CN112819242A - Civil transportation type airplane flight test task allocation optimization method - Google Patents

Abstract

The invention provides a civil transportation aircraft flight test task allocation optimization method, which comprises the steps of performing mutation operation and cross operation on chromosomes through parameter initialization, coding and population initialization, judging whether the current chromosomes accord with logic constraints or not, generating trial flight subject arrangement results on the chromosomes which accord with the logic constraints, traversing all chromosomes in a population through objective function value calculation, and performing selection operation in a binary tournament selection mode to obtain a new population until an optimal subject arrangement result is selected. The method can automatically obtain the flight test arrangement result meeting the test flight constraint condition according to the flight test task information of the civil transport aircraft, and optimizes the subject sequence by adopting a genetic algorithm so as to achieve the effect of shortening the test flight period.

Description

Civil transportation type airplane flight test task allocation optimization method

Technical Field

The invention relates to the field of flight tests of civil transport airplanes, in particular to a task allocation method.

Background

The civil transportation type airplane needs to carry out flight tests according to airworthiness requirements specified by airworthiness authorities so as to verify whether the design, manufacture and use of the civil transportation type airplane meet the flight safety standards. The flight test needs to consider the test requirements, the flight test conditions, the flight test safety and other preposed factors to determine the flight test subject sequence, so that the flight test task is distributed. The optimization effect of the flight test task allocation directly influences the period of the flight test, so that whether the airplane can be delivered according to the schedule or not is determined. Therefore, the method for optimizing the flight test task allocation is a key problem to be solved in the field of flight tests of civil transport airplanes.

The flight test task allocation optimization problem is a complex sequence optimization problem. The trial-flight subjects are numerous, the attribute categories are complicated, and complicated constraint conditions and merging relations exist among the subjects. The problem of optimizing the flight test task allocation needs to be solved how to sequence the trial flight subjects under the condition of meeting the constraint conditions, so that the total flight test period is shortest. The conventional flight test task allocation mainly adopts an expert experience method, and manually allocates a test flight task by utilizing the expert experience. The method has the problems of poor distribution result, strong subjectivity, low efficiency, difficulty in considering all constraints and large workload. Therefore, the method for optimizing the flight test task allocation is designed, and the method is particularly critical for quickly obtaining the optimal flight test task allocation scheme through the constructed mathematical model.

Disclosure of Invention

In order to overcome the defects of the prior art, the invention provides a method for optimizing the distribution of flight test tasks of civil transport aircrafts. The invention discloses a method for optimizing the distribution of flight test tasks of civil transport aircrafts, which is used for solving the problems of poor distribution result, strong subjectivity, low efficiency, difficulty in considering all constraints and large workload in the prior art.

The technical scheme adopted by the invention for solving the technical problem comprises the following steps:

step 1: initializing parameters:

initializing parameters used by a flight test task allocation optimization problem, and initializing test flight subject parameters and genetic algorithm parameters;

step 2: coding and population initialization:

coding a chromosome by using a subject number, obtaining NP subject sequences by generating NP random arrangements with non-repetitive natural numbers of 1-N, obtaining NP chromosomes after coding, wherein each chromosome represents one arrangement of trial-flight subject numbers of 1-N, and each gene in the chromosome is a corresponding subject number in the subject sequence; population refers to a collection of several chromosomes, thus obtaining an initial population of size NP, denoted chromosome K S_kAs follows:

in formula (1), k is 1, 2.. NP,

expressed as chromosome ith gene corresponding to subject number of ith subject in the subject sequence, i is 1, 2.

And step 3: carrying out chromosome mutation operation;

carrying out mutation operation on chromosomes in the current population, wherein the mutation probability is p _ mut, which means that the probability of p _ mut exists in each chromosome is carried out, and the mutation method is that a section of gene sequence in the chromosomes is randomly selected, the position and the length are random, and the gene sequence is randomly disordered;

and 4, step 4: performing chromosome crossing operation;

performing crossover operation on chromosomes in the population subjected to mutation operation in the step 3, and selecting chromosome S_2tAnd chromosome S_2t-1Wherein t is 1,2, NP/2, called chromosome S_2tAnd chromosome S_2t-1Performing cross operation on the two parent chromosomes for the parent chromosomes, wherein the cross probability is p _ cro, and the probability that the two parent chromosomes have the p _ cro is indicated to perform the cross operation;

and 5: judging the current chromosome S_crtWhether the chromosomes accord with the logic constraint or not and generating a trial flight subject arrangement result for the chromosomes which accord with the logic constraint;

step 6: and (3) calculating an objective function value:

the flight test task allocation optimization target is that the total period of the flight test is shortest, so that the target function is the total period of the flight test, the target function value of the subject sequence corresponding to each chromosome is calculated, and for the test flight arrangement result X, the calculation steps of the target function are as follows:

and 7: if all chromosomes in the population have been traversed, go to step 8, otherwise, go to chromosome S_crt+1Repeating the step 5;

and 8: selecting operation is carried out in a binary championship selection mode to obtain a new population;

and step 9: repeating the steps 3-8 until the iteration number reaches the maximum iteration number NM, outputting the chromosome with the minimum objective function value in the population as the optimal subject arrangement result, and recording as X^*Outputting the shortest flight test total period f (X)^*)。

In step 1, the parameter initialization includes:

a) initializing trial flight subject parameters;

the number of trial flying subjects is N, the number of the trial flying subjects is N, N is 1,2,_nthe merged title number set of the titles is combsbj_nThe preset subject number set of subjects is prsbj_n；

b) Initializing parameters of a genetic algorithm;

the initialization population size is NP, the variation probability is p _ mut, the cross probability is p _ cro, and the maximum iteration number is NM.

In step 4, the cross operation method comprises: a segment of the gene on one parent chromosome is randomly selected, rearranged with reference to the relative position of the same gene on the other parent chromosome, and the same operation is performed on the other chromosome.

The detailed steps in step 5 are as follows:

a) initializing a subject arrangement set, and setting a trial flight subject arrangement result X as follows:

X＝[x₁ x₂ ... x_t ... x_K] (3)

in the formula (3), X is the subject arrangement result of the testing machine, and X_tDefining the t item subject set for the tester, K the total number of subject sets for the tester, and mainsbj_tIs a subject set x_tSubject of (1), timesg _ st_tAnd timed _ ed_tAre respectively a subject sequence x_tThe start and end times of (c); initializing X as a null vector, arranging a subject set oversbj as a null set, setting a plane _ ed value of a current test machine test flight task ending time as 0, setting a current subject set sequence number t as 1, and setting X as a current subject set sequence number₁Is an empty set;

b) sequentially selecting subjects according to the current chromosome gene order, wherein the selected subject number is s; if the preposed subject set prsbj of the subject s_sThe subject in (1) is arranged, namely the subject satisfies:

then step c) is performed; if the current chromosome does not accord with the formula (4), the current chromosome does not accord with the logic constraint, and the step 7 is carried out;

c) if the current subject set x_tFor an empty set, s is added to x_tZhong, and make mainsbj_tD, performing step s; if the current subject set x_tIf not, then determine if the subject s can be merged into x_tIf the subject s satisfies formula (5):

then subject s is added to x_tContinuing the step d; if equation (5) is not satisfied, the plane _ ed is made timeg_ed_tT plus 1, initialize x_tRepeating step c for the empty set;

d) determining a subject sequence x_tIs started and ended at the time timesg _ st_tAnd timed _ ed_tThe calculation formula is as follows:

and if the distribution of all the subject sequences corresponding to the current chromosome is finished, the chromosome accords with the logic constraint, the step 6 is continued, otherwise, the s is added into the set oversbj, and the step b) is repeated.

In step 8, the binary tournament selection mode adopts a return sampling mode, two chromosome sequences are randomly extracted from the previous generation population, and if the two chromosome sequences both accord with logic constraint, the chromosome sequence with the minimum objective function value calculated by the formula (6) is selected and reserved in the next generation population set; if the two chromosome sequences which accord with the logic constraint have the same objective function value, randomly selecting one chromosome sequence and reserving the selected chromosome sequence into the next generation of population; if one of the two chromosomes accords with the logic constraint, the chromosomes which accord with the logic constraint are reserved in the next generation of population; if the two chromosomes do not accord with the logic constraint, one of the chromosomes is randomly selected and reserved in the next generation of population, and the operation of the step is repeated until the next generation of population reaches NP.

The method has the advantages that the flight test arrangement result meeting the test flight constraint condition can be automatically obtained according to the flight test task information of the civil transport aircraft, and the subject sequence is optimized by adopting a genetic algorithm so as to achieve the effect of shortening the test flight period.

Drawings

Fig. 1 is a front-end relationship diagram of each test flight subject in the example.

FIG. 2 is a schematic diagram of an exemplary chromosomal mutation procedure.

FIG. 3 is a schematic diagram of chromosome crossing operations in an example.

FIG. 4 shows the results of the sequence 2 subject arrangement of the simplified example.

FIG. 5 is an example optimal task allocation result.

FIG. 6 is a flowchart of flight test mission allocation optimization steps.

Detailed Description

The invention is further illustrated with reference to the following figures and examples.

The specific implementation steps are as follows:

1. parameter initialization

a) Test flight subject parameter initialization

The number N of initial subjects is 40, and the detailed information of the trial-flight subjects is shown in table 1:

TABLE 1 details of trial flight subjects

The preposition relationship among the trial subjects is shown in fig. 1, in which the arrows point from the preposition subject to the postposition subject. According to the prepositive relationship, the subjects 1-30 are divided into 2 branches, each branch has no prepositive relationship, and the subjects 31-40 are scattered subjects, i.e. no prepositive subject or postpositive subject.

b) Initializing parameters of a genetic algorithm;

the initial population size NP is 500, the variation probability p _ mut is 0.1, the cross probability p _ cro is 0.3, and the maximum number of iterations NM is 100.

2. And (3) coding and population initialization, generating 500 random arrays with non-repeated natural numbers of 1-40, and taking each array as a chromosome to obtain an initial population with the scale of 500.

3. And (3) carrying out mutation operation on each chromosome, wherein the mutation probability is 0.1, and randomly determining the position of a mutation fragment in the chromosome sequence. By chromosome S_kFor example, the gene sequence is

The mutation operation is shown in FIG. 2. Randomly selecting variant fragments to be 3-4-5-6, randomly disordering the sequence to be 5-4-6-3, and obtaining the variant gene sequence

4. Performing crossover operation on the 2t chromosome and the 2t-1 chromosome (t is 1, 2.., 250), wherein the crossover probability is 0.3, and the gene sequence is obtained

(parent chromosome S)_2t) And the gene sequence is

(parent chromosome S)_2t-1) For example, the crossover operation is shown in FIG. 3: selection of the parent chromosome S_2tThe start position of the cross-over segment is the third position, the end position is the fifth position, i.e. 3-4-5, according to gene 3, gene 4, gene 5 in the parent chromosome S_2t-1The order of (1) is that the rearranged chromosomes S are rearranged firstly by 3, then 5 and finally 4_2tIs composed of

For parent chromosome S_2t-1Performing the same operation to select the parent chromosome S_2t-1The crossing fragment of (1), the starting position of which is the third position and the ending position of which is the fifth position, i.e., 5-7-2, is the gene in the parent chromosome S_2tThe order of the above is 2, 5 and 7, and the rearranged chromosome S_2t-1The gene sequence is

5. For each chromosome it is determined whether it meets the logical constraints. A simplified example is shown to illustrate how to determine whether a chromosome meets a logical constraint and perform objective function value calculation with chromosomes that meet the logical constraint. The number of trial flight subjects in this example is 5, and the detailed information of the subjects is shown in the table:

table 2 example trial flight subject details

For subject sequence 1: 4-1-5-3-2, and family sequence 2: 1-2-3-4-5, performing flight test task allocation, and calculating objective function value

a) Sequence 1: 4-1-5-3-2, chromosome S₁＝[4 1 5 3 2]

1) Judging whether the subject 4 meets the logic constraint, wherein the plane _ ed is 0 and x since the subject 4 has no preposed subject₁＝[4]，oversbj＝[4],timesg_ed₁＝10。

2) It is determined whether subject 1 meets the logical constraint. Since subject 1 has no previous subjects and subject 1 is not a merged subject of subject 4, plane _ ed is 10, x₂＝[1]，timesg_ed₂＝40，oversbj＝[4 1]。

3) Determining whether subject 5 meets the logic constraint, since subject 5 has no previous subject and subject 5 is not the merged subject of subject 4, plane _ ed is 40, x₃＝[5]，timesg_ed₃＝45，oversbj＝[4 1 5]。

4) Determining whether subject 3 meets the logical constraint due to the pre-set prsbj of subject 3₃＝[1 2]Where subject 2 is not in the oversbj set, i.e.

Thus S₁Not a chromosome that meets logical constraints.

b) Sequence 2: 1-2-3-4-5, chromosome S₂＝[1 2 3 4 5]

1) Subject 1, subject 2, subject 3 in the sequence all meet the logical constraint. Subject 1, subject 2, subject 3 are assigned sequentially, in which case x₁＝[1]，x₂＝[2]，x₃＝[3]，oversbj＝[1 2 3]，timesg_ed₁＝30，timesg_ed₂＝45，timesg_ed₃＝60，plane_ed＝45。

2) And judging whether the subject 4 meets the logic constraint, wherein the subject 4 meets the logic constraint because the subject 4 does not have a preposed subject. Since subject 4 is the merged subject of subject 3, i.e., 4 ∈ combsbj₃Therefore, if the merged subjects need to be considered, then the plane _ ed is 45, x₃＝[3 4]，timesg_ed₃＝60。

3) And judging whether the subject 5 meets the logic constraint, wherein the subject 5 meets the logic constraint because the subject 5 does not have a front subject. Since subject 5 is the merged subject of subject 3, i.e., 5 ∈ combsbj₃Therefore, if the merged subjects need to be considered, then the plane _ ed is 45, x₃＝[3 4 5]，timesg_ed₃＝60。

4) The subject assignment result X [ [1 ] can be obtained] [2] [3 4 5]]，

I.e., the objective function value is 60, the result is shown in fig. 4.

6. Selecting operation: and (5) performing 500 times of selection operation to obtain a new population.

7. And (5) repeating the steps 3-6, knowing that the iteration times reach 100, outputting the chromosome with the minimum objective function value in the population as an optimal subject sequence. The obtained optimal sequence is as follows:

34-32-33-1-16-36-2-17-4-19-37-35-3-18-6-21-22-5-20-8-24-38-11-26-27-39-10-12-7-23-9-25-31-13-28-29-40-14-30-15

8. and performing task arrangement aiming at the optimal subject sequence to obtain an optimal task allocation result, as shown in fig. 5. In the figure, each second row rectangle represents a trial-flight subject set, the number in the rectangle represents a subject number included in the subject set, the length of the rectangle is proportional to the duration of the trial-flight subject set, the number above the rectangle represents the time corresponding to the corresponding scale, for example, the 4 th module includes a subject 1, a subject 16 and a subject 36, and the detailed arrangement of the subjects is as shown in the table:

TABLE 3 task assignment results Table

The operational flow is shown in fig. 6.

Claims (5)

1. A civil transportation type airplane flight test task allocation optimization method is characterized by comprising the following steps:

step 1: initializing parameters:

initializing parameters used by a flight test task allocation optimization problem, and initializing test flight subject parameters and genetic algorithm parameters;

step 2: coding and population initialization:

coding a chromosome by using a subject number, obtaining NP subject sequences by generating NP random arrangements with non-repetitive natural numbers of 1-N, obtaining NP chromosomes after coding, wherein each chromosome represents one arrangement of trial-flight subject numbers of 1-N, and each gene in the chromosome is a corresponding subject number in the subject sequence; population refers to a collection of several chromosomes, thus obtaining an initial population of size NP, denoted chromosome K S_kAs follows:

in formula (1), k is 1, 2.. NP,

expressed as chromosome ith gene corresponding to subject number of ith subject in the subject sequence, i is 1, 2.

And step 3: carrying out chromosome mutation operation;

carrying out mutation operation on chromosomes in the current population, wherein the mutation probability is p _ mut, which means that the probability of p _ mut exists in each chromosome is carried out, and the mutation method is that a section of gene sequence in the chromosomes is randomly selected, the position and the length are random, and the gene sequence is randomly disordered;

and 4, step 4: performing chromosome crossing operation;

staining in the population after performing the mutation operation in step 3Crossing over the body to select chromosome S_2tAnd chromosome S_2t-1Wherein t is 1,2, NP/2, called chromosome S_2tAnd chromosome S_2t-1Performing cross operation on the two parent chromosomes for the parent chromosomes, wherein the cross probability is p _ cro, and the probability that the two parent chromosomes have the p _ cro is indicated to perform the cross operation;

and 5: judging the current chromosome S_crtWhether the chromosomes accord with the logic constraint or not and generating a trial flight subject arrangement result for the chromosomes which accord with the logic constraint;

step 6: and (3) calculating an objective function value:

the flight test task allocation optimization target is that the total period of the flight test is shortest, so that the target function is the total period of the flight test, the target function value of the subject sequence corresponding to each chromosome is calculated, and for the test flight arrangement result X, the calculation steps of the target function are as follows:

and 7: if all chromosomes in the population have been traversed, go to step 8, otherwise, go to chromosome S_crt+1Repeating the step 5;

and 8: selecting operation is carried out in a binary championship selection mode to obtain a new population;

and step 9: repeating the steps 3-8 until the iteration number reaches the maximum iteration number NM, outputting the chromosome with the minimum objective function value in the population as the optimal subject arrangement result, and recording as X^*Outputting the shortest flight test total period f (X)^*)。

2. The civil transportation type aircraft flight test task allocation optimization method according to claim 1, characterized in that:

in step 1, the parameter initialization includes:

a) initializing trial flight subject parameters;

the number of trial flying subjects is N, the subject number is N, N is 1,2The duration is timesbj_nThe merged title number set of the titles is combsbj_nThe preset subject number set of subjects is prsbj_n；

b) Initializing parameters of a genetic algorithm;

the initialization population size is NP, the variation probability is p _ mut, the cross probability is p _ cro, and the maximum iteration number is NM.

3. The civil transportation type aircraft flight test task allocation optimization method according to claim 1, characterized in that:

in step 4, the cross operation method comprises: a segment of the gene on one parent chromosome is randomly selected, rearranged with reference to the relative position of the same gene on the other parent chromosome, and the same operation is performed on the other chromosome.

4. The civil transportation type aircraft flight test task allocation optimization method according to claim 1, characterized in that:

the detailed steps in step 5 are as follows:

a) initializing a subject arrangement set, and setting a trial flight subject arrangement result X as follows:

X＝[x₁ x₂ ... x_t ... x_K] (3)

in the formula (3), X is the subject arrangement result of the testing machine, and X_tDefining the t item subject set for the tester, K the total number of subject sets for the tester, and mainsbj_tIs a subject set x_tSubject of (1), timesg _ st_tAnd timed _ ed_tAre respectively a subject sequence x_tThe start and end times of (c); initializing X as a null vector, arranging a subject set oversbj as a null set, setting a plane _ ed value of a current test machine test flight task ending time as 0, setting a current subject set sequence number t as 1, and setting X as a current subject set sequence number₁Is an empty set;

b) sequentially selecting subjects according to the current chromosome gene order, wherein the selected subject number is s; if the preposed subject set prsbj of the subject s_sThe subject in (1) is arranged, namely the subject satisfies:

then step c) is performed; if the current chromosome does not accord with the formula (4), the current chromosome does not accord with the logic constraint, and the step 7 is carried out;

c) if the current subject set x_tFor an empty set, s is added to x_tZhong, and make mainsbj_tD, performing step s; if the current subject set x_tIf not, then determine if the subject s can be merged into x_tIf the subject s satisfies formula (5):

then subject s is added to x_tContinuing the step d; if equation (5) is not satisfied, let plane _ ed be timesg _ ed_tT plus 1, initialize x_tRepeating step c for the empty set;

d) determining a subject sequence x_tIs started and ended at the time timesg _ st_tAnd timed _ ed_tThe calculation formula is as follows:

and if the distribution of all the subject sequences corresponding to the current chromosome is finished, the chromosome accords with the logic constraint, the step 6 is continued, otherwise, the s is added into the set oversbj, and the step b) is repeated.

5. The civil transportation type aircraft flight test task allocation optimization method according to claim 1, characterized in that:

in step 8, the binary tournament selection mode adopts a return sampling mode, two chromosome sequences are randomly extracted from the previous generation population, and if the two chromosome sequences both accord with logic constraint, the chromosome sequence with the minimum objective function value calculated by the formula (6) is selected and reserved in the next generation population set; if the two chromosome sequences which accord with the logic constraint have the same objective function value, randomly selecting one chromosome sequence and reserving the selected chromosome sequence into the next generation of population; if one of the two chromosomes accords with the logic constraint, the chromosomes which accord with the logic constraint are reserved in the next generation of population; if the two chromosomes do not accord with the logic constraint, one of the chromosomes is randomly selected and reserved in the next generation of population, and the operation of the step is repeated until the next generation of population reaches NP.

Images