Mine digital twin system roaming path planning method based on differential evolution algorithm

Abstract

The application discloses a mine digital twin system roaming path planning method based on a differential evolution algorithm, according to a longitudinal section view of the inside of a mine, adopts a grid data model under a specified mine gridding density to represent a map environment, and represents reachable path points in the mine as a gridded mine reachable position matrix through discrete squares; determines an optimization target and a constraint condition of a task; and uses an improved differential evolution algorithm to solve an optimal mine digital twin system roaming path planning scheme, can replace a fixed differential strategy with adaptive selection from four differential strategies with the aid of the characteristics of the differential evolution algorithm, such as strong global convergence ability, high robustness and no need of the aid of characteristic information of a problem, and adds a polynomial mutation as an additional mutation method, improves the ability of the algorithm to jump out of a local optimal solution, and solves the technical problems that the roaming path planning method in the prior art is prone to falling into a local optimal solution and has a slow convergence speed.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of path planning, and relates to a planning method, in particular to a mine digital twin system roaming path planning method based on a differential evolution algorithm. BACKGROUND

[0002] The digital twin system platform focuses on the reproduction, analysis and decision of physical objects in the data world, and can effectively monitor and analyze the mine. In the digital twin system platform, in order to enable users to roam various scenes of the digital twin system, roaming path planning needs to be performed. Path planning refers to roaming all scenes that users need to view under the condition that the starting point and the ending point are given, and ensures multiple optimization targets such as the shortest path and smooth path. Path planning is a key problem that directly affects the scene exploration and immersion of users. In addition, multiple constraint conditions such as path perspective and safe distance between the path and obstacles also need to be considered. These constraint conditions bring certain difficulties to the roaming path planning.

[0003] Most of the existing path planning algorithms are not suitable for the mine digital twin system with a large area and complex road conditions, and have problems such as being easy to fall into a local optimal solution and slow convergence speed. SUMMARY

[0004] In view of the problems in the prior art, the purpose of the present application is to provide a mine digital twin system roaming path planning method based on a differential evolution algorithm, so as to solve the technical problems that the existing roaming path planning method is easy to fall into a local optimal solution and has a slow convergence speed.

[0005] In order to solve the above technical problems, the present application adopts the following technical solutions:

[0006] A mine digital twin system roaming path planning method based on a differential evolution algorithm, specifically comprising the following steps:

[0007] Step 1: According to the longitudinal section view of the mine, a grid data model under a specified mine grid density is used to represent the map environment, and the reachable path points in the mine are represented as a grid mine reachable position matrix through discrete grids.

[0008] Each discrete grid contains numbers 0 and 1, 0 represents reachability, and 1 represents unreachability.

[0009] Step 2: Determine the optimization target and constraint condition of the task.

[0010] The optimization target includes the shortest path and the smooth path.

[0011] The constraint condition includes the path perspective and the safe distance between the path and the obstacle.

[0012] Step 3: Establish the optimization objective evaluation function F(x) for the task;

[0013] st

[0014] in:

[0015] X = (x1, x2, ..., x) T ), where x i Represents the coordinates of the discrete square corresponding to the roaming position at time i;

[0016] T represents the total number of roaming moments;

[0017] f m (x) represents the evaluation function of the m-th optimization objective of this task;

[0018] This represents the target weight for optimization;

[0019] g i (X) represents the i-th inequality constraint related to the path view or the safe distance between the path and obstacles in the roaming path planning problem, where 1≤i≤p and p is the number of inequality constraints.

[0020] h j (X) represents the j-th equality constraint related to the path view or the safe distance between the path and obstacles in the roaming path planning problem, where 1≤j≤q and q is the number of equality constraints.

[0021] Step four involves using an improved differential evolution algorithm to solve for the optimal roaming path planning scheme for the mine digital twin system. This includes the following steps:

[0022] Step 4.1, set initial parameters;

[0023] The algorithm is set to have a maximum function evaluation count of maxFE, an initial evaluation count of FE = 0, a population size of N, a crossover probability of rmp, a difference mutation probability of mu1, a polynomial mutation probability of mu2, and four candidate difference strategies: DE / rand / 1 / bin, DE / rand-to-best / 2 / bin, DE / rand / 2 / bin, and DE / current-to-rand / 1. The initial difference strategy selection probability for each strategy is P. i =0.25;

[0024] Step 4.2: Based on the rasterized mine reachability matrix to be solved, generate an initial population according to the neighborhood method or the random method;

[0025] Step 4.3: Determine whether each individual needs to undergo differential mutation using the differential strategy based on the differential mutation probability mu1. If so, proceed to step 4.4; otherwise, retain the corresponding individual directly.

[0026] Step 4.4: Select the probability P according to the initial difference strategy of the four difference strategies. i Choose the difference strategy for each individual that needs difference variation;

[0027] Step 4.5: Perform differential mutation on each individual that needs differential mutation according to the selected differential mutation strategy to obtain the set of offspring individuals after differential mutation.

[0028] Step 4.6: Merge the individuals retained in Step 4.3 with the set of offspring individuals obtained after differential mutation in Step 4.5 to obtain the offspring population;

[0029] Step 4.7: Determine whether each individual in the offspring population needs to undergo polynomial mutation according to the polynomial mutation probability mu2. If so, proceed to step 4.8; otherwise, directly retain the corresponding individual.

[0030] Step 4.8: Perform polynomial mutation on the individuals that need to be mutated to obtain the set of offspring individuals after polynomial mutation.

[0031] Step 4.9: Merge the individuals directly retained in Step 4.7 with the set of offspring individuals obtained after polynomial mutation in Step 4.8 to obtain a new offspring population;

[0032] Step 4.10: Evaluate the new offspring population obtained in Step 4.9 by optimizing the objective evaluation function F(x), and update the evaluation count FE;

[0033] Step 4.11: Select offspring individuals from the new offspring population obtained in Step 4.9 to enter the next generation population using the individual selection method, and obtain the next generation population;

[0034] Step 4.12: Based on the retention rates of various differential strategies among offspring individuals in the next generation population obtained in Step 4.11, update the differential strategy selection probability P. i ;

[0035] Step 4.13: Determine whether the current number of evaluations FE satisfies FE≥maxFE. If it does, the algorithm ends. Based on the evaluation results, select the best offspring individual in the next generation population obtained in Step 4.11 as the optimal mine digital twin system roaming path planning scheme obtained by the algorithm. If it does not satisfy the condition, return to Step 4.3.

[0036] This invention also includes the following technical features:

[0037] In step one, the grid density of the mine can be determined by the user.

[0038] In step 4.2, the specific process of generating the initial population according to the neighborhood method is as follows:

[0039] Starting from a certain coordinate point, the next coordinate point to be visited is a randomly selected reachable coordinate point from an annulus with radius r as the inner circle and radius r′ as the outer circle.

[0040] In step 4.10, the method for updating the evaluation count FE is as follows:

[0041] FE = FE + the number of times the evaluation function for the optimization objective is called.

[0042] The individual selection method described is either the method of directly selecting the top N optimal individuals or the roulette wheel selection method.

[0043] Compared with the prior art, the beneficial technical effects of this invention are:

[0044] (I) The method of this invention can improve upon the traditional differential evolution algorithm by leveraging its strong global convergence ability, high robustness, and the fact that it does not require the characteristic information of the problem. It replaces the fixed differential strategy with an adaptive selection from four differential strategies and adds polynomial mutation as an additional mutation method, which improves the algorithm's ability to escape local optima. By applying this algorithm to solve the roaming path planning problem of a mine digital twin system, a high-quality roaming path planning scheme for a mine digital twin system is obtained, which solves the technical problems of existing roaming path planning methods being prone to getting trapped in local optima and having slow convergence speed.

[0045] (II) The roaming path planning method for mine digital twin system based on differential evolution algorithm proposed in this invention can model the problem according to the mine structure and adaptively select mutation operators, which effectively improves the practicality of roaming path planning for mine digital twin system. Attached Figure Description

[0046] Figure 1 This is a flowchart illustrating the method of the present invention.

[0047] The specific content of the present invention will be further explained in detail below with reference to the embodiments. Detailed Implementation

[0048] It should be noted that, unless otherwise specified, all components in this invention are those known in the art.

[0049] The following are specific embodiments of the present invention. It should be noted that the present invention is not limited to the following specific embodiments. All equivalent modifications made based on the technical solutions of this application fall within the protection scope of the present invention.

[0050] A method for roaming path planning in a mine digital twin system based on differential evolution algorithm, specifically including the following steps:

[0051] Step 1: Based on the longitudinal section map inside the mine, the map environment is represented by a raster data model with a specified mine raster density, and the reachable path points in the mine are represented by discrete squares as a rasterized mine reachable location matrix.

[0052] Each discrete square contains the numbers 0 and 1, where 0 represents reachable and 1 represents unreachable;

[0053] Step two: Determine the optimization objective and constraints of the task;

[0054] The optimization objectives include finding the shortest path and ensuring a smooth path.

[0055] Constraints include the path viewpoint and the safe distance between the path and obstacles;

[0056] Step 3: Establish the optimization objective evaluation function F(x) for the task;

[0057] st

[0058] in:

[0059] X = (x1, x2, ..., x) T ), where x i Represents the coordinates of the discrete square corresponding to the roaming position at time i;

[0060] T represents the total number of roaming moments;

[0061] f m (x) represents the evaluation function of the m-th optimization objective of this task;

[0062] This represents the target weight for optimization;

[0063] g i (X) represents the i-th inequality constraint related to the path view or the safe distance between the path and obstacles in the roaming path planning problem, where 1≤i≤p and p is the number of inequality constraints.

[0064] h j (X) represents the j-th equality constraint related to the path view or the safe distance between the path and obstacles in the roaming path planning problem, where 1≤j≤q and q is the number of equality constraints.

[0065] Step four involves using an improved differential evolution algorithm to solve for the optimal roaming path planning scheme for the mine digital twin system. This includes the following steps:

[0066] Step 4.1, set initial parameters;

[0067] The algorithm is set to have a maximum function evaluation count of maxFE, an initial evaluation count of FE = 0, a population size of N, a crossover probability of rmp, a difference mutation probability of mu1, a polynomial mutation probability of mu2, and four candidate difference strategies: DE / rand / 1 / bin, DE / rand-to-best / 2 / bin, DE / rand / 2 / bin, and DE / current-to-rand / 1. The initial difference strategy selection probability for each strategy is P. i =0.25;

[0068] Step 4.2: Based on the rasterized mine reachability matrix to be solved, generate an initial population according to the neighborhood method or the random method;

[0069] Step 4.3: Determine whether each individual needs to undergo differential mutation using the differential strategy based on the differential mutation probability mu1. If so, proceed to step 4.4; otherwise, retain the corresponding individual directly.

[0070] Step 4.4: Select the probability P according to the initial difference strategy of the four difference strategies. i Choose the difference strategy for each individual that needs difference variation;

[0071] Step 4.5: Perform differential mutation on each individual that needs differential mutation according to the selected differential mutation strategy to obtain the set of offspring individuals after differential mutation.

[0072] Step 4.6: Merge the individuals retained in Step 4.3 with the set of offspring individuals obtained after differential mutation in Step 4.5 to obtain the offspring population;

[0073] Step 4.7: Determine whether each individual in the offspring population needs to undergo polynomial mutation according to the polynomial mutation probability mu2. If so, proceed to step 4.8; otherwise, directly retain the corresponding individual.

[0074] Step 4.8: Perform polynomial mutation on the individuals that need to be mutated to obtain the set of offspring individuals after polynomial mutation.

[0075] Step 4.9: Merge the individuals directly retained in Step 4.7 with the set of offspring individuals obtained after polynomial mutation in Step 4.8 to obtain a new offspring population;

[0076] Step 4.10: Evaluate the new offspring population obtained in Step 4.9 by optimizing the objective evaluation function F(x), and update the evaluation count FE;

[0077] Step 4.11: Select offspring individuals from the new offspring population obtained in Step 4.9 to enter the next generation population using the individual selection method, and obtain the next generation population;

[0078] Step 4.12: Based on the retention rates of various differential strategies among offspring individuals in the next generation population obtained in Step 4.11, update the differential strategy selection probability P. i ;

[0079] Step 4.13: Determine whether the current number of evaluations FE satisfies FE≥maxFE. If it does, the algorithm ends. Based on the evaluation results, select the best offspring individual in the next generation population obtained in Step 4.11 as the optimal mine digital twin system roaming path planning scheme obtained by the algorithm. If it does not satisfy the condition, return to Step 4.3.

[0080] In the above technical solution, the strong global convergence ability, high robustness, and lack of need for problem feature information of the differential evolution algorithm can be leveraged to improve the traditional differential evolution algorithm. The fixed differential strategy is replaced with an adaptive selection from four differential strategies. At the same time, polynomial mutation is added as an additional mutation method to improve the algorithm's ability to escape local optima. By applying this algorithm to solve the roaming path planning problem of the mine digital twin system, a high-quality roaming path planning scheme for the mine digital twin system is obtained, which solves the technical problems of existing roaming path planning methods being prone to getting trapped in local optima and having slow convergence speed.

[0081] Path smoothing refers to avoiding frequent path changes during roaming, which can negatively impact the roaming experience.

[0082] Furthermore, the roaming path planning method for mine digital twin systems based on differential evolution algorithm can model the problem according to the mine structure and adaptively select mutation operators, effectively improving the practicality of roaming path planning for mine digital twin systems.

[0083] Based on the retention rates of various differential strategies among the offspring individuals in the next generation population obtained in step 4.11, update the differential strategy selection probability P. i The method is as follows:

[0084] Assuming that among the retained individuals, the DE / rand / 1 / bin differencing strategy is the most frequent and the DE / current-to-rand / 1 differencing strategy is the least frequent, then for P... iThe update is performed to update the selection probabilities of the best and worst performing difference strategies respectively, and then the individual selection probabilities are adjusted proportionally so that the sum of the total probabilities is 1.

[0085]

[0086]

[0087] in: A user-defined parameter greater than 1;

[0088] A user-defined parameter less than 1;

[0089] In step one, the grid density of the mine can be determined by the user.

[0090] In step 4.2, the specific process of generating the initial population according to the neighborhood method is as follows:

[0091] Starting from a certain coordinate point, the next coordinate point to be visited is a randomly selected reachable coordinate point from an annulus with radius r as the inner circle and radius r′ as the outer circle.

[0092] Preferably, the radius r is 1 and r′ is 4.

[0093] In step 4.10, the method for updating the evaluation count FE is as follows:

[0094] FE = FE + the number of times the evaluation function for the optimization objective is called.

[0095] Individual selection method is either the method of directly selecting the top N optimal ones or the roulette wheel selection method.

Claims

1. A method for roaming path planning in a mine digital twin system based on differential evolution algorithm, characterized in that, Specifically, the following steps are included: Step 1: Based on the longitudinal section map inside the mine, the map environment is represented by a raster data model with a specified mine raster density, and the reachable path points in the mine are represented by discrete squares as a rasterized mine reachable location matrix. Each discrete square contains the numbers 0 and 1, where 0 represents reachable and 1 represents unreachable; Step two: Determine the optimization objective and constraints of the task; The optimization objectives include shortest path and smoothest path; The constraints include the path viewpoint and the safe distance between the path and obstacles; Step 3: Establish the optimization objective evaluation function F(x) for the task; s.t.g i ()≤0,i=1,2,…p, h j ()=0,j=1,2,…q, in: X = (x1, x2, ..., x) T ), where x i Represents the coordinates of the discrete square corresponding to the roaming position at time i; T represents the total number of roaming moments; f m (x) represents the evaluation function of the m-th optimization objective of this task; This represents the target weight for optimization; g i (X) represents the i-th inequality constraint related to the path view or the safe distance between the path and obstacles in the roaming path planning problem, where 1≤i≤p and p is the number of inequality constraints. h j (X) represents the j-th equality constraint related to the path view or the safe distance between the path and obstacles in the roaming path planning problem, where 1≤j≤q and q is the number of equality constraints. Step four involves using an improved differential evolution algorithm to solve for the optimal roaming path planning scheme for the mine digital twin system. This includes the following steps: Step 4.1, set initial parameters; The algorithm is set to have a maximum function evaluation count of maxFE, an initial evaluation count of FE = 0, a population size of N, a crossover probability of rmp, a difference mutation probability of mu1, a polynomial mutation probability of mu2, and four candidate difference strategies: DE / rand / 1 / bin, DE / rand-to-best / 2 / bin, DE / rand / 2 / bin, and DE / current-to-rand / 1. The initial difference strategy selection probability for each strategy is P. i =0.25; Step 4.2: Based on the rasterized mine reachability matrix to be solved, generate an initial population according to the neighborhood method or the random method; Step 4.3: Determine whether each individual needs to undergo differential mutation using the differential strategy based on the differential mutation probability mu1. If so, proceed to step 4.4; otherwise, retain the corresponding individual directly. Step 4.4: Select the probability P according to the initial difference strategy of the four difference strategies. i Choose the difference strategy for each individual that needs difference variation; Step 4.5: Perform differential mutation on each individual that needs differential mutation according to the selected differential mutation strategy to obtain the set of offspring individuals after differential mutation. Step 4.6: Merge the individuals retained in Step 4.3 with the set of offspring individuals obtained after differential mutation in Step 4.5 to obtain the offspring population; Step 4.7: Determine whether each individual in the offspring population needs to undergo polynomial mutation according to the polynomial mutation probability mu2. If so, proceed to step 4.8; otherwise, directly retain the corresponding individual. Step 4.8: Perform polynomial mutation on the individuals that need to be mutated to obtain the set of offspring individuals after polynomial mutation. Step 4.9: Merge the individuals directly retained in Step 4.7 with the set of offspring individuals obtained after polynomial mutation in Step 4.8 to obtain a new offspring population; Step 4.10: Evaluate the new offspring population obtained in Step 4.9 by optimizing the objective evaluation function F(x), and update the evaluation count FE; Step 4.11: Select offspring individuals from the new offspring population obtained in Step 4.9 to enter the next generation population using the individual selection method, and obtain the next generation population; Step 4.12: Based on the retention rates of various differential strategies among offspring individuals in the next generation population obtained in Step 4.11, update the differential strategy selection probability P. i ; Step 4.13: Determine whether the current number of evaluations FE satisfies FE≥maxFE. If it does, the algorithm ends. Based on the evaluation results, select the best offspring individual in the next generation population obtained in Step 4.11 as the optimal mine digital twin system roaming path planning scheme obtained by the algorithm. If it does not satisfy the condition, return to Step 4.

3.

2. The roaming path planning method for a mine digital twin system based on differential evolution algorithm as described in claim 1, characterized in that, In step 4.2, the specific process of generating the initial population according to the neighborhood method is as follows: Starting from a certain coordinate point, the next coordinate point to be visited is a randomly selected reachable coordinate point from an annulus with radius r as the inner circle and radius r′ as the outer circle.

3. The roaming path planning method for a mine digital twin system based on differential evolution algorithm as described in claim 1, characterized in that, In step 4.10, the method for updating the evaluation count FE is as follows: FE = FE + the number of times the evaluation function for the optimization objective is called.

4. The roaming path planning method for a mine digital twin system based on differential evolution algorithm as described in claim 1, characterized in that, The individual selection method described is either the method of directly selecting the top N optimal individuals or the roulette wheel selection method.

Images