CN115391991A - Loader shoveling process optimization method based on Gaussian process regression and genetic algorithm - Google Patents
Loader shoveling process optimization method based on Gaussian process regression and genetic algorithm Download PDFInfo
- Publication number
- CN115391991A CN115391991A CN202210905903.9A CN202210905903A CN115391991A CN 115391991 A CN115391991 A CN 115391991A CN 202210905903 A CN202210905903 A CN 202210905903A CN 115391991 A CN115391991 A CN 115391991A
- Authority
- CN
- China
- Prior art keywords
- population
- gene sequence
- evolution
- loader
- gaussian
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Genetics & Genomics (AREA)
- Data Mining & Analysis (AREA)
- Physiology (AREA)
- Geometry (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Computer Hardware Design (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a loader shoveling process optimization method based on Gaussian process regression and genetic algorithm, which comprises the following steps of: the method includes the steps that manual operation is carried out for multiple shovel loading field operations, loader displacement, movable arm oil cylinder displacement, rotary bucket oil cylinder displacement and insertion instantaneous speed of a point A of four stages of AB, BC, CD and DE of each operation are extracted, and multiple groups of track characteristic parameters are obtained; the method comprises the following steps: introducing a fitness function, wherein the fitness function is a Gaussian process regression model, performing population evolution, and setting population evolution times, and each evolution process is as follows: (1) selecting; (2) and a crossing process; (3) and a mutation process; (4) and an evolution process; (5) repeating the steps (1) - (4) until the set population evolution times are reached to obtain the population after the evolution is finished; fifthly, decoding: and obtaining the optimal track characteristic parameters. The shovel loading automatic operation track of the loader optimized by the method is more accurate, and the working efficiency is higher.
Description
Technical Field
The invention belongs to the technical field of machinery, and particularly relates to a loader shoveling process optimization method based on Gaussian process regression and a genetic algorithm.
Background
Many researchers have studied how to improve the operating efficiency of loaders. Optimizing the loader shovel trajectory without modifying the loader hardware or providing an auxiliary system is one of the effective ways to improve the efficiency of loader operations. In the research of the optimization of a plurality of tracks, most methods are based on simulation experiments, establish simulation models and optimize the shovel loading track. However, a certain deviation exists between the theoretical trajectory and the actual shovel loading operation trajectory, and the environment where the loader is in the shovel loading operation is complex, so that the problems cannot be well solved through simulation experiments.
Disclosure of Invention
The invention provides a loader shoveling process optimization method based on Gaussian process regression and genetic algorithm.
In order to achieve the purpose, the technical scheme of the invention is as follows:
the loader shoveling process optimization method based on Gaussian process regression and genetic algorithm comprises the following steps:
the method includes the steps that a single shoveling operation of a loader is decomposed into an AB stage of inserting a material pile, a BC stage of shoveling, a CD stage of idling the loader and a DE stage of lifting a bucket; manually operating to carry out multiple on-site operations, extracting displacement of a loader, displacement of a movable arm oil cylinder, displacement of a rotating bucket oil cylinder and insertion instantaneous speed of a point A at four stages of AB, BC, CD and DE of each operation, and taking the parameters of each operation as a group of track characteristic parameters to obtain multiple groups of track characteristic parameters; wherein, A is the point that the bucket contacts the material pile, B is the point that the bucket finishes inserting, C is the point that the shovel finishes loading, D is the point that begins to lift the bucket, E is the end point of the whole shovel loading operation;
the method comprises the following steps: taking a plurality of groups of track characteristic parameters obtained in the step, taking each track characteristic parameter in each group as chromosome information, sequencing the chromosome information to arrange the chromosome information into a row of gene sequences, obtaining a plurality of rows of gene sequences, setting evolution times, and generating an initial population, wherein the number of the groups of the track characteristic parameters is the scale number n of the initial population;
introducing a fitness function, wherein the fitness function is a Gaussian process regression model, and the formula is as follows:
K yy* ＝(k(x 1 ,x * ) k(x 2 ,x * ) … k(x n ,x * )) T (3)
K y*y* ＝k(x*,x*) (4)
wherein y is an output variable, and y is a predicted value;
inputting each gene sequence in the initial population into a fitness function to obtain efficiency values of each gene sequence, wherein the efficiency values are the fitness of each gene sequence;
fourth, carrying out population evolution, and setting the number of times of population evolution, wherein each evolution process is as follows:
(1) and a selection process: based on a roulette method, the higher the fitness value of the gene sequences, the higher the probability of selection, the n times of rotation of the roulette wheel, and each time of rotation of the roulette wheel, a row of gene sequences are selected from a target population to generate a new population containing n groups of gene sequences as a parent population; the target population is the initial population at the beginning;
(2) and a crossing process: setting the crossover probability to be 0.4-1, randomly selecting two groups of gene sequences in the parent population, randomly selecting positions on the gene sequences to exchange chromosome information to obtain new gene sequences, and performing n times of crossover to generate a new population containing n groups of gene sequences to obtain a crossover sub-population;
(3) and a mutation process: setting the variation probability to be 0.01-0.1, randomly selecting a certain gene sequence in the cross sub-population, carrying out variation at a certain position on the gene sequence, changing chromosome information to obtain a new gene sequence, carrying out n-time variation to generate a new population containing n groups of gene sequences, and obtaining a variation sub-population;
(4) and the evolution process: calculating the fitness of each gene sequence in the variant sub-population and the parent population, and selecting the gene sequences of the first n groups with the maximum fitness in the variant sub-population and the parent population to obtain a new generation population;
(5) inputting the new generation of population generated in the step (4) into the step (1) as a target population, and repeating the steps (1) - (4) until the set population evolution times are reached to obtain a population after the evolution is finished;
fifthly, decoding: and outputting the gene sequence corresponding to the optimal fitness in the population after evolution is completed to obtain the optimal track characteristic parameters, and completing the optimization process.
The step three is that the derivation process of the regression model of the Gaussian process is as follows:
the gaussian process consists of a mean function and a kernel function, and the expression is:
y～N(μ(x),k(x,x')) (5)
in the formula: μ (x) is the mean function and k (x, x') is the kernel function.
Using a gaussian kernel as kernel function:
in the formula: sigma f 、σ l Is a hyperparameter, δ is a dirac function, δ =1 when x = x', otherwise is zero;
to simplify the calculation, the mean function is chosen to be 0, i.e., μ (x) =0;
setting the value to be predicted as y, and according to the definition of the Gaussian process, obeying the Gaussian distribution to any limited linear combination of random variables in the Gaussian process, and then conforming the output variable y and the predicted value y to the joint Gaussian distribution, wherein the joint Gaussian distribution function is as follows:
obtaining a Gaussian process regression model, namely a formula (1), through Bayesian derivation;
predicted mean of Gaussian process regression model
Comprises the following steps:
in the step (1), the selection process specifically comprises the following steps:
the calculation formula of each gene sequence in the target population, namely the individual selection probability is as follows:
randomly numbering each gene sequence in a target population to obtain n groups of numbered gene sequences, respectively calculating the selection probability of each gene sequence based on a formula (9), then accumulatively adding the selection probabilities of all the gene sequences to obtain n groups of accumulated addition probabilities respectively corresponding to each numbered gene sequence, and rotating a wheel disc for n times; randomly generating a random number between 0 and 1 by rotating each time, checking the interval of the accumulative addition probability in which the random number falls, selecting the accumulative addition probability upwards, and extracting the gene sequence with the number corresponding to the accumulative addition probability; finally obtaining n groups of gene sequences, and carrying out random numbering to obtain the parent population.
In the step (2), the crossing process specifically comprises:
a. performing interleaved individual selection
Randomly generating two numbers between 0 and 1, then multiplying n, performing upward integer to obtain two integer values, and selecting the gene sequences with numbers corresponding to the two integer values in the parent population as individuals for performing cross;
b. whether to perform a cross operation determination
Randomly generating a number from 0 to 1, if the number is greater than the crossover probability, not executing crossover operation, and if the number is less than the crossover probability, executing crossover operation;
c. cross location selection
Randomly generating a number from 0 to 1, multiplying the number by the length of the gene sequence, namely multiplying the number of chromosomes contained in the gene sequence, performing upward integer extraction to obtain the position of the chromosome to be exchanged in the gene sequence, and exchanging the chromosomes of the position of two individuals performing intersection;
d. and (5) performing the operation of the steps a-c for n times, wherein the serial number of the gene sequence after each exchange is unchanged, and obtaining a cross-over sub-population.
In the step (3), the mutation process specifically comprises:
a. individual selection of variants
Randomly generating a number from 0 to 1, then multiplying n, performing upward integer acquisition, and selecting a gene sequence with a number corresponding to the integer value in the cross sub-population as an individual for executing variation;
b. whether to perform mutation operations
Randomly generating a number from 0 to 1, if the number is greater than the mutation probability, not executing mutation operation, and if the number is less than the mutation probability, executing mutation operation;
c. variant location selection
Randomly generating a number of 0 to 1, multiplying the number by the length of the gene sequence, namely multiplying the number of chromosomes contained in the gene sequence, and performing upward integer acquisition to obtain the position of the chromosome to be mutated in the gene sequence;
d. generation of variant values
A ij = value at jth position of ith individual, i is determined by individual selection of variant in step a, j is determined by variant position selection in step c; a. The 1j Is the lower limit value at the j position; a. The 2j Is the upper limit value at the j position;
randomly generating a number from 0 to 1:
if the number is greater than 0.5, the resulting variance value is calculated as:
if the number is less than 0.5, the resulting variance value is calculated as follows:
replacing the chromosome to be mutated in the mutated individual with the calculated mutation value;
e. and (5) performing the operations of the steps a-d for n times, wherein the serial number of the gene after each mutation is unchanged, and obtaining a variant sub population.
The value range of n is 20-50.
The invention has the beneficial effects that:
according to the invention, the characteristic parameters of the track and the operation efficiency value of each unit of the loader at each stage of the shoveling operation are obtained through a large number of experiments, a proxy model is established by utilizing Gaussian process regression, the shoveling track of the loader is optimized by using a genetic algorithm and aiming at improving the operation efficiency value, the optimal displacement of the loader, the displacement of a movable arm oil cylinder, the displacement of a rotary bucket oil cylinder and the insertion instantaneous speed of a point A in the shoveling process are realized, the shoveling automatic operation track of the optimized loader is more accurate, the working efficiency is higher, and the reliability of the automatic shoveling operation is further improved.
Drawings
FIG. 1 is an exploded view of a single scooping operation of the loader of the present invention;
FIG. 2 is a schematic representation of a regression model of the Gaussian process of the present invention;
FIG. 3 is a flow chart of a genetic algorithm of the present invention;
FIG. 4 is a genetic algorithm optimization map of the present invention;
Detailed Description
The present invention will be described in detail below with reference to specific embodiments in conjunction with the accompanying drawings.
Example 1
The loader shoveling process optimization method based on Gaussian process regression and genetic algorithm comprises the following steps:
the method includes the steps that as shown in fig. 1, a single shoveling operation of a loader is decomposed into a material pile inserting AB stage, a shoveling BC stage, a loader idling CD stage and a bucket lifting DE stage; manually operating to carry out multiple on-site operations, extracting loader displacement, movable arm oil cylinder displacement, rotary bucket oil cylinder displacement and insertion instantaneous speed of point A of four stages of AB, BC, CD and DE of each operation, and taking the parameters of each operation as a group of track characteristic parameters to obtain 100 groups of track characteristic parameters; wherein, A is the point that the bucket contacts the material pile, B is the point that the bucket finishes inserting, C is the point that the shovel finishes loading, D is the point that begins to lift the bucket, E is the end point of the whole shovel loading operation;
the method comprises the following steps: taking 20 groups of the obtained track characteristic parameters at random, taking each track characteristic parameter in each group as chromosome information, sequencing the chromosome information to arrange the chromosome information into a row of gene sequences, obtaining a plurality of rows of gene sequences, setting evolution times, and generating an initial population t 0 The number of groups of the track characteristic parameters is the scale number n of the initial population;
introducing a fitness function, wherein the fitness function is a Gaussian process regression model, a schematic diagram of the Gaussian process regression model is shown in FIG. 2, and the formula is as follows:
K yy* ＝(k(x 1 ,x * ) k(x 2 ,x * ) … k(x n ,x * )) T (3)
K y*y* ＝k(x*,x*) (4)
wherein y is an output variable, and y is a predicted value;
inputting each gene sequence in the initial population into a fitness function to obtain efficiency values of each gene sequence, wherein the efficiency values are the fitness of each gene sequence;
fourthly, performing population evolution, wherein the number of times of population evolution is set to be 1000, and the evolution process of each time is as follows:
(1) and a selection process: based on a roulette method, the higher the fitness value of the gene sequences, the higher the probability of selection, the n times of rotation of the roulette wheel, and each time of rotation of the roulette wheel, a row of gene sequences are selected from a target population to generate a new population containing n groups of gene sequences as a parent population; at the beginning, the target population is an initial population;
the selection process specifically comprises the following steps:
the calculation formula of each gene sequence in the target population, namely the individual selection probability is as follows:
randomly numbering each gene sequence in a target population to obtain n groups of numbered gene sequences, respectively calculating the selection probability of each gene sequence based on a formula (9), then accumulatively adding the selection probabilities of all the gene sequences to obtain n groups of accumulated addition probabilities respectively corresponding to each numbered gene sequence, and rotating a wheel disc for n times; randomly generating a random number between 0 and 1 by rotating each time, checking the interval of the accumulative addition probability in which the random number falls, selecting the accumulative addition probability upwards, and extracting the gene sequence with the number corresponding to the accumulative addition probability; finally obtaining n groups of gene sequences, and carrying out random numbering to obtain a parent population;
(2) and a crossing process: setting the crossover probability to be 0.6, randomly selecting two groups of gene sequences in the parent population, randomly selecting positions on the gene sequences to exchange chromosome information to obtain new gene sequences, and performing n times of crossover to generate a new population containing n groups of gene sequences to obtain a crossover sub-population;
the crossing process specifically comprises the following steps:
a. performing interleaved individual selection
Randomly generating two numbers between 0 and 1, then multiplying n, performing upward integer extraction to obtain two integer values, and selecting the gene sequences with numbers corresponding to the two integer values in the parent population as individuals for executing intersection;
b. whether to perform a cross operation determination
Randomly generating a number from 0 to 1, if the number is greater than the crossover probability, not executing crossover operation, and if the number is less than the crossover probability, executing crossover operation;
c. cross location selection
Randomly generating a number from 0 to 1, multiplying the number by the length of the gene sequence, namely multiplying the number of chromosomes contained in the gene sequence, performing upward integer extraction to obtain the position of the chromosome to be exchanged in the gene sequence, and exchanging the chromosomes of the position of two individuals performing intersection;
d. and (5) performing the operation of the steps a-c for n times, wherein the serial number of the gene sequence after each exchange is unchanged, and obtaining a cross-over sub-population.
(3) And a mutation process: setting the variation probability to be 0.1, randomly selecting a certain gene sequence in the cross sub-population, carrying out variation at a certain position on the gene sequence, changing chromosome information to obtain a new gene sequence, carrying out n-time variation to generate a new population containing n groups of gene sequences, and obtaining a variation sub-population;
the mutation process specifically comprises the following steps:
a. individual selection of variants
Randomly generating a number from 0 to 1, then multiplying n, performing upward integer extraction, and selecting a gene sequence with a number corresponding to the integer value in the cross sub-population as an individual for performing variation;
b. whether to perform mutation operations
Randomly generating a number from 0 to 1, if the number is greater than the mutation probability, not executing mutation operation, and if the number is less than the mutation probability, executing mutation operation;
c. variant location selection
Randomly generating a number of 0 to 1, multiplying the number by the length of the gene sequence, namely multiplying the number of chromosomes contained in the gene sequence, and performing upward integer acquisition to obtain the position of the chromosome to be mutated in the gene sequence;
d. generation of variant values
A ij = value at jth position of ith individual, i is determined by individual selection of variant in step a, j is determined by variant position selection in step c; a. The 1j Is the lower limit value at the j position; a. The 2j Is the upper limit value at the j position;
randomly generating a number from 0 to 1:
if the number is greater than 0.5, the resulting variance value is calculated as:
if the number is less than 0.5, the resulting variance value is calculated as follows:
replacing the chromosome to be mutated in the mutated individual with the calculated mutation value;
e. performing the operation of the steps a-d for n times, wherein the serial number of the gene sequence after each mutation is unchanged, and obtaining a variant sub population;
(4) and the evolution process: calculating the fitness of each gene sequence in the variation subspecies and the parent population, and selecting the gene sequences of the first n groups with the maximum fitness in the variation subspecies and the parent population to obtain a new generation population t 1 ；
(5) Inputting the new generation of population generated in the step (4) into the step (1) as a target population, and repeating the steps (1) - (4) until the set population evolution times are reached to obtain a population after evolution is finished;
fifthly, decoding: and outputting the gene sequence corresponding to the optimal fitness in the population after evolution is completed to obtain the optimal track characteristic parameters, and completing the optimization process.
Fifthly, optimizing a result by a genetic algorithm:
as shown in fig. 4, the efficiency corresponding to the optimal trajectory characteristic parameter is improved by 13.6% relative to the highest efficiency in 100 samples.
Example 2
Example 1 gaussian process regression model:
K yy* ＝(k(x 1 ,x * ) k(x 2 ,x * ) … k(x n ,x * )) T (3)
K y*y* ＝k(x*,x*) (4)
wherein y is an output variable, and y is a predicted value;
the derivation process of the gaussian process regression model is as follows:
the gaussian process consists of a mean function and a kernel function, and the expression is:
y～N(μ(x),k(x,x')) (5)
in the formula: μ (x) is the mean function and k (x, x') is the kernel function.
Using a gaussian kernel as kernel function:
in the formula: sigma f 、σ l Is a hyper-parameter, δ is a dirac function, δ =1 when x = x', otherwise is zero;
to simplify the calculation, the mean function is chosen to be 0, i.e., μ (x) =0;
and setting the value to be predicted as y, and according to the definition of the Gaussian process, subjecting the linear combination of any limited random variables in the Gaussian process to Gaussian distribution, so that the output variable y and the predicted value y also conform to joint Gaussian distribution, and the joint Gaussian distribution function is as follows:
obtaining a Gaussian process regression model, namely a formula (1), through Bayesian derivation;
predicted mean of Gaussian process regression model
Comprises the following steps:
gaussian process regression model error:
using a determining coefficient R 2 Mean absolute percent error MAPE and root mean square error RMSE to evaluate the fit of gaussian regression:
in the formula, y represents a predicted value and y represents a true value.
In example 1, a gaussian process regression model was constructed from 100 sets of test samples, and model errors were calculated based on equations (12), (13), and (14) as shown in table 1 below:
TABLE 1 Gaussian Process regression model error
The number of test samples can be further increased to construct a Gaussian process regression model, so that the model error can be further reduced.
Claims (6)
1. A loader shovel loading process optimization method based on Gaussian process regression and genetic algorithm is characterized by comprising the following steps:
the method includes the steps that a single shoveling operation of a loader is decomposed into an AB stage for inserting a material pile, a BC stage for shoveling, a CD stage for idling of the loader and a DE stage for lifting a bucket; manually operating to carry out multiple on-site operations, extracting loader displacement, movable arm oil cylinder displacement, rotary bucket oil cylinder displacement and insertion instantaneous speed of point A of four stages of AB, BC, CD and DE of each operation, and taking the parameters of each operation as a group of track characteristic parameters to obtain multiple groups of track characteristic parameters; wherein, A is the point that the bucket contacts the material pile, B is the point that the bucket finishes inserting, C is the point that the shovel finishes loading, D is the point that begins to lift the bucket, E is the end point of the whole shovel loading operation;
secondly, encoding: taking multiple sets of track characteristic parameters obtained in the step, taking each track characteristic parameter in each set as chromosome information, sequencing the chromosome information to arrange the chromosome information into a row of gene sequences, obtaining multiple rows of gene sequences, setting evolution times, and generating an initial population, wherein the number of the sets of the track characteristic parameters is the scale number n of the initial population;
introducing a fitness function, wherein the fitness function is a Gaussian process regression model, and the formula is as follows:
wherein y is an output variable, and y is a predicted value;
inputting each gene sequence in the initial population into a fitness function to obtain efficiency values of each gene sequence, wherein the efficiency values are the fitness of each gene sequence;
fourth, carrying out population evolution, and setting the number of times of population evolution, wherein each evolution process is as follows:
(1) and selecting the process: based on a roulette method, the higher the fitness value of the gene sequences, the higher the probability of selection, the n times of rotation of the roulette wheel, and each time of rotation of the roulette wheel, a row of gene sequences are selected from a target population to generate a new population containing n groups of gene sequences as a parent population; at the beginning, the target population is an initial population;
(2) and a crossing process: setting the crossover probability to be 0.4-1, randomly selecting two groups of gene sequences in the parent population, randomly selecting positions on the gene sequences to exchange chromosome information to obtain new gene sequences, and performing n times of crossover to generate a new population containing n groups of gene sequences to obtain a crossover sub-population;
(3) and a mutation process: setting the variation probability to be 0.01-0.1, randomly selecting a certain gene sequence in the cross sub-population, carrying out variation at a certain position on the gene sequence, changing chromosome information to obtain a new gene sequence, carrying out n-time variation to generate a new population containing n groups of gene sequences, and obtaining a variation sub-population;
(4) and the evolution process: calculating the fitness of each gene sequence in the variant sub-population and the parent population, and selecting the gene sequences of the first n groups with the maximum fitness in the variant sub-population and the parent population to obtain a new generation population;
(5) inputting the new generation of population generated in the step (4) into the step (1) as a target population, and repeating the steps (1) - (4) until the set population evolution times are reached to obtain a population after the evolution is finished;
fifthly, decoding: and outputting the gene sequence corresponding to the optimal fitness in the population after evolution is completed to obtain the optimal track characteristic parameters, and completing the optimization process.
2. The loader shovel process optimization method based on gaussian process regression and genetic algorithm of claim 1, wherein:
the step three is that the derivation process of the regression model of the Gaussian process is as follows:
the gaussian process consists of a mean function and a kernel function, and the expression is:
y～N(μ(x),k(x,x')) (5)
in the formula: μ (x) is the mean function and k (x, x') is the kernel function.
Using a gaussian kernel function as the kernel function:
in the formula: sigma f 、σ l Is a hyperparameter, δ is a dirac function, δ =1 when x = x', otherwise is zero;
to simplify the calculation, the mean function is chosen to be 0, i.e., μ (x) =0;
and setting the value to be predicted as y, and according to the definition of the Gaussian process, subjecting the linear combination of any limited random variables in the Gaussian process to Gaussian distribution, so that the output variable y and the predicted value y also conform to joint Gaussian distribution, and the joint Gaussian distribution function is as follows:
obtaining a Gaussian process regression model, namely a formula (1), through Bayesian derivation;
predicted mean of Gaussian process regression model
Comprises the following steps:
3. the loader shovel process optimization method based on gaussian process regression and genetic algorithm according to claim 1, characterized in that:
in the step (1), the selection process is specifically as follows:
the calculation formula of each gene sequence in the target population, namely the individual selection probability is as follows:
randomly numbering each gene sequence in a target population to obtain n groups of numbered gene sequences, respectively calculating the selection probability of each gene sequence based on a formula (9), then accumulatively adding the selection probabilities of all the gene sequences to obtain n groups of accumulated addition probabilities respectively corresponding to each numbered gene sequence, and rotating a wheel disc for n times; randomly generating a random number between 0 and 1 by rotating each time, checking the interval of the accumulative addition probability in which the random number falls, selecting the accumulative addition probability upwards, and extracting the gene sequence with the number corresponding to the accumulative addition probability; finally obtaining n groups of gene sequences, and carrying out random numbering to obtain the parent population.
4. The loader shovel process optimization method based on gaussian process regression and genetic algorithm according to claim 3, characterized in that:
in the step (2), the crossing process specifically comprises:
a. performing interleaved individual selection
Randomly generating two numbers between 0 and 1, then multiplying n, performing upward integer extraction to obtain two integer values, and selecting the gene sequences with numbers corresponding to the two integer values in the parent population as individuals for executing intersection;
b. whether to perform a cross operation determination
Randomly generating a number from 0 to 1, if the number is greater than the crossover probability, not executing crossover operation, and if the number is less than the crossover probability, executing crossover operation;
c. cross location selection
Randomly generating a number from 0 to 1, multiplying the number by the length of the gene sequence, namely multiplying the number of chromosomes contained in the gene sequence, performing upward integer extraction to obtain the position of the chromosome to be exchanged in the gene sequence, and exchanging the chromosomes of the position of two individuals performing intersection;
d. and (5) performing the operation of the steps a-c for n times, wherein the serial number of the gene after each exchange is unchanged, and thus a crossed sub population is obtained.
5. The loader shovel process optimization method based on gaussian process regression and genetic algorithm according to claim 4, characterized in that:
in the step (3), the mutation process specifically comprises:
a. individual selection of variants
Randomly generating a number from 0 to 1, then multiplying n, performing upward integer extraction, and selecting a gene sequence with a number corresponding to the integer value in the cross sub-population as an individual for performing variation;
b. whether to perform mutation operations
Randomly generating a number from 0 to 1, if the number is greater than the mutation probability, not executing mutation operation, and if the number is less than the mutation probability, executing mutation operation;
c. variant location selection
Randomly generating a number of 0 to 1, multiplying the number by the length of the gene sequence, namely multiplying the number of chromosomes contained in the gene sequence, and performing upward integer acquisition to obtain the position of the chromosome to be mutated in the gene sequence;
d. generation of variant values
A ij = value at jth position of ith individual, i is determined by individual selection of variant in step a, j is determined by variant position selection in step c; a. The 1j Is the lower limit value at the j position; a. The 2j Is the upper limit value at the j position;
randomly generating a number from 0 to 1:
if the number is greater than 0.5, the resulting variance value is calculated as:
if the number is less than 0.5, the resulting variance value is calculated as:
replacing the chromosome to be mutated in the mutated individual with the calculated mutation value;
e. and (5) performing the operations of the steps a-d for n times, wherein the serial number of the gene after each mutation is unchanged, and obtaining a variant sub population.
6. The loader shovel process optimization method based on gaussian process regression and genetic algorithm according to claim 1, characterized in that:
the value range of n is 20-50.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210905903.9A CN115391991A (en) | 2022-07-29 | 2022-07-29 | Loader shoveling process optimization method based on Gaussian process regression and genetic algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210905903.9A CN115391991A (en) | 2022-07-29 | 2022-07-29 | Loader shoveling process optimization method based on Gaussian process regression and genetic algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115391991A true CN115391991A (en) | 2022-11-25 |
Family
ID=84116935
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210905903.9A Pending CN115391991A (en) | 2022-07-29 | 2022-07-29 | Loader shoveling process optimization method based on Gaussian process regression and genetic algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115391991A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116924287A (en) * | 2023-09-18 | 2023-10-24 | 临工重机股份有限公司 | Control method, device, equipment and medium of hydraulic compensation leveling mechanism |
| CN118551503A (en) * | 2024-07-29 | 2024-08-27 | 南京农业大学 | Method, device, equipment and medium for determining working parameters of seed supply device |
-
2022
- 2022-07-29 CN CN202210905903.9A patent/CN115391991A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116924287A (en) * | 2023-09-18 | 2023-10-24 | 临工重机股份有限公司 | Control method, device, equipment and medium of hydraulic compensation leveling mechanism |
| CN116924287B (en) * | 2023-09-18 | 2023-12-08 | 临工重机股份有限公司 | Control method, device, equipment and medium of hydraulic compensation leveling mechanism |
| CN118551503A (en) * | 2024-07-29 | 2024-08-27 | 南京农业大学 | Method, device, equipment and medium for determining working parameters of seed supply device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115391991A (en) | Loader shoveling process optimization method based on Gaussian process regression and genetic algorithm | |
| Haidar et al. | Genetic algorithms application and testing for equipment selection | |
| Tavakkoli-Moghaddain et al. | Facilities layout design by genetic algorithms | |
| CN102073311B (en) | Method for scheduling machine part processing line by adopting discrete quantum particle swarm optimization | |
| CN109635495B (en) | Arterial road phase difference simulation optimization method based on neural network and genetic algorithm | |
| CN105975701A (en) | Parallel scheduling disassembly path forming method based on mixing fuzzy model | |
| CN104732067A (en) | Industrial process modeling forecasting method oriented at flow object | |
| CN106773700A (en) | The Multi model Predictive Controllers of genetic algorithm optimization | |
| CN117195449A (en) | Wind farm arrangement optimization method, device, equipment and medium | |
| KR100930652B1 (en) | Pile placement design method of foundation under the structure using genetic algorithm | |
| CN113408839A (en) | Intelligent production model of industrial machinery based on block chain | |
| CN110069851B (en) | Rock-soil parameter random field inversion method | |
| CN116307139A (en) | Wind power ultra-short-term prediction method for optimizing and improving extreme learning machine | |
| CN115456286A (en) | Short-term photovoltaic power prediction method | |
| CN113378276B (en) | Composite foundation intelligent design method based on genetic algorithm and gene expression programming | |
| CN118333161B (en) | Optimization method based on multi-group competition multi-objective optimization confidence rule base | |
| CN110288726B (en) | Sugarcane transfer vehicle danger prediction method based on BP neural network | |
| CN110096741B (en) | Pre-forging forming die design method based on prediction model and improved genetic algorithm | |
| CN109902007B (en) | Test case generation method based on point dyeing model | |
| Honiden | Tree structure modeling and genetic algorithm-based approach to unequal-area facility layout problem | |
| CN114037006B (en) | Method for generating typical daily load curve of power system | |
| CN117236195B (en) | Machine learning offline agent model production optimization method for reducing development risk | |
| Wu et al. | Multiple sequence alignment using ga and nn | |
| CN118428408B (en) | Mountain torrent forecasting method based on multi-AI model fusion and parameter optimization | |
| CN118711692B (en) | Ornamental plant morphological feature inversion soil heavy metal content model generation method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |