Strip mine transportation scheduling model optimization method, system, equipment and medium considering large key equipment fault

Abstract

The invention discloses a strip mine transportation scheduling model optimization method, system, equipment and medium considering a large key equipment fault, and the method comprises the steps: building an M / M / m queuing model according to a queuing theory principle when the mine large key equipment has a fault, taking the queuing length of a whole transportation system truck as an optimization index, solving the model through employing an iteration method, and solving the model according to the optimization index; determining the optimal number of trucks capable of meeting the production requirements during the equipment failure period; an available truck set is determined again, and a rescheduling model with the maximum yield in the fault time, the minimum idle time of the electric shovel or the crusher and the minimum grade fluctuation deviation is constructed; according to the characteristics of the established model, a differential evolution algorithm is improved by adopting a two-stage optimization mechanism, and a multi-objective solving algorithm is designed on the basis of the improved differential evolution algorithm; solving the rescheduling model to obtain an optimal scheduling scheme; the system, the equipment and the medium are used for implementing the method. The operation continuity of the mine transportation system is guaranteed, the transportation efficiency is improved, and the transportation cost is reduced.

Description

Technical Field

[0001] The present invention belongs to the field of mining system engineering and mine optimization technology, and particularly relates to an open-pit mine transportation scheduling model optimization method, system, equipment and medium taking into account large-scale key equipment failures. Background Art

[0002] In recent years, with the continuous deepening of the application of emerging technologies such as artificial intelligence, big data, the Internet of Things, and cloud computing in the mining field, the global mining industry is undergoing a new technological revolution. As an important way to mine mineral resources in my country, open-pit mining accounts for about 50% of the total cost of production and operation. Therefore, the reasonable scheduling of transportation equipment is a direct way to improve the economic benefits of enterprises, save transportation costs, and reduce energy consumption. However, in the actual production of mines, some dynamic disturbance events often affect the scheduling plans made in advance, among which the impact of equipment failure is the most typical. In large open-pit mines, electric shovels and crushers are key equipment in the entire production and transportation link. Once a failure occurs, if the scheduling plan is not corrected in time, it may cause chaos in truck scheduling, delay working hours, and waste of transportation resources, thereby affecting the progress of mining. Therefore, ensuring the continuity of mine production and the rational use of transportation resources to achieve cost reduction and efficiency improvement in production operations are issues that every mining enterprise pays close attention to and urgently needs to solve.

[0003] The existing related technologies do not have in-depth research on scheduling considering failures of key mine equipment.

[0004] In the literature [LiN, Feng S, Lei T, et al. Rescheduling Plan Optimization of Underground Mine Haulage Equipment Based on Random Breakdown Simulation [J]. Sustainability, 2022, 14 (6): 3448.], Li et al. [1] established a transportation equipment scheduling model based on random simulation of equipment failure. The model aims to achieve the maximum completion rate of the original mining plan and the minimum fluctuation of ore grade during rescheduling. By improving the wolf pack algorithm, the position update formula of individuals in the wolf pack is changed for optimization. Then, the optimal model solution is used to optimize the rescheduling of the transportation plan.

[0005] In the literature [Zhang Jun, Fan Yongkai, Li Dongqi. Research and Application of Intelligent Truck Dispatching System in Open-pit Mines [J]. Modern Mining, 2022, 38(09): 184-188+199.], Zhang Jun et al. [2] established a mathematical model for mine traffic planning based on a new generation of wireless communication technology, high-precision positioning technology, intelligent vehicle-mounted terminal equipment and linear programming theory. Based on the principle of maximizing production under the premise of a certain number of vehicles, the traffic calculation on each optimal path was completed, thereby achieving the goal of efficient and free dispatch of open-pit trucks.

[0006] The current research on open-pit mine truck dispatching systems is diversified and involves various angles. However, there are few related technologies that consider the equipment. In actual mine production, equipment failures occur from time to time, so it is of great significance to study how to ensure the orderly production in the event of key equipment failures. In addition, the description of the scene should be in line with the actual situation. When the equipment fails, the number of trucks in the entire dispatching system will be redundant, so the number of trucks must be re-determined to ensure the effective use of resources. Summary of the invention

[0007] In order to overcome the deficiencies of the above-mentioned prior art, the purpose of the present invention is to provide an open-pit mine transportation scheduling model optimization method, system, equipment and medium considering the failure of large key equipment, using the principle of queuing theory to ensure the rational use of truck resources with the leader of the transportation system truck team as an indicator, and establish a rescheduling model with the maximum output during the failure time, the minimum idle time of the electric shovel or crusher and the minimum deviation of grade fluctuation, and design a multi-objective evolutionary algorithm to solve the rescheduling model, so as to ensure the orderly production in the event of equipment failure. The present invention has the characteristics of realizing the anti-disturbance and orderliness of the mine transportation scheduling system.

[0008] In order to achieve the above object, the technical solution adopted by the present invention is:

[0009] An open-pit mine transportation scheduling model optimization method considering large key equipment failures includes the following steps:

[0010] Step 1: When a large key equipment in a mine fails, an M / M / m queuing model is constructed based on the principles of queuing theory. The length of the truck queue in the entire transportation system is used as the optimization indicator, and the M / M / m queuing model is solved by an iterative method to determine the optimal number of trucks that can meet production needs during the equipment failure.

[0011] Step 2: Based on the optimal number of trucks obtained in step 1, the available truck set is re-determined, and a rescheduling model is constructed to maximize the output during the failure time, minimize the idle time of the shovel or crusher, and minimize the grade fluctuation deviation;

[0012] Step 3: According to the characteristics of the rescheduling model established in step 2, a two-stage optimization mechanism is used to improve the differential evolution algorithm, and a multi-objective solution algorithm is designed based on this;

[0013] Step 4: Solve the rescheduling model obtained in step 2 according to the multi-objective solution algorithm designed in step 3 to obtain the optimal scheduling plan for the fleet within a working cycle.

[0014] In the M / M / m queue model of step 1, M represents that the arrival time of trucks and the loading and unloading time follow negative exponential distribution, and m represents the number of service stations, i.e., electric shovels and crushers;

[0015] Taking the queue length of the trucks in the transportation system as the optimization index of the entire transportation system, the formula is as follows:

[0016]

[0017] Where K is the total number of trucks in the transportation system after the failure of the key equipment; i represents the service point, i = 1…N; λ i represents the flow at the i-th service point, 0≤λ i ≤μ i m i , μ i represents the service rate of each service station at the i-th service point; m i is the number of identical service stations contained in the i-th service point; ρ i is the service intensity of the ith service point, that is, the probability that the service station (electric shovel, crusher) is busy, P mi is the probability that all servers in the system are occupied; a i is the correction factor; f i represents the queue leader at the i-th service point;

[0018]

[0019] Among them, C Bi represents the coefficient of variation of truck loading and unloading times;

[0020] In step 1, an iterative method is used to solve the M / M / m queue model. The specific steps are as follows:

[0021] Step 1.1, initialization, take λ l =0 is the lower limit of flow rate, λ u =min{μ i m i} is the upper limit of traffic flow;

[0022] Step 1.2, let Substitution , calculate the size of trucks in the transportation system;

[0023] Step 1.3, take ε as the precision. If K-ε≤g(λ)≤K+ε, stop the iteration and get the required flow λ. If g(λ)>K, let λ u =λ, if g(λ)<K, let λ l =λ, and repeat step 1.2.

[0024] The rescheduling model in step 2 includes model objectives and model constraints, which are expressed as the following formula:

[0025] The model objectives include:

[0026] (1) Maximum output

[0027]

[0028] Among them, f1 represents the output, C r represents the truck's carrying capacity, x rij represents the number of trips made by the r-th truck between the i-th loading point and the j-th unloading point, N represents the set of loading points, M represents the set of unloading points, and R represents the set of available trucks;

[0029] (2) Minimum idle time for electric shovel or crusher

[0030]

[0031] Where f2 represents the idle time of the shovel or crusher, T represents the working time of a shift, and T z Indicates the loading time, T x Indicates the uninstallation time, x rij represents the number of trips made by the r-th truck between the i-th loading point and the j-th unloading point, T y1 With T y2 They represent the non-working time of the shovel or crusher due to rest and maintenance respectively; x1 and x2 are {0,1} decision variables, and when they are 0, they represent the failure of the shovel or crusher respectively;

[0032] (3) Minimum grade fluctuation deviation

[0033]

[0034] Among them, p y Indicates the original grade deviation, g i represents the ore grade at the i-th loading point, G j represents the target grade of the jth unloading point, K represents the set of available trucks;

[0035] The model constraints include:

[0036] (4) Production constraints at loading points

[0037]

[0038] g i represents the total load of the i-th loading point;

[0039] (5) Output constraints at unloading points

[0040]

[0041] q j represents the total amount of unloading at the jth unloading point;

[0042] (6) Planned production requirements at unloading points

[0043]

[0044] f j represents the planned unloading output of the jth unloading point;

[0045] (7) Ore grade constraints at unloading point

[0046]

[0047] e represents the grade limit, and a represents the maximum grade deviation allowed at the unloading point;

[0048] (8) Decision variable constraints:

[0049] x1,x2∈{0,1}

[0050] The decision variable takes the value of 0 or 1.

[0051] The multi-objective solution algorithm in step 3 specifically includes:

[0052] Step 3.1, preprocess the population to reduce the search pressure of the algorithm;

[0053] Step 3.2, generate reference vectors and associate them with individuals, divide them into Pareto levels, and take non-dominated individuals to generate direction vectors;

[0054] Step 3.3, search along the generated directions to generate promising offspring solutions;

[0055] Step 3.4, using differential evolution algorithm to ensure the diversity of offspring solutions;

[0056] Step 3.5, environmental selection, that is, selecting the optimal individuals whose number is the initial population size N from the parent population and the offspring solution;

[0057] Among them, an individual represents a scheduling plan of a fleet in a production cycle, a population is composed of multiple individuals, and a population represents a collection of multiple scheduling plans of a fleet in a production cycle.

[0058] The population preprocessing method in step 3.1 transforms the population using the characteristic matrix of the population covariance matrix, as shown below:

[0059] Step 3.1.1, calculate the mean m of the population

[0060]

[0061] Among them, x i represents the decision variable, N represents the population size;

[0062] Step 3.1.2, calculate the covariance C of the population

[0063]

[0064] Among them C i,j represents the element in the i-th row and j-th column of the covariance matrix; x i,k ,x j,k denotes the kth element in the i-th row and j-th column of the decision variable, m k represents the kth element of the mean m, D represents the dimension of the variable x, and N represents the population size;

[0065] Step 3.1.3, calculate the characteristic matrix of the covariance matrix

[0066] Cv i =λ i v i

[0067] Where C represents the covariance matrix, λ i represents the largest eigenvalue corresponding to the i-th row in the covariance matrix C, ν i The eigenvector representing the largest eigenvalue corresponding to the i-th row in the covariance matrix C;

[0068] Transformation matrix V = (v1, v2…, v m ) is the characteristic matrix of the covariance matrix, and P'=(Pm)V is used to transform the characteristics of the population to improve the quality of the offspring solution;

[0069] The specific steps for generating the direction vector in step 3.2 are as follows:

[0070] Step 3.2.1, generate reference vector W = (w1, w2...w N ) is associated with the individuals in the population, and the individuals with the smaller distance are selected according to the distance between the individuals in the population and the reference vector, which is recorded as set S;

[0071] Step 3.2.2, sort the individuals in set S by non-dominated order, divide them into Pareto levels, and select the non-dominated individuals as set Q;

[0072] Step 3.2.3, add the search boundary to the set Q to ensure the diversity of future solutions;

[0073] Step 3.2.4, combine the upper and lower boundaries of the search space to generate the direction vector, the formula is as follows:

[0074] v u,i =x i -U

[0075] v l,i =x j -L

[0076] Among them, U and L represent the upper and lower boundaries of the search space respectively, and x i ,x j ∈Q,v u,i ,v l,i are the direction vectors generated along the upper and lower boundaries respectively.

[0077] In step 3.3, a search is performed along the generated direction to generate a promising offspring solution. The specific steps are as follows:

[0078] According to the direction vector v generated in step 3.2 u,i ,v l,i , along v u,i ,v l,i The direction in which the solution is generated is promising. The specific method is as follows:

[0079]

[0080] Among them, c represents the search step size, x u,i ,x l,i represents the subsequent solution;

[0081] In step 3.4, a differential evolution algorithm is used to ensure the diversity of offspring solutions. The specific steps are as follows:

[0082] Step 3.4.1, generate mutation vector v based on individual differences i,G+1

[0083] v i,G+1 =x i,G +F(x r1,G -x r2,G ),i=1,...,N

[0084] Among them, G represents algebra, x i,G represents the i-th individual in the population, x r1,G With xr2,G Respectively represent two random individuals in the group, r1, r2∈[0,N]; F represents the scale factor, F∈[0,1];

[0085] Step 3.4.2, crossover operation between individuals

[0086]

[0087] Where CR is the crossover probability, CR∈[0,1]; v ij,G+1 Represents the population individual v ij,G The element of the jth dimension of ij,G+1 represents the population individual x ij,G The element of the jth dimension of ij,G+1 The element of the jth dimension representing the individual completed by the crossover;

[0088] Step 3.4.3, Greedy Selection Strategy

[0089]

[0090] Where f is the fitness function corresponding to the minimization problem, G represents the number of generations, and x i,G With x i,G+1 represent the individuals of the Gth generation and the G+1th generation, respectively, u i,G+1 Represents the individual after the mutation crossover operation;

[0091] The step 3.5, selecting the optimal individuals whose number is equal to the initial population size N from the parent population and the offspring solution, is specifically as follows:

[0092] Step 3.5.1, merge the parent population and the offspring solution, denoted as P h ;

[0093] Step 3.5.2, for P h Perform non-dominated sorting, divide the Pareto level and record it as R, R = 1, 2, ..., R max , where R max Indicates the maximum pareto level of an individual;

[0094] Step 3.5.3, sort the individuals according to the Pareto level from small to large, record the Pareto level of the Nth individual and record it as R l ;

[0095] Step 3.5.4, calculate the Pareto level from 1 to R l The number of individuals with -1 is recorded as N f , the pareto level is R l The number of individuals is denoted as N l ;

[0096] Step 3.5.5, if N f +N l = N, output the subsequent solutions, which contain Pareto levels from 1 to R l The optimal individual; if N f +N l >N, proceed to step 3.5.6;

[0097] Step 3.5.6, calculate the Pareto rank as R l The distance D between individuals, and arrange them from large to small distance D m ;

[0098] Step 3.5.7, calculate NN f , denoted as N q According to D m The size of the pareto level is R l Select the top N individuals q individuals;

[0099] Step 3.5.8, output the subsequent solutions, including the Pareto levels of 1 to R l -1 N f Individuals with pareto rank R l N q of individuals.

[0100] The solution process of step 4 is as follows:

[0101] Step 4.1, initialize the algorithm parameters, including the initial solution set population P, population size N, reference vector W, crossover and mutation parameters p c 、p m , search space upper and lower boundaries U, L, differential algorithm scale factor F, maximum number of evaluations FE max ;

[0102] Step 4.2, calculate the mean, covariance matrix and characteristic matrix of the initial scheduling solution set population P, and transform the population P according to the transformation formula, recorded as the transformed population P b ;

[0103] Step 4.3, transform the population P b Associated with the reference vector W and take the solutions associated with the reference vector in the set S;

[0104] Step 4.4, sort the individuals in set S by non-dominated order, divide them into Pareto levels, and select the non-dominated individuals in set Q;

[0105] Step 4.5, according to the generation criteria of the direction vector, take the individuals in Q to generate the direction vector based on the upper and lower boundaries of the search space;

[0106] Step 4.6, based on the direction vector, select an appropriate step size to generate promising offspring solutions;

[0107] Step 4.7, determine whether the number of iterations reaches the set threshold, if the condition is met, proceed to step 4.8, otherwise return to step 4.2;

[0108] Step 4.8, use the differential evolution algorithm to adjust the generated scheduling scheme again to increase its diversity;

[0109] Step 4.9, select the optimal scheduling solution with number N, denoted as P y , record the corresponding output, idle time of shovel or crusher and grade deviation;

[0110] Step 4.10, judge the iteration status. If the termination condition has been reached, output the current optimal Pareto scheduling solution; otherwise, clear the storage set P of the optimal scheduling solution. y , return to step 4.8.

[0111] The present invention also provides an open-pit mine transportation scheduling model optimization system considering large-scale key equipment failures, comprising:

[0112] Anti-disturbance module: used to determine the optimal number of trucks. When an electric shovel or crusher fails, the transportation scheduling system is disturbed. The M / M / m queue model is used to calculate the queue length of the entire transportation system trucks, thereby determining the optimal number of trucks to ensure the continuity of the entire transportation system operation.

[0113] Rescheduling model building module: used to build a rescheduling model; after re-determining the available truck set, a rescheduling model is constructed with the maximum output during the failure time, the minimum idle time of the shovel or crusher, and the minimum deviation from grade fluctuation;

[0114] Algorithm design module: A two-stage optimization mechanism is used to improve the differential evolution algorithm, which is used to implement the preprocessing operation of the population; the association operation between the reference vector and the individual and the determination of the search direction; the generation of promising offspring solutions; the improvement of the diversity of offspring solutions and the selection operation of parent solutions and offspring solutions;

[0115] Model solving module: used to solve the heavy scheduling model and output the optimal scheduling plan.

[0116] The present invention also provides an open-pit mine transportation scheduling model optimization device considering the failure of large key equipment, comprising:

[0117] A memory for storing a computer program for implementing an optimization method for an open-pit mine transportation scheduling model taking into account failures of large key equipment;

[0118] A processor is used to implement the open-pit mine transportation scheduling model optimization method considering the failure of large key equipment when executing the computer program.

[0119] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement an open-pit mine transportation scheduling model optimization method taking into account large-scale key equipment failures.

[0120] Compared with the prior art, the present invention has the following beneficial effects:

[0121] 1. The present invention adopts a queuing model to solve the number of trucks, thus bringing about a technical effect of being simple, easy to operate and having a high solving efficiency.

[0122] 2. Since the present invention adopts a population preprocessing method, it brings about the technical effect of being able to obtain a transportation plan that is more in line with the current production status of the mine.

[0123] 3. Since the present invention adopts a two-stage differential algorithm to solve the rescheduling model, it brings about the technical effect of higher quality of subsequent solutions.

[0124] In summary, the present invention proposes to use the principles of queuing theory to determine the scale of trucks when dealing with the rescheduling problem after a failure of key equipment in a mine. The model is simple and the solution efficiency is high, ensuring the rational use of truck resources. In addition, when solving the rescheduling model established after the failure occurs, a two-stage differential algorithm is used to consider the convergence and diversity of the solution to ensure the quality of the scheduling plan. At the same time, the transformation operation of the initial population makes the algorithm feasible and superior when dealing with complex problems. In the process of solving the rescheduling model established after the failure occurs, a transportation plan that is more in line with the current production status of the mine can be obtained. The present invention solves the current transportation scheduling problem after a failure of key equipment in a metal open-pit mine. Starting from the actual production situation of the mine, it has important significance for improving transportation efficiency, ensuring the continuity of mine operation, and improving the economic benefits of mining enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0125] Figure 1 It is a flow chart of the two-stage differential evolution algorithm used in the present invention to solve the model. DETAILED DESCRIPTION

[0126] The embodiments of the present invention are described in detail below with reference to the accompanying drawings and examples.

[0127] like Figure 1As shown in the figure, this open-pit mine transportation scheduling model optimization method considering the failure of large key equipment mainly includes the following steps:

[0128] Step 1: When a large key equipment in a mine fails, an M / M / m queuing model is constructed based on the principles of queuing theory. The length of the truck queue in the entire transportation system is used as the optimization indicator, and the M / M / m queuing model is solved by an iterative method to determine the optimal number of trucks that can meet production needs during the equipment failure.

[0129] In the M / M / m queue model of step 1, M represents that the arrival time of trucks and the loading and unloading time follow negative exponential distribution, and m represents the number of service stations, i.e., electric shovels and crushers;

[0130] Taking the queue length of the trucks in the transportation system as the optimization index of the entire transportation system, the formula is as follows:

[0131]

[0132] Where K is the total number of trucks in the transportation system after the failure of the key equipment; i represents the service point, i = 1…N; λ i represents the flow at the i-th service point, 0≤λ i ≤μ i m i , μ i represents the service rate of each service station at the i-th service point; m i is the number of identical service stations contained in the i-th service point; ρ i is the service intensity of the ith service point, that is, the probability that the service station (electric shovel, crusher) is busy, P mi is the probability that all servers in the system are occupied; a i is the correction factor; f i represents the queue leader at the i-th service point;

[0133]

[0134] Among them, C Bi represents the coefficient of variation of truck loading and unloading times;

[0135] In step 1, the simulation iteration method is used to solve the M / M / m queue model. The specific steps are as follows:

[0136] Step 1.1, initialization, take λ l =0 is the lower limit of flow rate, λ u =min{μ i m i} is the upper limit of traffic flow;

[0137] Step 1.2, let Substitution , calculate the size of trucks in the transportation system;

[0138] Step 1.3, take ε as the precision. If K-ε≤g(λ)≤K+ε, stop the iteration and get the required flow λ. If g(λ)>K, let λ u =λ, if g(λ)<K, let λ l =λ, and repeat step 1.2.

[0139] Step 2: Based on the optimal number of trucks obtained in step 1, the available truck set is re-determined, and a rescheduling model is constructed to maximize the output during the failure time, minimize the idle time of the shovel or crusher, and minimize the grade fluctuation deviation;

[0140] The rescheduling model in step 2 includes model objectives and model constraints, which are expressed as the following formula:

[0141] The model objectives include:

[0142] (1) Maximum output

[0143]

[0144] Among them, f1 represents the output, C r represents the truck's carrying capacity, x rij represents the number of trips made by the r-th truck between the i-th loading point and the j-th unloading point, N represents the set of loading points, M represents the set of unloading points, and R represents the set of available trucks;

[0145] (2) Minimum idle time for electric shovel or crusher

[0146]

[0147] Where f2 represents the idle time of the shovel or crusher, T represents the working time of a shift, and T z Indicates the loading time, T x Indicates the uninstallation time, x rij represents the number of trips made by the r-th truck between the i-th loading point and the j-th unloading point, T y1 With T y2 They represent the non-working time of the shovel or crusher due to rest and maintenance respectively; x1 and x2 are {0,1} decision variables, and when they are 0, they represent the failure of the shovel or crusher respectively;

[0148] (3) Minimum grade fluctuation deviation

[0149]

[0150] Among them, p yIndicates the original grade deviation, g i represents the ore grade at the i-th loading point, G j represents the target grade of the jth unloading point, K represents the set of available trucks;

[0151] The model constraints include:

[0152] (4) Production constraints at loading points

[0153]

[0154] g i represents the total load of the i-th loading point;

[0155] (5) Output constraints at unloading points

[0156]

[0157] q j represents the total amount of unloading at the jth unloading point;

[0158] (6) Planned production requirements at unloading points

[0159]

[0160] f j represents the planned unloading output of the jth unloading point;

[0161] (7) Ore grade constraints at unloading point

[0162]

[0163] e represents the grade limit, and a represents the maximum grade deviation allowed at the unloading point;

[0164] (8) Decision variable constraints:

[0165] x1,x2∈{0,1}

[0166] The decision variable takes the value of 0 or 1.

[0167] Step 3: According to the characteristics of the rescheduling model established in step 2, a two-stage optimization mechanism is used to improve the differential evolution algorithm, and a multi-objective solution algorithm is designed based on this;

[0168] In the first stage of the algorithm, the population is preprocessed, non-dominated individuals are selected to generate direction vectors, and promising offspring solutions are generated along the direction vectors to ensure the convergence of the offspring solutions; in the second stage of the algorithm, the differential evolution algorithm is used to ensure the diversity of the offspring solutions;

[0169] The multi-objective solution algorithm in step 3 includes:

[0170] Step 3.1, preprocess the population to reduce the search pressure of the algorithm;

[0171] The population is mainly transformed by using the characteristic matrix of the population's covariance matrix, as shown below:

[0172] Step 3.1.1, calculate the mean m of the population

[0173]

[0174] Among them, x i represents the decision variable, N represents the population size;

[0175] Step 3.1.2, calculate the covariance C of the population

[0176]

[0177] Among them, C i,j represents the element in the i-th row and j-th column of the covariance matrix; x i,k ,x j,k denotes the kth element in the i-th row and j-th column of the decision variable, m k represents the kth element of the mean m, D represents the dimension of the variable x, and N represents the population size;

[0178] Step 3.1.3, calculate the characteristic matrix of the covariance matrix

[0179] Cv i =λ i v i

[0180] Where C represents the covariance matrix, λ i represents the largest eigenvalue corresponding to the i-th row in the covariance matrix C, ν i The eigenvector representing the largest eigenvalue corresponding to the i-th row in the covariance matrix C.

[0181] Transformation matrix V = (v1, v2…, v m ) is the characteristic matrix of the covariance matrix. For the entire population, the strong connection between variables leads to increased search pressure. Therefore, P'=(Pm)V is used to perform characteristic transformation on the population to improve the quality of the offspring solution.

[0182] Step 3.2, generate reference vectors and associate them with individuals, divide them into Pareto levels, and take non-dominated individuals to generate direction vectors;

[0183] The direction vector is used to improve the search efficiency of the algorithm. The generation of the direction vector and the generation of the sampling solution are as follows:

[0184] Step 3.2.1, generate reference vector W = (w1, w2...w N ) is associated with the individuals in the population, and the individuals with the smaller distance are selected according to the distance between the individuals in the population and the reference vector, which is recorded as set S;

[0185] Step 3.2.2, sort the individuals in set S by non-dominated order, divide them into Pareto levels, and select the non-dominated individuals as set Q;

[0186] Step 3.2.3, in order to ensure the diversity of future solutions, add the search boundary to the set Q;

[0187] Step 3.2.4, combine the upper and lower boundaries of the search space to generate the direction vector, the formula is as follows:

[0188] v u,i =x i -U

[0189] v l,i =x j -L

[0190] Among them, U and L represent the upper and lower boundaries of the search space respectively, and x i ,x j ∈Q,v u,i ,v l,i are the direction vectors generated along the upper and lower boundaries respectively;

[0191] Step 3.3, search along the generated direction to generate promising offspring solutions; the specific steps are as follows:

[0192] According to the direction vector v generated in step 3.2 u,i ,v l,i , along v u,i ,v l,i The direction in which the solution is generated is promising. The specific method is as follows:

[0193]

[0194] Among them, c represents the search step size, x u,i ,x l,i Indicates the later solution.

[0195] Step 3.4: Use differential evolution algorithm to ensure the diversity of offspring solutions. The specific steps are as follows:

[0196] Step 3.4.1, generate mutation vector v based on individual differences i,G+1

[0197] v i,G+1 =x i,G +F(x r1,G -xr2,G ),i=1,...,N

[0198] Where G represents algebra, x i,G represents the i-th individual in the population, x r1,G With x r2,G Respectively represent two random individuals in the group, r1, r2∈[0,N]; F represents the scale factor, F∈[0,1];

[0199] Step 3.4.2, crossover operation between individuals

[0200]

[0201] Where CR is the crossover probability, CR∈[0,1]; v ij,G+1 Represents the population individual v ij,G The element of the jth dimension of ij,G+1 represents the population individual x ij,G The element of the jth dimension of ij,G+1 The element of the jth dimension representing the individual completed by the crossover;

[0202] Step 3.4.3, Greedy Selection Strategy

[0203]

[0204] Where f is the fitness function corresponding to the minimization problem, G represents the number of generations, and x i,G With x i,G+1 represent the individuals of the Gth generation and the G+1th generation, respectively, u i,G+1 Represents the individual after the mutation crossover operation.

[0205] Step 3.5, environmental selection, that is, selecting the optimal individuals whose number is the initial population size N from the parent population and the offspring solution;

[0206] The specific steps are as follows:

[0207] Step 3.5.1, merge the parent population and the offspring solution, denoted as P h ;

[0208] Step 3.5.2, for P h Perform non-dominated sorting, divide the Pareto level and record it as R, R = 1, 2, ..., R max , where R max Indicates the maximum pareto level of an individual;

[0209] Step 3.5.3, sort the individuals according to the Pareto level from small to large, record the Pareto level of the Nth individual and record it as R l ;

[0210] Step 3.5.4, calculate the Pareto level from 1 to R l The number of individuals with -1 is recorded as N f , the pareto level is R l The number of individuals is denoted as N l ;

[0211] Step 3.5.5, if N f +N l = N, output the subsequent solutions, which contain Pareto levels from 1 to R l The optimal individual; if N f +N l >N, proceed to step 3.5.6;

[0212] Step 3.5.6, calculate the Pareto rank as R l The distance D between individuals, and arrange them from large to small distance D m ;

[0213] Step 3.5.7, calculate NN f , denoted as N q According to D m The size of the pareto level is R l Select the top N individuals q individuals;

[0214] Step 3.5.8, output the subsequent solutions, including the Pareto levels of 1 to R l -1 N f Individuals with pareto rank R l N q Individual.

[0215] Among them, an individual represents a scheduling plan of a fleet in a production cycle, a population is composed of multiple individuals, and a population represents a collection of multiple scheduling plans of a fleet in a production cycle.

[0216] Since what is generated is the scheduling plan of the fleet within a production cycle, the individual represents the scheduling plan of the fleet within a production cycle. The population is composed of multiple individuals, and the population represents the set of multiple scheduling plans of the fleet within a production cycle. These plans compete with each other, that is, individuals compete with each other, and the best plan is selected through iterative update through optimization criteria.

[0217] like Figure 1 As shown, in step 4, the rescheduling model obtained in step 2 is solved according to the multi-objective solution algorithm designed in step 3 to obtain the optimal scheduling plan for the fleet within a working cycle.

[0218] Step 4.1, initialize the algorithm parameters, including the initial solution set population P, population size N, reference vector W, crossover and mutation parameters p c 、p m , search space upper and lower boundaries U, L, differential algorithm scale factor F, maximum number of evaluations FE max ;

[0219] Step 4.2, calculate the mean, covariance matrix and characteristic matrix of the initial scheduling solution set population P, and transform the population P according to the transformation formula, recorded as the transformed population P b ;

[0220] Step 4.3, transform the population P b Associated with the reference vector W and take the solutions associated with the reference vector in the set S;

[0221] Step 4.4, sort the individuals in set S by non-dominated order, divide them into Pareto levels, and select the non-dominated individuals in set Q;

[0222] Step 4.5, according to the generation criteria of the direction vector, take the individuals in Q to generate the direction vector based on the upper and lower boundaries of the search space;

[0223] Step 4.6, based on the direction vector, select an appropriate step size to generate promising offspring solutions;

[0224] Step 4.7, determine whether the number of iterations reaches the set threshold, if the condition is met, proceed to step 4.8, otherwise return to step 4.2;

[0225] Step 4.8, use the differential evolution algorithm to adjust the generated scheduling scheme again to increase its diversity;

[0226] Step 4.9, select the optimal scheduling solution with number N, denoted as P y , record the corresponding output, idle time of shovel or crusher and grade deviation;

[0227] Step 4.10, judge the iteration status. If the termination condition has been reached, output the current optimal Pareto scheduling solution; otherwise, clear the storage set P of the optimal scheduling solution. y , return to step 4.8.

[0228] The present invention also provides an open-pit mine transportation scheduling model optimization system considering large-scale key equipment failures, comprising:

[0229] Anti-disturbance module: used to determine the optimal number of trucks in step 1. When the shovel or crusher fails, the transportation scheduling system is disturbed. The M / M / m queue model is used to calculate the queue length of the trucks in the entire transportation system, thereby determining the optimal number of trucks to ensure the continuity of the operation of the entire transportation system.

[0230] Rescheduling model building module: used to implement the establishment of the rescheduling model in step 2; after re-determining the available truck set, a rescheduling model is constructed with the maximum output during the failure time, the minimum idle time of the shovel or crusher, and the minimum deviation of grade fluctuation;

[0231] Algorithm design module: A two-stage optimization mechanism is used to improve the differential evolution algorithm, which is used to implement the preprocessing operation of the population in step 3.1, the association operation between the reference vector and the individual and the determination of the search direction in step 3.2, the generation of promising offspring solutions in step 3.3, the improvement of the diversity of offspring solutions in step 3.4, and the selection operation of parent solutions and offspring solutions in step 3.5.

[0232] Model solving module: used to solve the rescheduling model in step 4 and output the optimal scheduling solution.

[0233] The present invention also provides an open-pit mine transportation scheduling model optimization device considering the failure of large key equipment, comprising:

[0234] A memory for storing a computer program for implementing an optimization method for an open-pit mine transportation scheduling model taking into account failures of large key equipment;

[0235] A processor is used to implement the open-pit mine transportation scheduling model optimization method considering the failure of large key equipment when executing the computer program.

[0236] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement an open-pit mine transportation scheduling model optimization method taking into account large-scale key equipment failures.

[0237] Aiming at the open-pit mine transportation scheduling problem with key equipment failure, the present invention constructs a rescheduling model with the maximum output during the failure time, the shortest idle time of the electric shovel or crusher and the smallest grade fluctuation deviation according to the actual production requirements and indicators of the mine; and adopts an improved two-stage differential evolution algorithm to solve the model, so as to obtain a transportation plan that is more in line with the current production status of the mine. Starting from the actual production situation of the mine, the transportation efficiency is effectively improved, the continuity of the mine production is guaranteed, and it is of great significance to improve the economic benefits of the mining enterprise.

Claims

1. An open-pit mine transportation scheduling model optimization method considering large-scale key equipment failures, characterized in that: The following steps are involved: Step 1: When a large key equipment in a mine fails, an M / M / m queuing model is constructed based on the principles of queuing theory. The length of the truck queue in the entire transportation system is used as the optimization indicator, and the M / M / m queuing model is solved by an iterative method to determine the optimal number of trucks that can meet production needs during the equipment failure. Step 2: Based on the optimal number of trucks obtained in step 1, the available truck set is re-determined, and a rescheduling model is constructed to maximize the output during the failure time, minimize the idle time of the shovel or crusher, and minimize the grade fluctuation deviation; Step 3: According to the characteristics of the rescheduling model established in step 2, a two-stage optimization mechanism is used to improve the differential evolution algorithm, and a multi-objective solution algorithm is designed based on this; Step 4: Solve the rescheduling model obtained in step 2 according to the multi-objective solution algorithm designed in step 3 to obtain the optimal scheduling plan for the fleet within a working cycle.

2. The open-pit mine transportation scheduling model optimization method considering large-scale key equipment failure according to claim 1 is characterized in that: In the M / M / m queue model of step 1, M represents that the arrival time of trucks and the loading and unloading time follow negative exponential distribution, and m represents the number of service stations, i.e., electric shovels and crushers; Taking the queue length of the trucks in the transportation system as the optimization index of the entire transportation system, the formula is as follows: Where K is the total number of trucks in the transportation system after the failure of the key equipment; i represents the service point, i = 1…N; λ i represents the flow at the i-th service point, 0≤λ i ≤μ i m i , μ i represents the service rate of each service station at the i-th service point; m i is the number of identical service stations contained in the i-th service point; ρ i is the service intensity of the ith service point, that is, the probability that the service station (electric shovel, crusher) is busy, P mi is the probability that all servers in the system are occupied; a i is the correction factor; f i represents the queue leader at the i-th service point; Among them, C Bi represents the coefficient of variation of truck loading and unloading times; In step 1, an iterative method is used to solve the M / M / m queue model. The specific steps are as follows: Step 1.1, initialization, take λ l =0 is the lower limit of flow rate, λ u =min{μ i m i } is the upper limit of traffic flow; Step 1.2, let Substitution , calculate the size of trucks in the transportation system; Step 1.3, take ε as the precision. If K-ε≤g(λ)≤K+ε, stop the iteration and get the required flow λ. If g(λ)>K, let λ u =λ, if g(λ)<K, let λ l =λ, and repeat step 1.

2.

3. The open-pit mine transportation scheduling model optimization method considering large-scale key equipment failure according to claim 1 is characterized in that: The rescheduling model in step 2 includes model objectives and model constraints, which are expressed as the following formula: The model objectives include: (1) Maximum output Among them, f1 represents the output, C r represents the truck's carrying capacity, x rij represents the number of trips made by the r-th truck between the i-th loading point and the j-th unloading point, N represents the set of loading points, M represents the set of unloading points, and R represents the set of available trucks; (2) Minimum idle time for electric shovel or crusher Where f2 represents the idle time of the shovel or crusher, T represents the working time of a shift, T z Indicates the loading time, T x Indicates the uninstallation time, x rij Indicates that the rth truck arrives at the i-th loading point The number of transports between the jth unloading points, T y1 With T y2 They represent the non-working time of the electric shovel or crusher due to rest and maintenance respectively; x1 and x2 are {0,1} decision variables, and when they are 0, they represent the non-working time of the electric shovel or crusher due to rest and maintenance respectively. or crusher failure; (3) Minimum grade fluctuation deviation Among them, p y Indicates the original grade deviation, g i represents the ore grade at the i-th loading point, G j represents the target grade of the jth unloading point, K represents the set of available trucks; The model constraints include: (4) Production constraints at loading points g i represents the total load of the i-th loading point; (5) Output constraints at unloading points q j represents the total amount of unloading at the jth unloading point; (6) Planned production requirements at unloading points f j represents the planned unloading output of the jth unloading point; (7) Ore grade constraints at unloading point e represents the grade limit, and a represents the maximum grade deviation allowed at the unloading point; (8) Decision variable constraints: x1,x2∈{0,1} The decision variable takes the value of 0 or 1.

4. The open-pit mine transportation scheduling model optimization method considering large-scale key equipment failure according to claim 1 is characterized in that: The multi-objective solution algorithm in step 3 specifically includes: Step 3.1, preprocess the population to reduce the search pressure of the algorithm; Step 3.2, generate reference vectors and associate them with individuals, divide them into Pareto levels, and take non-dominated individuals to generate direction vectors; Step 3.3, search along the generated directions to generate promising offspring solutions; Step 3.4, using differential evolution algorithm to ensure the diversity of offspring solutions; Step 3.5, environmental selection, that is, selecting the optimal individuals whose number is the initial population size N from the parent population and the offspring solution; Among them, an individual represents a scheduling plan of a fleet in a production cycle, a population is composed of multiple individuals, and a population represents a collection of multiple scheduling plans of a fleet in a production cycle.

5. The method for optimizing the open-pit mine transportation scheduling model considering failure of large key equipment according to claim 4 is characterized in that: The population preprocessing method in step 3.1 transforms the population using the characteristic matrix of the population covariance matrix, as shown below: Step 3.1.1, calculate the mean m of the population Among them, x i represents the decision variable, N represents the population size; Step 3.1.2, calculate the covariance C of the population Among them, C i,j represents the element in the i-th row and j-th column of the covariance matrix; x i,k ,x j,k denotes the kth element in the i-th row and j-th column of the decision variable, m k represents the kth element of the mean m, D represents the dimension of the variable x, and N represents the population size; Step 3.1.3, calculate the characteristic matrix of the covariance matrix Cv i =λ i v i Where C represents the covariance matrix, λ i represents the largest eigenvalue corresponding to the i-th row in the covariance matrix C, ν i The eigenvector representing the largest eigenvalue corresponding to the i-th row in the covariance matrix C; Transformation matrix V = (v1, v2…, v m ) is the characteristic matrix of the covariance matrix, and P'=(Pm)V is used to transform the characteristics of the population to improve the quality of the offspring solution; The specific steps for generating the direction vector in step 3.2 are as follows: Step 3.2.1, generate reference vector W = (w1, w2...w N ) is associated with the individuals in the population, and the individuals with the smaller distance are selected according to the distance between the individuals in the population and the reference vector, which is recorded as set S; Step 3.2.2, sort the individuals in set S by non-dominated order, divide them into Pareto levels, and select the non-dominated individuals as set Q; Step 3.2.3, add the search boundary to the set Q to ensure the diversity of future solutions; Step 3.2.4, combine the upper and lower boundaries of the search space to generate the direction vector, the formula is as follows: v u,i =x i -U v l,i =x j -L Among them, U and L represent the upper and lower boundaries of the search space respectively, and x i ,x j ∈Q,v u,i ,v l,i are the direction vectors generated along the upper and lower boundaries respectively.

6. The method for optimizing the open-pit mine transportation scheduling model considering failure of large key equipment according to claim 4 is characterized in that: In step 3.3, a search is performed along the generated direction to generate a promising offspring solution. The specific steps are as follows: According to the direction vector v generated in step 3.2 u,i ,v l,i , along v u,i ,v l,i The direction in which the solution is generated is promising. The specific method is as follows: Among them, c represents the search step size, x u,i ,x l,i represents the subsequent solution; In step 3.4, a differential evolution algorithm is used to ensure the diversity of offspring solutions. The specific steps are as follows: Step 3.4.1, generate mutation vector v based on individual differences i,G+1 v i,G+1 =x i,G +F(x r1,G -x r2,G ),i=1,...,N Among them, G represents algebra, x i,G represents the i-th individual in the population, x r1,G With x r2,G Respectively represent two random individuals in the group, r1, r2∈[0,N]; F represents the scale factor, F∈[0,1]; Step 3.4.2, crossover operation between individuals Where CR is the crossover probability, CR∈[0,1]; v ij,G+1 Represents the population individual v ij,G The element of the jth dimension of ij,G+1 represents the population individual x ij,G The element of the jth dimension of ij,G+1 The element of the jth dimension representing the individual completed by the crossover; Step 3.4.3, Greedy Selection Strategy Where f is the fitness function corresponding to the minimization problem, G represents the number of generations, and x i,G With x i,G+1 represent the individuals of the Gth generation and the G+1th generation, respectively, u i,G+1 Represents the individual after the mutation crossover operation; The step 3.5, selecting the optimal individuals whose number is equal to the initial population size N from the parent population and the offspring solution, is specifically as follows: Step 3.5.1, merge the parent population and the offspring solution, denoted as P h ; Step 3.5.2, for P h Perform non-dominated sorting, divide the Pareto level and record it as R, R = 1, 2, ..., R max , where R max Indicates the maximum pareto level of an individual; Step 3.5.3, sort the individuals according to the Pareto level from small to large, record the Pareto level of the Nth individual and record it as R l ; Step 3.5.4, calculate the Pareto level from 1 to R l The number of individuals with -1 is recorded as N f , the pareto level is R l The number of individuals is denoted as N l ; Step 3.5.5, if N f +N l = N, output the subsequent solutions, which contain Pareto levels from 1 to R l The optimal individual; if N f +N l >N, proceed to step 3.5.6; Step 3.5.6, calculate the Pareto rank as R l The distance D between individuals, and arrange them from large to small distance D m ; Step 3.5.7, calculate NN f , denoted as N q According to D m The size of the pareto level is R l Select the top N individuals q individuals; Step 3.5.8, output the subsequent solutions, including the Pareto levels of 1 to R l -1 N f Individuals with pareto rank R l N q Individual.

7. The method for optimizing the open-pit mine transportation scheduling model considering failure of large key equipment according to claim 1 is characterized in that: The solution process of step 4 is as follows: Step 4.1, initialize the algorithm parameters, including the initial solution set population P, population size N, reference vector W, crossover and mutation parameters p c 、p m , search space upper and lower boundaries U, L, differential algorithm scale factor F, maximum number of evaluations FE max ; Step 4.2, calculate the mean, covariance matrix and characteristic matrix of the initial scheduling solution set population P, and transform the population P according to the transformation formula, recorded as the transformed population P b ; Step 4.3, transform the population P b Associated with the reference vector W and take the solutions associated with the reference vector in the set S; Step 4.4, sort the individuals in set S by non-dominated order, divide them into Pareto levels, and select the non-dominated individuals in set Q; Step 4.5, according to the generation criteria of the direction vector, take the individuals in Q to generate the direction vector based on the upper and lower boundaries of the search space; Step 4.6, based on the direction vector, select an appropriate step size to generate promising offspring solutions; Step 4.7, determine whether the number of iterations reaches the set threshold, if the condition is met, proceed to step 4.8, otherwise return to step 4.2; Step 4.8, use the differential evolution algorithm to adjust the generated scheduling scheme again to increase its diversity; Step 4.9, select the optimal scheduling solution with number N, denoted as P y , record the corresponding output, idle time of shovel or crusher and grade deviation; Step 4.10, judge the iteration status. If the termination condition has been reached, output the current optimal Pareto scheduling solution; otherwise, clear the storage set P of the optimal scheduling solution. y , return to step 4.

8.

8. An open-pit mine transportation scheduling model optimization system considering large-scale key equipment failures, characterized in that: include: Anti-disturbance module: used to determine the optimal number of trucks; When an electric shovel or crusher fails, the transportation scheduling system is disturbed. The M / M / m queue model is used to calculate the queue length of the entire transportation system trucks, thereby determining the optimal number of trucks to ensure the continuity of the entire transportation system operation. Rescheduling model building module: used to build a rescheduling model; after re-determining the available truck set, a rescheduling model is constructed with the maximum output during the failure time, the minimum idle time of the shovel or crusher, and the minimum deviation from grade fluctuation; Algorithm design module: A two-stage optimization mechanism is used to improve the differential evolution algorithm, which is used to implement the preprocessing operation of the population; the association operation between the reference vector and the individual and the determination of the search direction; the generation of promising offspring solutions; the improvement of the diversity of offspring solutions and the selection operation of parent solutions and offspring solutions; Model solving module: used to solve the heavy scheduling model and output the optimal scheduling plan.

9. An open-pit mine transportation scheduling model optimization device considering the failure of large key equipment, characterized in that: include: A memory for storing a computer program for implementing a method for optimizing an open-pit mine transportation scheduling model taking into account a large-scale key equipment failure as described in any one of claims 1 to 7; A processor is used to implement an open-pit mine transportation scheduling model optimization method considering large-scale key equipment failures as described in any one of claims 1-7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it can implement the open-pit mine transportation scheduling model optimization method considering the failure of large key equipment described in claims 1-7.

Images