeny population to form new species;

For example, the mutated progeny population is arranged in ascending order of fitness value, the first SonNum chromosomes are taken out, the parent population is arranged in descending order of fitness value, and the FatherNum chromosomes are taken out to form a new population.

The solution corresponding to the new population is used as the first solution.

In specific implementation, the process of resetting mutation processing on the chromosome obtained by the crossover operation may include: generating a random number between 0 and 1, if the random number is less than a preset mutation probability, according to the A chromosome is generated according to the method for generating the initial solution; a chromosome is randomly selected in the progeny population, and the randomly selected chromosome is replaced by the chromosome generated according to the initial solution, and the other chromosomes remain unchanged.

In a specific implementation, the use of the third genetic algorithm to solve the current interruption model may include multiple iterative processes until the number of iterations reaches a preset number of times, and the solution corresponding to the final population is used as the second solution; wherein, Each iterative process includes:

Taking the objective function of the interruption model that minimizes the total completion time of distributing mandatory task information as the current optimization objective as the current fitness function, calculates the fitness function value of the chromosomes in the population;

The roulette selection method is used to select the preset number of chromosomes with the highest fitness function value from the parent population and inherit them into the offspring population;

Perform pairwise single-point crossover operations on chromosomes in the population;

Delete the forwarding node mutation processing on the chromosome obtained by the crossover operation;

Performing a second update operation on the chromosomes obtained by deleting the mutation processing of the forwarding node, specifically combining the first preset number of chromosomes with the highest fitness in the progeny population and the second preset number of chromosomes with the highest fitness in the progeny population, form new species.

In the above process, the deletion and forwarding node mutation processing on the chromosome obtained by the crossover operation may include: generating a random number between 0 and 1, and if the random number is less than a preset mutation probability, randomly select A gene in the chromosome, and the forwarding node of the mandatory task information corresponding to the randomly selected gene is set to -1, and the moment when the mandatory task information reaches each forwarding node is set to -1.

During specific implementation, before the second update operation is performed on the chromosome obtained by deleting the mutation processing of the forwarding node, the method may further include:

Whether the distribution and transmission attributes corresponding to the genes on the chromosomes after the mutation processing of the deleted forwarding node satisfies the preset constraints;

If so, execute the second update operation;

Otherwise, the second update operation is performed after adjusting the fitness function value of the chromosome after the mutation processing of the deleted forwarding node.

Here, judging whether the distribution and transmission attributes corresponding to the genes on the chromosome after the mutation processing of the deletion forwarding node satisfies the preset constraint conditions is actually a constraint check to ensure that the scheme corresponding to the chromosome satisfies the preset constraint conditions.

To sum up, the method for dynamically distributing and processing unmanned-manned aircraft formation information provided by the present invention adopts a pre-planning model for a task pool that does not include mandatory task information, and then uses the pre-planning scheme obtained by solving the pre-planning model to analyze the tasks. The task information in the pool is distributed and delivered. However, if the mandatory task information is received in the task pool, the interrupt model is called. The model first takes maximizing the distribution quantity of mandatory task information as the optimization goal. When the goal is achieved, it will minimize the distribution of mandatory task information. The total completion time is the goal, to achieve immediate distribution and transmission on the basis of all mandatory task information distribution and transmission, to ensure the timeliness of task information, and to avoid delays in the coordinated operation of drones and manned aircraft. It can be seen that the present invention can take into account the particularity of mandatory task information in the process of distributing and arranging the task information to be distributed in the task pool, so that the task information in the task pool can be reasonably arranged to form optimal information distribution and transmission. sequence.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, but not to limit them; although the embodiments of the present invention have been described in detail with reference to the foregoing embodiments, ordinary The skilled person should understand that it is still possible to modify the technical solutions described in the foregoing embodiments, or to perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the present invention. The scope of the technical solutions of the embodiments of each embodiment.

Claims (9)

1. An unmanned-manned-organized information dynamic distribution processing method is characterized by comprising the following steps:

calling a pre-established pre-planning model aiming at a task pool not including mandatory task information, wherein the optimization goal of the pre-planning model is to maximize the total weight value of the task information to be distributed in the task pool under a first preset constraint condition;

initializing the distribution and transmission attributes of the task information to be distributed by adopting a first coding method to obtain a first initial solution;

on the basis of the first initial solution, solving the pre-planning model by adopting a first genetic algorithm to obtain a pre-planning scheme for distributing and transmitting the task information to be distributed;

if the task pool receives mandatory task information in the process of distributing and transmitting the to-be-distributed task information in the task pool according to the pre-planning scheme, stopping the distribution and transmission of the to-be-distributed task information and calling a pre-established interrupt model;

initializing the distribution and transmission attributes of the mandatory task information by adopting a second coding method to obtain a second initial solution;

setting the current optimization target of the interrupt model to maximize the distribution quantity of mandatory task information under a second preset constraint condition; on the basis of the second initial solution, a second genetic algorithm is adopted to solve the current interrupt model, and a first scheme for mandatory task information distribution and transmission is obtained;

judging whether the distribution quantity of the mandatory task information in the first scheme is equal to the total quantity of the mandatory task information in the task pool or not;

if so, setting the current optimization target of the interrupt model to minimize the total completion time of distributing the mandatory task information under the second preset constraint condition, solving the current interrupt model by adopting a third genetic algorithm to obtain a second scheme for distributing and transmitting the mandatory task information, and distributing and transmitting the mandatory task information in the task pool according to the second scheme;

the initializing the distribution and transmission attributes of the task information to be distributed by adopting the first coding method to obtain a first initial solution comprises the following steps:

decoding the pre-planning model into a chromosome by adopting a first coding method, wherein the chromosome comprises genes which are in one-to-one correspondence with task information to be distributed in a task pool;

setting the first identifier of each gene on the chromosome as 1, wherein the set first identifier represents that the task information to be distributed corresponding to the gene can be distributed and transmitted;

acquiring a host node and a weight value of each piece of task information to be distributed, and randomly generating a source node different from the host node for each piece of task information to be distributed;

judging whether each task information to be distributed needs to be forwarded or not; for the task information to be distributed which needs to be forwarded, randomly generating a plurality of different forwarding nodes to form a forwarding path; setting a forwarding node of task information to be distributed, which does not need to be forwarded, as-1;

reading the time window of each task information to be distributed; for each task information to be distributed, randomly generating a time point in the time window, and taking the time point as the time when the task information to be distributed reaches the host node; for the task information to be distributed which needs to be forwarded, the time when the task information to be distributed reaches each forwarding node and the time when the task information to be distributed is sent from the source node are calculated according to the forwarding path; for the task information to be distributed which does not need to be forwarded, calculating the time when the task information to be distributed is sent from the source node, and setting the forwarding time of each forwarding node as-1;

taking the first identification, the source node, the forwarding node, the destination node, the time sent from the source node, the time of reaching each forwarding node, the time of reaching the destination node and the weight value of each piece of task information to be distributed as the distribution and transmission attribute of the task information to be distributed, and forming a first initial solution for the distribution and transmission attribute of each piece of task information to be distributed;

and/or initializing the distribution and transmission attribute of the mandatory task information by adopting a second coding method to obtain a second initial solution, wherein the method comprises the following steps:

decoding the interrupt model into a chromosome by adopting a second coding method, wherein the chromosome comprises genes which are in one-to-one correspondence with mandatory task information to be distributed in a task pool;

setting the first identifier of each gene on the chromosome as 1, wherein the first identifier set as 1 represents mandatory task information corresponding to the gene to be distributable and transferable;

acquiring a host node of each mandatory task information, and randomly generating a source node different from the host node of each mandatory task information;

judging whether each mandatory task information needs to be forwarded or not; for mandatory task information needing to be forwarded, randomly generating a plurality of different forwarding nodes to form a forwarding path; setting a forwarding node of mandatory task information which does not need to be forwarded as-1;

reading the time window of each mandatory task information; for each mandatory task information, randomly generating a time point in the time window, and taking the time point as the time when the mandatory task information reaches the host node; for the mandatory task information needing to be forwarded, the time when the mandatory task information reaches each forwarding node and the time when the mandatory task information is sent from the source node are calculated according to the forwarding path; for the mandatory task information which does not need to be forwarded, the time when the mandatory task information is sent from the source node is calculated, and the forwarding time of each forwarding node is set to be-1;

and taking the first identifier of each piece of mandatory task information, the source node, the forwarding node, the destination node, the time sent from the source node, the time arriving at each forwarding node and the time arriving at the destination node as the distribution and transmission attributes of the mandatory task information, wherein the distribution and transmission attributes of each piece of mandatory task information form a second initial solution.

2. The method of claim 1, wherein solving the pre-planned model using a first genetic algorithm comprises:

s1, setting the initial value of the iteration times k as 1;

s2, taking the target function of the pre-planning model as a fitness function, and calculating a fitness function value of chromosomes in the initial population;

s3, selecting a preset number of chromosomes with the highest fitness function value from the parent population to be inherited into the offspring population by adopting a roulette selection method;

s4, performing pairwise single-point crossing operation on chromosomes in the population;

s5, carrying out replacement mutation treatment on the chromosomes obtained by the crossover operation;

s6, performing a first updating operation on the chromosomes obtained through the repeated mutation processing, specifically combining the chromosomes with the lowest fitness in the filial generation population by a first preset number and the chromosomes with the lowest fitness in the filial generation population by a second preset number to form a new population;

s7, judging whether the current iteration number reaches the preset maximum iteration number k_max；

If so, taking a solution corresponding to the new population obtained in the last iteration process as a pre-planning scheme;

otherwise, taking the new population as the initial population, adding 1 to the iteration number, and returning to S2.

3. The method of claim 1,

circularly executing the operation of solving the current interrupt model by adopting a second genetic algorithm, judging whether the distribution quantity of the mandatory task information in the first scheme is equal to the total quantity of the mandatory task information in the task pool and judging whether the circulation frequency reaches the preset frequency or not until the distribution quantity of the mandatory task information in the first scheme after the solving operation is equal to the total quantity of the mandatory task information in the task pool or the circulation frequency reaches the preset frequency;

and if the cycle times reach the preset times, distributing and transmitting the mandatory task information in the task pool according to the first scheme after the last solving operation.

4. The method of claim 1, wherein solving the current interrupt model using the second genetic algorithm comprises:

taking an objective function of an interrupt model taking the distribution quantity of the maximum mandatory task information as a current optimization objective as a current fitness function, and calculating a fitness function value of chromosomes in a population;

selecting a preset number of chromosomes with the highest fitness function value from a parent population to be inherited into a child population by adopting a roulette selection method;

performing pairwise single-point crossing operation on chromosomes in the population;

carrying out replacement mutation treatment on the chromosomes obtained by the cross operation;

performing a first updating operation on the chromosomes obtained through the repeated mutation treatment, specifically combining a first preset number of chromosomes with the lowest fitness in the offspring population and a second preset number of chromosomes with the lowest fitness in the offspring population to form a new population;

and taking the solution corresponding to the new population as the first scheme.

5. The method of claim 4, wherein the performing the re-mutation on the chromosomes obtained by the crossover operation comprises: generating a random number between 0 and 1, and if the random number is smaller than a preset variation probability, generating a chromosome according to the generation method of the initial solution; randomly selecting one chromosome from the population of offspring and replacing the randomly selected chromosome with the generated chromosome from the initial solution, the other chromosomes remaining unchanged.

6. The method of claim 4,

solving the current interrupt model by adopting a third genetic algorithm, wherein the solution comprises a plurality of iteration processes until the iteration times reach preset times, and taking a final solution corresponding to the population as the second scheme; wherein each iterative process comprises:

taking an objective function of an interrupt model taking the total completion time of the minimum distribution mandatory task information as a current optimization objective as a current fitness function, and calculating a fitness function value of chromosomes in a population;

selecting a preset number of chromosomes with the highest fitness function value from a parent population to be inherited into a child population by adopting a roulette selection method;

performing pairwise single-point crossing operation on chromosomes in the population;

carrying out deletion forwarding node mutation processing on chromosomes obtained through the cross operation;

and performing second updating operation on the chromosomes obtained by the mutation processing of the deleted forwarding nodes, specifically combining the chromosomes with the highest fitness in the offspring population by the first preset number and the chromosomes with the highest fitness in the offspring population by the second preset number to form a new population.

7. The method according to claim 6, wherein the performing deletion forwarding node mutation processing on the chromosome obtained by the crossover operation comprises: and generating a random number between 0 and 1, if the random number is smaller than the preset variation probability, randomly selecting a gene in the chromosome, setting a forwarding node of the mandatory task information corresponding to the randomly selected gene as-1, and setting the time when the mandatory task information reaches each forwarding node as-1.

8. The method according to any one of claims 1 to 7,

the objective function of the pre-planning model is:

and/or the interrupt model takes the maximum distribution quantity of mandatory task information under the preset constraint condition as an objective function of the current optimization goal as follows:

and/or the interrupt model takes the total completion time for minimizing the distribution of mandatory task information under the preset constraint condition as an objective function of the current optimization goal as follows:

in the formula, F1 is the sum of the weight values of all the task information to be distributed; t represents any task information to be distributed; t is^bRepresenting a three-level task information set; t is^cRepresenting a secondary task information set; t is^dRepresenting a first-level task information set, and sequentially reducing the importance degree of third-level task information, second-level task information and first-level task information; decision variables

1 or 0 is selected, wherein 1 is selected to indicate that the task information t to be distributed is sent from the node i to the node j, and 0 is selected to indicate that the task information t to be distributed is not sent from the node i to the node j; v ═ {1,2, …, m } represents a set of nodes in the communication network topology, m represents a total number of nodes in the communication network topology; g_tA weight value representing the task information t to be distributed; f2 is the distribution quantity of mandatory task information; t is^aExpressing mandatory task information which is four-level task information, wherein the importance degree of the mandatory task information is higher than that of the three-level task information; f3 is the total completion time of mandatory task information to be distributed;

indicating the transfer of task information t to be distributed from node i to nodeThe transmission delay occurring at point j;

representing the propagation delay of the task information t to be distributed from the node i to the node j; n is a radical of^aThe amount of mandatory task information in the task pool.

9. The method according to any one of claims 1 to 7,

the first preset constraint condition comprises:

ET_t≤l_t,t∈T

ET_t≥e_t,t∈T

ET_t-ST_t≤D,t∈T

and/or the second preset constraint condition comprises:

ET_t≤l_t,t∈T^a

ET_t≥e_t,t∈T^a

ET_t-ST_t≤D,t∈T^a

in the formula, ET_tRepresenting the time when the task information t to be distributed actually reaches the information sink; ST (ST)_tRepresenting the moment when the task information t to be distributed actually starts to be distributed from the information source; v ═ {1,2, …, m } represents a set of nodes in the communication network topology, m represents a total number of nodes in the communication network topology;

representing transmission delay of the task information t to be distributed from the node i to the node j;

representing the propagation delay of the task information t to be distributed from the node i to the node j; t is a set of task information to be distributed, and the set does not comprise four levels of task information; l_tInformation for indicating that task information t to be distributed arrives at latestThe time of rest; e.g. of the type_tRepresenting the arrival time of the earliest information sink of the task information t to be distributed; d represents the maximum time delay acceptable in the communication network topology; TW (time-lapse launching) device_tRepresenting the bandwidth required by the task information t to be distributed; NW_ijRepresenting directed edges in a communication network topology<i,j>The maximum bandwidth that can be tolerated; b is_vRepresenting the maximum amount of data that can be provided by node V, V representing any node in the communication network topology, V ∈ V, decision variables

Taking 1 or 0, wherein taking 1 indicates that the task information t to be distributed is sent from the node i to the node j, and taking 0 indicates that the task information t to be distributed is not sent from the node i to the node j; t is^aIndicating mandatory task information.

Images