Cross-region operation path optimization method for harvesters based on intelligent optimization algorithms

Abstract

A method for optimizing the cross-region operation path of a harvester based on an intelligent optimization algorithm, the steps are as follows: (1) Parameter setting; (2) Encoding method, the hybrid improved real-coded emperor penguin optimization algorithm adopts a real-coded method; (3) Decoding method; (4) Output the optimal operation path and the cross-region operation length of N operation points, end the loop, otherwise, enter the next step; (5) Select whether to perform an oscillating adaptive perturbation search strategy on the individual with the minimum fitness value; (6) Select to perform a random strategy or a dynamic movement strategy; (7) Execute local search operations to improve the convergence speed and solution quality of the algorithm; (8) Update the optimal individual and the population, and return to (4). The present invention combines the advantages of multiple intelligent optimization algorithms, aiming to improve the global and local search capabilities of the algorithm, thereby improving the solution quality and efficiency of the problem of optimizing the cross-region operation path of the harvester.

Description

Technical Field

[0001] The present invention relates to the technical field of harvester path planning, and particularly to a method for optimizing the cross-region operation path of a harvester based on an intelligent optimization algorithm. Background Art

[0002] Traditional path planning methods, such as genetic algorithms, particle swarm optimization, ant colony algorithms, etc., although have achieved certain effects in some applications, still have some limitations when solving the problem of optimizing the cross-region operation path of a harvester. For example, these methods may fall into local optima, or cannot find a solution that meets the actual requirements in large-scale agricultural machinery scheduling.

[0003] To solve the above problems, researchers have begun to explore new optimization algorithms and methods. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method for optimizing the cross-region operation path of a harvester based on an intelligent optimization algorithm, which can improve the solution quality and efficiency of the cross-region operation path optimization problem of the harvester by enhancing the global and local search capabilities of the algorithm, in view of the deficiencies of the prior art.

[0005] The technical problem to be solved by the present invention is realized through the following technical solutions. The present invention is a method for optimizing the cross-region operation path of a harvester based on an intelligent optimization algorithm, and the steps of the method are as follows:

[0006] (1) Parameter setting:

[0007] Based on the hybrid improved real-coded emperor penguin optimization algorithm, the population size is n, the number of iterations of the algorithm is 200, the initial iteration number gen = 0, the number of cross-region operation points is N, and the objective function of the cross-region operation path planning problem of a harvester with N operation points is expressed as:

[0008]

[0009] In the formula, f(D) is the shortest distance of the harvester's cross-region operation; t i represents the i-th operation point; d(t i , t i+1 ) represents the distance between the i-th operation point and the (i + 1)-th operation point, and is expressed by the following formula:

[0010]

[0011] (x i , y i ) and (x i+1 , y i+1 ) are the coordinates of t i , t i+1coordinates;

[0012] (2) Encoding method:

[0013] The real-coded emperor penguin optimization algorithm based on hybrid improvement adopts the real-coded method;

[0014] If the number of working points for the harvester's cross-regional operation is N, then the individual is initially

[0015] X = N * rand(1, N) (3)

[0016] where X is an individual with a code length of N and a floating-point encoded value;

[0017] Generate n individuals according to formula (3);

[0018] (3) Decoding method:

[0019] Taking the kth individual X in the population k =(x1, x2, …, x N ) as an example, the specific decoding steps are as follows:

[0020] (3.1) Randomly generate an integer μ between [1, N], and use μ as the starting working point;

[0021] (3.2) After deleting the μth component in X k , the resulting vector is X’k, X’k = (x1, x2, …, x μ-1 , x μ+1 , …, x N ), and according to the distances d=(d μ,1 , d μ,2 , …, d μ,μ-1 , d μ,μ+1 , …, d μ,N ) between μ and the remaining working points, calculate the vector Tp related to the distances between the working points and the gene position values. The calculation formula of Tp is

[0022]

[0023] In the formula, represents the multiplication of the elements at the same positions of vector d and vector ;

[0024] (3.3) Find the minimum value in vector Tp, and take the working point i corresponding to the minimum value as the next working point;

[0025] (3.4) Let μ = i, and repeat Step 2 and Step 3 to determine the access order of all working points;

[0026] Decode n individuals according to Equation (4), calculate the fitness values of the n individuals according to Equation (1), and sort them in ascending order;

[0027] (4) If gen meets the maximum iteration condition, output the optimal job path and the cross-region job length of N job points, end the loop, otherwise, go to the next step;

[0028] (5) Select whether to perform a shock adaptive perturbation search strategy on the individual with the smallest fitness value. The calculation formula for the selection probability is as follows:

[0029]

[0030] In the formula, t is the current iteration number, and Maxgen is the maximum iteration number of the algorithm;

[0031] Randomly generate a random number between [0, 1]. If the random number is less than Pr1, perform a shock adaptive perturbation search strategy on the optimal individual, otherwise, go to the next step;

[0032] (6) Select to perform a random strategy or a dynamic movement strategy. The calculation formula for the selection probability is as follows:

[0033]

[0034] In the formula, t is the current iteration number, and Maxgen is the maximum iteration number of the algorithm;

[0035] Randomly generate a random number between [0, 1]. If the random number is greater than Pr2, perform a dynamic movement strategy on the population individuals, otherwise, execute the random strategy;

[0036] (7) Perform a local search operation

[0037] Randomly generate a random number m between [0, 1]. If m < P, execute the 2-Opt algorithm; otherwise, execute an adaptive combined perturbation; if the adaptive combined perturbation strategy is executed, then randomly generate a random number μ between [0, 1]. If μ < P s , execute the double bridge; otherwise, execute the neighborhood 2-opt operator. The calculation formulas for P and P s are as follows:

[0038]

[0039] In the formula, t is the current running time, t max is the maximum running time of the example, P min is the minimum selection probability, P max is the maximum selection probability, P min = 0.25, P max = 0.7;

[0040] (8) Update the optimal individual and population, and return to (4).

[0041] The technical problem to be solved by the present invention can also be further realized by the following technical solutions. For the above-mentioned method for optimizing the cross-regional operation path of a harvester based on an intelligent optimization algorithm, in step (5), the update formula of the oscillating adaptive perturbation search strategy is as follows:

[0042]

[0043]

[0044] In the formula, Xb(t) is the optimal individual, t represents the current iteration number, N represents the number of harvester operation points, q is an N-dimensional 0-1 row vector randomly generated, Maxgen is the maximum iteration number of the algorithm, β is the perturbation term, and λ is the adaptive change amount. denotes the multiplication of elements at the same positions of two vectors.

[0045] The technical problem to be solved by the present invention can also be further realized by the following technical solutions. For the above-mentioned method for optimizing the cross-regional operation path of a harvester based on an intelligent optimization algorithm, in step (6), the update formula of the dynamic movement strategy is as follows:

[0046]

[0047] In the formula, N is all the operation points of the harvester, and Xm r1(t - 1) and Xm r2(t - 1) are the optimal positions of two randomly selected emperor penguins.

[0048] The technical problem to be solved by the present invention can also be further realized by the following technical solutions. For the above-mentioned method for optimizing the cross-regional operation path of a harvester based on an intelligent optimization algorithm, in step (6), the update formula of the random strategy is as follows:

[0049] X i (t) = X i (t - 1) + A i (t) (15)

[0050]

[0051] In the formula, Xr(t - 1) is an individual randomly selected from the population, and Xm i(t - 1) is the optimal position in the memory of the i-th individual Xi(t - 1) in the population.

[0052] The technical problem to be solved by the present invention can also be further realized by the following technical solution. For the above-mentioned harvester cross-regional operation path optimization method based on the intelligent optimization algorithm, in step (7), the specific steps of the neighborhood 2-opt operator are as follows:

[0053] Taking a certain circuit X = (x1, x2,..., x N ) in the population as an example, N is the number of operation points,

[0054] Step 1: Randomly select an operation point a from N operation points. Taking a as the center and a circle with a radius of r1 as the neighborhood, denoted as U(a, r1). The calculation formula of r1 is as shown in formula (19):

[0055]

[0056] In the formula, Z is the route length of the optimal individual in the population;

[0057] Step 2: Determine the operation points u included in the neighborhood U(a, r1). If u≥2, randomly select 2 operation points aa1 and aa2 in the neighborhood; if u = 1, let aa1 = a, and then randomly select an operation point from [1, 2,..., a - 1, a + 1,..., N] as aa2;

[0058] Step 3: Find the operation point bb1 adjacent to aa1 and the operation point bb2 adjacent to aa2 in X, and bb1≠aa2, bb2≠aa1. If bb1 = bb2, repeat Step 1 and Step 2 until bb1≠bb2;

[0059] Step 4: Delete the edges (aa1, bb1) and (aa2, bb2), and connect the edges (bb1, aa2) and (aa1, bb2).

[0060] The technical problem to be solved by the present invention can also be further realized by the following technical solution. For the above-mentioned harvester cross-regional operation path optimization method based on the intelligent optimization algorithm, in step (7), the specific steps of the double-bridge movement are as follows:

[0061] Taking the individual X = (x1, x2,..., x N )(N is the number of operation points) in the population as an example,

[0062] Step 1: Generate a random integer a1 in [2, N - 6];

[0063] Step 2: Generate a random integer k0 according to [2 + a1, N - 4];

[0064] Step 3: Generate a random integer k1 according to [2 + k0, N - 2];

[0065] Step 4: Generate a random integer k2 in the range of [2 + k1, N];

[0066] Step 5: Sort k0, k1, and k2 in ascending order, and let a2 = min(k0, k1, k2), a3 = median(k0, k1, k2), and a4 = max(k0, k1, k2);

[0067] Step 6: Let b1 = a1 - 1, b2 = a2 - 1, b3 = a3 - 1, and b4 = a4 - 1;

[0068] Step 7: a1, b1, a2, b2, a3, b3, a4, and b4 are the index numbers of each component in individual X. Assume that the operation points corresponding to the index numbers a1, b1, a2, b2, a3, b3, a4, and b4 in X are aa1, bb1, aa2, bb2, aa3, bb3, aa4, and bb4 respectively;

[0069] Step 8: Delete the edges (aa1, bb1), (aa2, bb2), (aa3, bb3), and (aa4, bb4), and reconnect the edges (aa1, bb3), (aa2, bb4), (aa3, bb1), and (aa4, bb2).

[0070] The technical problem to be solved by the present invention can also be further realized through the following technical solution. For the above-mentioned method for optimizing the cross-region operation path of a harvester based on an intelligent optimization algorithm, in step (7), the 2-Opt algorithm is a local optimization algorithm. Specifically, for each route in the population, two non-adjacent edges in the given route are sequentially exchanged to obtain a set of all paths, and the exchanges that can improve the path are retained.

[0071] Compared with the prior art, the present invention can provide high-quality solutions in practical applications to meet the needs of different users and scenarios. Its beneficial effects are as follows:

[0072] 1. Balance between global and local search capabilities: The algorithm proposed by the present invention combines the advantages of global search and local search, ensuring that in the process of finding the optimal solution, it can widely explore the solution space and deeply search for better solutions in a certain area;

[0073] 2. High convergence speed: Compared with other algorithms, the algorithm of the present invention shows a faster convergence speed in most cases, which means that in a shorter time, the algorithm can find better or near-optimal solutions;

[0074] 3. Strong adaptability: By adjusting parameters in an adaptive manner, the algorithm can show good performance in different problem instances and different situations. Description of the Drawings

[0075] Figure 1 It is a flow chart of an algorithm of the present invention;

[0076] Figure 2 It is a schematic diagram of the method of the neighborhood 2-opt operator of the present invention;

[0077] Figure 3 It is a schematic diagram of the method of the double-bridge movement of the present invention. Specific implementation manners

[0078] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0079] Refer to Figure 1 , for an optimization method for the cross-regional operation path of a harvester based on an intelligent optimization algorithm, the steps are as follows:

[0080] (1) Parameter setting:

[0081] Based on the hybrid improved real-coded emperor penguin optimization algorithm, the population size is n, the number of iterations of the algorithm is 200, and the initial number of iterations gen = 0; the number of cross-regional operation points is N; for the problem of planning the cross-regional operation path of a harvester with N operation point positions

[0082] The objective function is expressed as:

[0083]

[0084] In the formula, f(D) is the shortest distance of the cross-regional operation of the harvester; t i represents the i-th operation point; d(t i , t i+1 ) represents the distance between the i-th operation point and the (i + 1)-th operation point, and is expressed by the following formula:

[0085]

[0086] (x i , y i ) and (x i+1 , y i+1 ) are the coordinates of t i , t i+1 respectively;

[0087] (2) Encoding method:

[0088] The real-coded emperor penguin optimization algorithm based on hybrid improvement adopts a real-coded method; if the number of working points for the harvester's cross-regional operation is N, then the initial individual is

[0089] X = N * rand(1, N) (3)

[0090] where X is an individual with a code length of N and a floating-point encoded value;

[0091] Generate n individuals according to Equation (3);

[0092] (3) Decoding method:

[0093] Taking the kth individual X in the population k =(x1, x2, …, x N ) as an example, the specific decoding steps are as follows:

[0094] 1) Randomly generate an integer μ between [1, N], and use μ as the starting working point;

[0095] 2) After deleting the μth component in X k , the resulting vector is X'k, X'k = (x1, x2, …, x μ-1 , x μ+1 , …, x N ), and according to the distances d = (d μ,1 , d μ,2 , …, d μ,μ-1 , d μ,μ+1 , …, d μ,N ) between μ and the remaining working points, calculate the vector Tp related to the distances between working points and the gene position values. The calculation formula for Tp is

[0096]

[0097] In the formula, represents the multiplication of the elements at the same positions of vector d and vector ;

[0098] 3) Find the minimum value in vector Tp, and take the working point i corresponding to the minimum value as the next working point;

[0099] 4) Let μ = i, and repeat Step 2 and Step 3 to determine the access order of all working points.

[0100] To facilitate the understanding of the solution method based on distance and position value, take N = 5 as an example for illustration. Let μ = 2, X k =(1.3575, 4.5155, 3.4863, 0.7845, 0.0637), and the distances between 5 working points are shown in Table 1. The specific decoding steps are shown in Table 2.

[0101] Table 1 Distances between Five Working Points

[0102]

[0103] Table 2 Decoding Examples

[0104]

[0105] Decode n individuals according to formula (4), calculate the fitness values of the n individuals according to formula (1), and sort them in ascending order;

[0106] (4) If gen meets the maximum iteration condition, output the optimal operation path and the cross-region operation length of N working points, end the loop, otherwise, go to the next step;

[0107] (5) Select whether to perform an oscillating adaptive perturbation search strategy on the individual with the smallest fitness value. The calculation formula for the selection probability is as follows:

[0108]

[0109] In the formula, t is the current iteration number, and Maxgen is the maximum iteration number of the algorithm;

[0110] Generate a random number between [0, 1]. If the random number is less than Pr1, perform an oscillating adaptive perturbation search strategy on the optimal individual, otherwise, go to the next step;

[0111] Among them, the update formula for the oscillating adaptive perturbation search strategy is as follows:

[0112]

[0113] In the formula, Xb(t) is the optimal individual, t represents the current iteration number, N represents the number of working points of the harvester, q is an N-dimensional 0-1 row vector randomly generated, Maxgen is the maximum iteration number of the algorithm, β is the perturbation term, and λ is the adaptive change amount. represents the multiplication of elements at the same positions of two vectors;

[0114] The oscillating adaptive perturbation search strategy uses a perturbation term that adaptively changes with the iteration number to update the optimal individual. The implementation method is simple, easy to understand, has a wide application range, has strong optimization ability and the ability to jump out of local optimal solutions, helps to reduce the time complexity of the algorithm, and improves the performance of the algorithm.

[0115] (6) Select to perform a random strategy or a dynamic movement strategy. The calculation formula for the selection probability is as follows:

[0116]

[0117] Where t is the current iteration number and Maxgen is the maximum iteration number of the algorithm.

[0118] Generate a random number between [0, 1]. If the random number is greater than Pr2, implement the dynamic movement strategy on the population individuals; otherwise, execute the random strategy.

[0119] Among them, the update formula of the dynamic movement strategy is as follows:

[0120]

[0121]

[0122] In the formula, N is all the working points of the harvester, and Xm r1(t - 1) and Xm r2(t - 1) are the optimal positions of two randomly selected emperor penguins.

[0123] The dynamic movement strategy is a flexible strategy that can dynamically adjust the search behavior of individuals according to the current state of the population. By combining local search and global search, the algorithm aims to effectively traverse the complex multi-dimensional search space while avoiding premature aggregation near the local optimal solution, thereby increasing the probability of finding the global optimal solution.

[0124] The update formula of the random strategy is as follows:

[0125] X i (t) = X i (t - 1) + A i (t) (13)

[0126]

[0127] In the formula, Xr(t - 1) is an individual randomly selected from the population, and Xm i(t - 1) is the optimal position in the memory of the i-th individual Xi(t - 1) in the population.

[0128] The random movement strategy combines the current position, the position of a randomly selected individual, and the historical optimal position in the individual's memory. By increasing the guidance of the randomly selected individual position to the individual, the individuals in the population have the ability to search in a directed random direction in the solution space, which helps the algorithm to conduct a more extensive exploration in the solution space. By dynamically adjusting the weight factors a1 and a2, the algorithm can adjust the behavior of individuals according to the current state of the population and the characteristics of the solution space, aiming to increase the probability of finding the global optimal solution. In summary, the random movement strategy can enhance the global search ability of the algorithm.

[0129] (7) Execute the local search operation to improve the convergence speed and solution quality of the algorithm.

[0130] Randomly generate a random number \(m\) in the range \([0, 1]\). If \(m < P\), execute the 2 - Opt algorithm; otherwise, execute the adaptive combined perturbation. If the adaptive combined perturbation strategy is executed, then randomly generate a random number \(\mu\) in the range \([0, 1]\). If \(\mu < P\) s , execute the double - bridge; otherwise, execute the neighborhood 2 - opt operator. The calculation formulas of \(P\) and \(P\) s are as follows.

[0131]

[0132] In the formula, \(t\) is the current running time, and \(t\) max is the maximum running time of the instance, \(P\) min is the minimum selection probability, and \(P\) max is the maximum selection probability, \(P\) min = 0.25, and \(P\) max = 0.7;

[0133] 1) Neighborhood 2 - opt operator

[0134] To facilitate the description of the neighborhood 2 - opt move, take a certain circuit \(X=(x_1,x_2,\cdots,x\) N ) in the population as an example, where \(N\) is the number of job points. The specific steps of the neighborhood 2 - opt operator are as follows:

[0135] Step 1: Randomly select a job point \(a\) from \(N\) job points. Take the circle with \(a\) as the center and radius \(r_1\) as the neighborhood, denoted as \(U(a,r_1)\). The calculation formula of \(r_1\) is as shown in Equation (19):

[0136]

[0137] In the formula, \(Z\) is the route length of the optimal individual in the population;

[0138] Step 2: Determine the job points \(u\) included in the neighborhood \(U(a,r_1)\). If \(u\geq2\), randomly select 2 job points \(aa_1\) and \(aa_2\) in the neighborhood; if \(u = 1\), let \(aa_1=a\), and then randomly select a job point from \([1,2,\cdots,a - 1,a + 1,\cdots,N]\) as \(aa_2\);

[0139] Step 3: Find the job point \(bb_1\) adjacent to \(aa_1\) and the job point \(bb_2\) adjacent to \(aa_2\) in \(X\), and \(bb_1\neq aa_2\), \(bb_2\neq aa_1\). If \(bb_1 = bb_2\), repeat Step 1 and Step 2 until \(bb_1\neq bb_2\);

[0140] Step 4: Delete the edges \((aa_1,bb_1)\) and \((aa_2,bb_2)\), and connect the edges \((bb_1,aa_2)\) and \((aa_1,bb_2)\).

[0141] To facilitate the description of the neighborhood 2-opt move, the job paths with geometric crossover edges and the job paths without geometric crossover edges in the population individuals are taken as examples for illustration respectively;

[0142] Randomly select an individual X=(1, 2, 10, 9, 8, 7, 6, 5, 4, 3) containing geometric crossover edges from the population. Suppose the job point randomly selected from X is 3. The job points included in the neighborhood with 3 as the center and r1 as the radius are 2 and 3. Then aa1 = 2, aa2 = 3. According to step 3, bb1 = 10, bb2 = 11. Delete the edges (aa1, bb1)=(2, 10), (aa2, bb2)=(3, 11), and connect the edges (bb1, aa2)=(10, 11) and (aa1, bb2)=(2, 3). The individual X and the route obtained after passing through the neighborhood 2-opt operator are as shown in Figure 2 a) in Figure 2 and Figure 2 b) in Figure 2 ; in

[0143] a) in Figure 2 and Figure 2 b) in Figure 2 d(2, 10)+d(3, 11)=96, d(2, 3)+d(10, 11)=41. The path length after reconnecting is better than the path length before perturbation. The individual generated by the neighborhood 2-opt operator is better than the individual X. Figure 2 Randomly select an individual X'=(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11) without geometric crossover edges from the population. Suppose the job point randomly selected from X' is 6. The job points included in the neighborhood with 6 as the center and r as the radius are 5, 6, 7, and 8. The two job points randomly selected from the neighborhood are 6 and 7 respectively. Then aa1 = 6, aa2 = 7. According to step 3, bb1 = 5, bb2 = 8. Delete the edges (aa1, bb1)=(6, 5), (aa2, bb2)=(7, 8), and connect the edges (bb1, aa2)=(5, 7) and (aa1, bb2)=(6, 8). The individual X and the route obtained after passing through the neighborhood 2-opt operator are as shown in

[0144] Note: In Figure 2 , the numbers in the circles are job points, the numbers on the line connecting two circles are the distances between two job points, and the dashed line between two circles represents the path after reconnecting.

[0145] 2) Double - bridge movement

[0146] The double - bridge movement can change the overall shape of the loop and reduce the probability that the algorithm falls into a local optimum. Taking the individual X=(x1, x2, …, x N )(N is the number of job points) in the population as an example, the specific steps of the double - bridge movement are described as follows.

[0147] Step 1: Generate a random integer a1 in the range of [2, N - 6];

[0148] Step 2: Generate a random integer k0 in the range of [2 + a1, N - 4];

[0149] Step 3: Generate a random integer k1 in the range of [2 + k0, N - 2];

[0150] Step 4: Generate a random integer k2 in the range of [2 + k1, N];

[0151] Step 5: Sort k0, k1, k2 in ascending order, and let a2 = min(k0, k1, k2), a3 = median(k0, k1, k2), a4 = max(k0, k1, k2);

[0152] Step 6: Let b1 = a1 - 1, b2 = a2 - 1, b3 = a3 - 1, b4 = a4 - 1;

[0153] Step 7: a1, b1, a2, b2, a3, b3, a4, and b4 are the index numbers of each component in the individual X. Suppose the job points corresponding to the index numbers a1, b1, a2, b2, a3, b3, a4, and b4 in X are aa1, bb1, aa2, bb2, aa3, bb3, aa4, and bb4 respectively.

[0154] Step 8: Delete the edges (aa1, bb1), (aa2, bb2), (aa3, bb3), (aa4, bb4), and reconnect the edges (aa1, bb3), (aa2, bb4), (aa3, bb1), and (aa4, bb2).

[0155] For the convenience of describing the specific steps of the double - bridge movement, taking the individual X=(3, 2, 1, 6, 5, 4, 7, 10, 9, 11, 8) without geometric cross - edges and the individual X'=(1, 10, 7, 4, 5, 11, 9, 2, 3, 8, 6) with geometric cross - edges as examples. The routes represented by X and X' are as shown in Figure 3 a) in Figure 3 c) in

[0156] Perform a double-bridge move on route X. Suppose the randomly selected 4 job points are aa1 = 1, aa2 = 5, aa3 = 9, aa4 = 8, then bb1 = 2, bb2 = 6, bb3 = 10, bb4 = 11. Delete the geometric crossing edges (1,2), (5,6), (9,10) and (8,11), and after connecting (1,10), (5,11), (9,2) and (8,6), a new route is obtained as shown in Figure 3 b) in

[0157] Perform a double-bridge move on route X'. aa1 = 1, aa2 = 5, aa3 = 9, aa4 = 8, then bb1 = 10, bb2 = 11, bb3 = 2, bb4 = 6. Delete the geometric crossing edges (1,10), (5,11), (9,2) and (8,6), and after connecting (1,2), (5,6), (9,10) and (8,11), a new route is obtained as shown in Figure 3 d) in

[0158] Note: In Figure 3 the job points in the circles are job points, the numbers on the line connecting two circles are the distances between two job points, and the dotted line between two circles represents the path after reconnection.

[0159] 3) 2-Opt algorithm

[0160] The 2-Opt algorithm is a local optimization algorithm. Its main idea is that for each route in the population, two non-adjacent edges in the given route are exchanged in turn to obtain a set of all paths, and the exchanges that can improve the path are retained. The 2-Opt algorithm can quickly eliminate the geometric crossing edges existing in each path, improve the quality of the solution and the convergence speed. At the initial stage of algorithm iteration, the possibility of geometric crossing edges existing in the path is large. As the number of iterations increases, the geometric crossing edges will become fewer and fewer, and the quality of the solution will become better and better.

[0161] (8) Update the optimal individual and the population, and return to (4).

[0162] The present invention is a method for optimizing the cross-region operation path of a harvester based on an intelligent optimization algorithm, and its advantages are as follows:

[0163] 1. Balance of global and local search capabilities: The algorithm proposed by the present invention combines the advantages of global search and local search, ensuring that in the process of finding the optimal solution, it can widely explore the solution space and deeply search for better solutions in a certain area;

[0164] 2. High convergence speed: Compared with other algorithms, the algorithm of the present invention shows a faster convergence speed in most cases, which means that in a shorter time, the algorithm can find a better or near-optimal solution;

[0165] 3. Strong adaptability: By adjusting parameters in an adaptive manner, the algorithm can exhibit good performance in different problem instances and different scenarios.

[0166] 4. The present invention can provide high-quality solutions in practical applications, meeting the needs of different users and scenarios.

[0167] In summary, the main objective of the present invention is to provide a hybrid improved real-coded emperor penguin optimization algorithm that can effectively solve the problem of optimizing the cross-region operation path of a harvester. This algorithm combines the advantages of multiple intelligent optimization algorithms, aiming to improve the global and local search capabilities of the algorithm, thereby enhancing the solution quality and efficiency for the problem of optimizing the cross-region operation path of a harvester.

Claims

1. A method for optimizing the cross-region operation path of a harvester based on an intelligent optimization algorithm, characterized in that: The steps of this method are as follows: (1) Parameter settings: The population size of the real number coding emperor penguin optimization algorithm based on hybrid improvement is n, the number of iterations of the algorithm is 200, the initial number of iterations gen = 0, the number of cross-region operation points is N, and the objective function of the cross-region operation path planning problem of the harvester with N operation points is expressed as: Where f(D) is the shortest distance for the harvester to operate across regions; t i represents the i-th operation point; d(t i ,t i+1 ) represents the distance between the i-th operating point and the i+1-th operating point, which is expressed by the following formula: (x i ,y i ) and (x i+1 ,y i+1 ) are t i , t i+1 The coordinates of (2) Encoding method: The emperor penguin optimization algorithm based on hybrid improved real number coding adopts real number coding; The number of operation points of the harvester's cross-region operation is N, so the individual initial X=N*rand(1,N) (3) Among them, X is an individual with a code length of N and a code value of a floating point number; Generate n individuals using formula (3); (3)Decoding method: Take the kth individual X in the population k =(x1,x2,…,x N ) as an example, the specific decoding steps are as follows: (3.1) Randomly generate an integer μ between [1, N] and use μ as the starting point; (3.2) Delete X k The vector after the μth component in is X'k, X'k=(x1,x2,…,x μ-1 ,x μ+1 ,…,x N ), according to the distance between μ and other working points d = (d μ,1 ,d μ,2 ,…,d μ,μ-1 ,d μ,μ+1 ,…,d μ,N ), calculate the vector Tp related to the distance between the operation points and the size of the gene position value. The calculation formula of Tp is In the formula, represents the vector d and the vector Multiply elements at the same position; (3.3) Find the minimum value in the vector Tp, and take the operation point i corresponding to the minimum value as the next operation point; (3.4) Let μ = i, repeat Step 2 and Step 3 to determine the visit order of all operating points; Decode n individuals using formula (4), calculate the fitness values ​​of n individuals using formula (1), and sort them in order from small to large; (4) If gen meets the maximum iteration condition, output the optimal operation path and the cross-zone operation length of N operation points, and end the loop; otherwise, proceed to the next step; (5) Choose whether to oscillate the adaptive perturbation search strategy for the individual with the smallest fitness value. The calculation formula for the selection probability is as follows: In the formula, t is the current iteration number, and Maxgen is the maximum iteration number of the algorithm; Generate a random number between [0, 1]. If the random number is less than Pr1, perform the oscillating adaptive perturbation search strategy on the optimal individual. Otherwise, proceed to the next step. (6) Choose to use a random strategy or a dynamic movement strategy. The formula for calculating the selection probability is as follows: In the formula, t is the current iteration number, and Maxgen is the maximum iteration number of the algorithm; A random number is randomly generated between [0, 1]. If the random number is greater than Pr2, a dynamic mobile strategy is used on the individuals in the population. Otherwise, a random strategy is executed. (7) Perform local search operations Randomly generate a random number \(m\) in the interval \([0, 1]\). If \(m < P\), execute the 2 - Opt algorithm; otherwise, execute the adaptive combined perturbation. If the adaptive combined perturbation strategy is executed, then randomly generate a random number \(\mu\) between \([0, 1]\). If \(\mu < P\) s , execute the double - bridge; otherwise, execute the neighborhood 2 - opt operator. The calculation formulas of \(P\) and \(P\) s are as follows: Where t is the current running time, t max is the maximum running time of the example, P min is the minimum selection probability, P max is the maximum selection probability, P min =0.25, P max =0.7; (8) Update the optimal individuals and populations, and return to (4).

2. The method for optimizing the cross-region operation path of a harvester based on an intelligent optimization algorithm according to claim 1, characterized in that: In step (5), the update formula of the oscillating adaptive perturbation search strategy is as follows: In the formula, Xb(t) is the optimal individual, t is the current iteration number, N is the number of harvester operation points, q is a randomly generated N-dimensional 0-1 row vector, Maxgen is the maximum number of iterations of the algorithm, β is the disturbance term, and λ is the adaptive change amount. Represents the multiplication of the elements at the same position of two vectors.

3. The method for optimizing the cross-region operation path of a harvester based on an intelligent optimization algorithm according to claim 1, characterized in that: In step (6), the update formula of the dynamic mobility strategy is as follows: Where N is the total number of operating points of the harvester, and Xm r1(t-1) and Xm r2(t-1) are the optimal positions of two randomly selected emperor penguins.

4. The method for optimizing the cross-region operation path of a harvester based on an intelligent optimization algorithm according to claim 1 or 3, characterized in that: In step (6), the update formula of the random strategy is as follows: X i (t)=X i (t-1)+A i (t) (15) Where Xr(t-1) is an individual randomly selected from the population, and Xmi(t-1) is the optimal position in the memory of the i-th individual Xi(t-1) in the population.

5. The method for optimizing the cross-region operation path of a harvester based on an intelligent optimization algorithm according to claim 1, characterized in that: In step (7), the specific steps of the neighborhood 2-opt operator are as follows: Take a certain circuit X in the population = (x1, x2, ..., x N ) as an example, N is the number of operation points, Step 1: Randomly select a work point a from N work points, take a as the center and a circle with a radius of r1 as the neighborhood, denoted as U(a,r1), and the calculation formula of r1 is as follows: In the formula, Z is the route length of the optimal individual in the population; Step 2: Determine the work point u contained in the neighborhood U(a,r1). If u≥2, randomly select two work points aa1 and aa2 in the neighborhood. If u=1, let aa1=a, and then randomly select a work point in [1,2,…,a-1,a+1,…,N] as aa2. Step 3: Find the operating point bb1 adjacent to aa1 and the operating point bb2 adjacent to aa2 in X, and bb1≠aa2, bb2≠aa1. If bb1=bb2, repeat Step 1 and Step 2 until bb1≠bb2. Step 4: Delete the edges (aa1, bb1) and (aa2, bb2), and connect the edges (bb1, aa2) and (aa1, bb2).

6. The method for optimizing the cross-region operation path of a harvester based on an intelligent optimization algorithm according to claim 1, characterized in that: In step (7), the specific steps of double bridge movement are as follows: Take the individuals in the population X=(x1,x2,…,x N )(N is the number of operation points) as an example, Step 1: Generate a random integer a1 in [2,N-6]; Step 2: Generate a random integer k0 according to [2+a1,N-4]; Step 3: Generate a random integer k1 according to [2+k0,N-2]; Step 4: Generate a random integer k2 according to [2+k1,N]; Step 5: Sort k0, k1, k2 from small to large, let a2 = min(k0, k1, k2), a3 = median(k0, k1, k2), a4 = max(k0, k1, k2); Step 6: Let b1 = a1-1, b2 = a2-1, b3 = a3-1, b4 = a4-1; Step 7: a1, b1, a2, b2, a3, b3, a4 and b4 are the index numbers of each component in individual X. Let the operation points corresponding to the index numbers a1, b1, a2, b2, a3, b3, a4 and b4 in X be aa1, bb1, aa2, bb2, aa3, bb3, aa4 and bb4 respectively; Step 8: Delete the edges (aa1,bb1), (aa2,bb2), (aa3,bb3), (aa4,bb4) and reconnect the edges (aa1,bb3), (aa2,bb4), (aa3,bb1) and (aa4,bb2).

7. The method for optimizing the cross-region operation path of a harvester based on an intelligent optimization algorithm according to claim 1, characterized in that: In step (7), the 2-Opt algorithm is a local optimization algorithm. Specifically, for each route in the population, two non-adjacent edges in the given route are exchanged in turn to obtain a set of all paths, and the exchange that can improve the path is retained.

Images