CN117634772A - Strip mine unmanned truck scheduling method considering truck performance - Google Patents

Abstract

The invention discloses a method for dispatching unmanned trucks in open-pit mines considering truck performance, which includes the following steps; Step 1: Establishment of open-pit mine truck operating status database; Step 2: Establishment of truck operating efficiency prediction model; Step 3: According to The open-pit mine truck operating status database and the truck operating efficiency prediction model are used to establish a multi-objective open-pit mine driverless truck scheduling optimization model; Step 4: Use the multi-objective particle swarm algorithm to solve the scheduling optimization model established in step 3 for the final Optimization of scheduling plans. Use historical data of mine production to establish a prediction model for truck transportation efficiency, and bring the prediction results into the solution model for real-time mine truck scheduling optimization to obtain a scheduling plan that is closer to actual production, thereby reducing transportation costs in production.

Description

A method for dispatching driverless trucks in open pit mines considering truck performance

Technical field

The present invention relates to the technical field of open-pit mine operation scheduling methods, and in particular to a method for dispatching unmanned trucks in open-pit mines that considers truck performance.

Background technique

The implementation of the open pit mine production plan is completed by allocating transportation equipment between the blasting and unloading areas. The main task of mine cart transportation is to transport ore from the blast pile to the unloading station or storage yard. Therefore, optimizing the scheduling of mine cars and improving the transportation efficiency of mine cars are of great significance to reducing production costs and improving production efficiency. The optimization of mine car scheduling is also an important link in the optimization of open pit mine systems.

Traditional mine car scheduling optimization usually assumes that the performance and transportation efficiency of mine cars are at the same level and unchanged. But this is not consistent with the actual production situation in reality. In actual production, the trucks working in each shift have different models, loads, kilometers, engine conditions, tire wear conditions, and maintenance conditions. These factors cause each truck to have different transport efficiency. Therefore, the different performance of each truck should be considered in the scheduling optimization of mine trucks, so that the optimization results can be closer to the actual production situation. This requires establishing a new method for autonomous truck dispatching in open pit mines that considers truck performance. Analyzing historical data about mine truck conditions in open-pit mine production and establishing a predictive model are of constructive significance for establishing this new scheduling method that considers truck performance.

Contents of the invention

In order to overcome the shortcomings of the above existing technologies, the purpose of the present invention is to provide a method for dispatching unmanned trucks in open-pit mines that considers truck performance, uses historical data of mine production to establish a prediction model for truck transportation efficiency, and brings the prediction results to the The solution model for real-time scheduling optimization of mine cars can obtain a scheduling plan that is closer to actual production, thereby reducing transportation costs in production.

In order to achieve the above objects, the technical solution adopted by the present invention is:

A method for dispatching driverless trucks in open pit mines considering truck performance, including the following steps;

Step 1: Establishment of open-pit mine truck operating status database;

Step 2: Establishment of truck operating efficiency (speed) prediction model;

Step 3: Establish a multi-objective open-pit mine driverless truck scheduling optimization model based on the open-pit mine truck operating status database and truck operating efficiency (speed) prediction model;

Step 4: Use the multi-objective particle swarm algorithm to solve the scheduling optimization model for optimizing the final scheduling solution.

The specific step 1 is: collect the historical data of the truck load, completed kilometers, engine condition, tire wear, repair and maintenance status and average speed of each shift of each truck in the open-pit mine production, and establish the operating status of the open-pit mine trucks. Database; develop a statistical table of operating status data of open-pit mine trucks to collect historical data on the operating status of open-pit mine trucks.

The specific step 2 is: through the analysis of factors affecting truck transportation performance, five influencing factors are established, including truck load, kilometers traveled, number of engine failures, kilometers used by tires, and number of daily maintenance times. A prediction model with the average vehicle speed per shift as the output parameter.

In step 3, based on the actual production and transportation process of the open pit mine, with the goals of minimizing truck waiting time, minimizing truck waiting time, minimizing remaining production, and minimizing ore grade deviation, an optimization model for driverless truck dispatching in the open pit mine is established. ; The open-pit mine driverless truck scheduling optimization model includes an objective function and constraints.

The objective function is specifically:

Minimize trucking costs:

The transportation cost of a truck consists of the transportation cost of a heavy truck from the loading point to the unloading point, the transportation cost of an empty truck from the unloading point to the loading point, and the transportation cost of an empty truck to and from the refueling point;

Minimize truck waiting time:

The waiting time of a truck is composed of the truck shift working time minus the actual running time of the truck. The actual running time includes the running time between loading and unloading points, loading and unloading time, time to and from the gas station, and refueling time;

Minimize remaining production:

In order to ensure that the ore produced in one shift at the mining site can be transported as much as possible, that is, the arranged transportation capacity is as close as possible to the production capacity, the absolute value of the output minus the transportation volume is required to be minimum;

Minimize ore grade deviation:

The sum of the ore grade deviations at each unloading point is the smallest;

The constraints are specifically:

Production capacity constraints at the unloading point:

The total amount of ore shipped to each unloading point shall not exceed the maximum production capacity of that unloading point;

Production planning requirements for unloading points:

The total amount of ore unloaded at each unloading point must at least meet the production demand of that unloading point;

Production capacity constraints for load points:

The total amount of ore loaded at each loading point shall not exceed the maximum production capacity of that loading point;

Ore grade requirements at the unloading point:

The grade deviation of ore unloaded at each unloading point shall not exceed the maximum grade deviation;

Traffic flow continuity constraints ensure that the incoming and outgoing traffic flow of each loading/unloading point is equal;

Remaining fuel constraints:

Self-driving trucks are no longer like traditional trucks. There is a driver who can keep an eye on the remaining fuel level at any time while driving. Therefore, when planning and dispatching driverless trucks for traffic flow, it is necessary to monitor their remaining fuel levels. Its specific manifestation is that the remaining fuel quantity of each truck must not be less than K in order to return to the refueling point for refueling. Among them, the minimum remaining oil amount K = the maximum oil consumption from the unloading point to the loading point + the maximum oil consumption from the loading point to the unloading point + the maximum oil consumption from the unloading point to the refueling point O;

The number of truck transportation requirements, the number of truck transportation must be a positive integer;

X _ijkl ,Y _ijkl ∈{0,1,2,3...}

Constraints in conflict avoidance situations;

When two trucks are at a certain loading or unloading point at the same time, the truck routes will conflict, which will have a greater impact on the truck's operating costs and queuing time. In order to reduce the impact caused by truck task conflicts as much as possible, it is necessary to set priorities for trucks under the constraints of the initial state. Given that fully loaded trucks are more expensive to operate than empty trucks, priority is given to fully loaded trucks. In addition, the priority determination also applies to when two trucks meet at the same intersection, and the unloaded driverless truck should give way.

Constraints in truck failure conditions;

Once a truck fails and cannot move, the truck should exit the traffic flow planning and dispatching system in a timely manner. Therefore, when a driverless truck alerts the general dispatch center, the truck should be immediately excluded from the system and the priority of the path, loading point or unloading point where the truck is located should be reduced so that the planning and dispatching plan can be adjusted in a timely manner. Achieve the company’s established goals;

The parameter symbols used in the above formula are defined as follows:

i: The index number of the loading point, indicating the i-th loading point (i.e., the excavator position), i = 1, 2,...,I,; j: The index number of the unloading point, indicating the j-th unloading point (i.e., crushing station location), j=1,2,...,J,; k: truck model index number, indicating the k-th model of truck, k=1,2,...,K, type; l: truck number index number, Represents the l-th k-type truck, l= _{1,2 ,} ...,L,; The number of times a k-type truck of l is empty from unloading point j to loading point i, times; d _ij : the distance from loading point i to unloading point j, km; d _jo : the distance from unloading point j to refueling point O, km ; d _oi , the optimal distance from refueling point O to loading point i, km; C _k : loading capacity of k-type truck, t; CE _kl : unit distance cost of k-type truck numbered l when overloaded, yuan /km; CL _kl : The unit distance cost of the k-type truck numbered l when empty, yuan/km; E _kl : The fuel tank capacity of the k-type truck numbered l, L; EE _kl : The k-type truck numbered l Fuel consumption per unit distance when the truck is heavily loaded, L/km; EL _kl , fuel consumption per unit distance when the k-type truck numbered l is unloaded, L/km; K: minimum remaining fuel volume, L; g _i : loading point i The maximum production capacity, t; f _j : the minimum production demand of unloading point j (within one shift), t; q _j : the maximum production capacity of unloading point j, t; e: the limit of ore grade; α _i : loading point The ore grade of i; β: the allowable error of the ore grade; G _j : the target grade of the unloading point j; T _lim : the shift working time; SE _kl : the average speed of the K-type truck numbered l when heavy load, km/h; SL _kl : the average speed of the K-type truck numbered l when empty, km/h; T _O , the average time for truck refueling, min; T _z : the average time for truck loading, min; T _q : the average time for truck unloading Time, min; K _jOkl : The number of runs of the k-type truck numbered l from the unloading point j to the refueling point O when empty, times; K _Oikl : The k-type truck numbered l from the refueling point O to the loading point when empty The number of runs of point i, times.

The specific step 4 is:

Set the basic parameters of the multi-objective particle swarm algorithm such as inertia weight ω and flight speed c1, c2, and solve the open-pit mine driverless truck scheduling optimization model through the multi-objective particle swarm algorithm;

Determine whether the current solution (scheme) meets the three constraints of production capacity, grade requirements, and traffic flow control, and then update the position and speed of the particles; the above entire process is completed as one iteration. When the algorithm iteration is completed, the result is output and the calculation process ends. Otherwise, continue with the multi-objective particle swarm algorithm solution; the specific steps are as follows:

Step1: Initialize the particle swarm P; for each particle, determine its initial position and speed and calculate the target vector of each particle in the particle swarm P;

Step2: Save some particles in the particle swarm P in the external particle swarm NP, and the positions of these particles are non-inferior solutions;

Step3: Determine the best initial position of each particle, that is, the initial position of each particle itself;

Step 4: Divide the target space into many grids (hypercubes), and determine the grid where each particle is located based on the target vector corresponding to the particle;

Step5: Define the fitness value (the ratio of a number equal to or greater than 1 to the number of NP members contained in the grid) for each grid that contains at least one external particle swarm individual, and then for each particle, according to the roulette method Select a grid and randomly select an individual from the external particle group as the gbest of the particle;

Step6: Update the position and velocity of all particles according to the basic formula of PSO;

Step7: Use the following measures to prevent particles from flying out of the search space: Once a particle flies out of the boundary of a certain decision variable, the particle stays on the boundary and changes its flight direction at the same time;

Step8: Calculate the target vector of each particle in the particle swarm P;

Step9: Use the adaptive grid method to update and maintain the external particle swarm NP;

Step10: Update the pbest of the particles. According to the comparison between the new solution obtained during the particle flight and its own best position pbest, if the new solution dominates pbest, the new solution is the new pbest; otherwise, pbest remains unchanged; if the new solution and pbest are not dominated by each other, then Randomly select one of the two as the new best position;

Step11: If the termination condition is established, stop the search and determine whether the plan meets the three constraints of production capacity, grade limit and traffic flow control in the model; when the above conditions are met at the same time, output the optimization plan; otherwise, go to Step6.

Beneficial effects of the present invention:

The dispatch model constructed in the present invention is more in line with the actual production situation of the mine, and the dispatch plan obtained from the plan can effectively reduce transportation costs and improve transportation efficiency.

This invention aims at the optimization problem of unmanned truck dispatching in open-pit mines considering truck performance, with the goal of minimizing truck waiting time, minimizing truck waiting time, minimizing remaining production, and minimizing ore grade deviation, and taking into account the different transportation efficiencies of trucks. , constructed an optimization model for driverless truck dispatching in open-pit mines. The multi-objective particle swarm algorithm is used for optimization and solution, which effectively solves the optimization problem of driverless truck dispatching in open-pit mines and meets the needs of actual mine transportation.

Description of drawings

Figure 1 is a truck production scheduling flow chart.

Figure 2 is a flow chart of unmanned truck dispatching for open pit mines taking into account truck performance according to the present invention.

Detailed ways

The present invention will be further described in detail below with reference to examples.

A method for dispatching driverless trucks in open pit mines considering truck performance, including the following steps;

First, establish an open-pit mine truck operating status database based on historical data collected during open-pit mine production on each truck's load capacity, completed kilometers, engine condition, tire wear, repair and maintenance status, and average speed of each shift;

Secondly, a prediction model based on BP neural network is trained based on historical data, and the real-time running speed of the truck is obtained through prediction;

Finally, a driverless truck dispatching system for open-pit mines was established with production capacity, grade requirements, and traffic flow control as constraints, with the goal of minimizing truck waiting time, minimizing truck waiting time, minimizing remaining production, and minimizing ore grade deviation. model, and the multi-objective particle swarm algorithm is used to solve the model.

As shown in Figure 1, it is an open pit mine truck production scheduling flow chart. In actual production, trucks transport to and from loading points, unloading points, and gas stations. On-site truck dispatching is one of the main ways to improve production efficiency.

As shown in Figure 2, the present invention proposes a method for dispatching unmanned trucks in open-pit mines that considers truck performance, which mainly includes the following steps:

Step 1. Establishment of open pit truck operating status database:

Collect historical data on the truck load, completed kilometers, engine condition, tire wear, repair and maintenance status, and average speed of each shift of each truck in open-pit mine production, and establish an open-pit mine truck operating status database; formulate open-pit mine truck operation The status data statistics table collects historical data on the operating status of open-pit mine trucks. The truck operating status data statistical table designed in the present invention is as follows:

Step 2. Establishment of truck operating efficiency (speed) prediction model:

Through the analysis of factors that affect truck transportation performance, we established five influencing factors as input parameters: truck load, kilometers traveled, number of engine failures, kilometers used by tires, and number of daily maintenance times. Based on the average speed of the vehicle per shift, is the prediction model for the output parameters.

In this invention, 100 groups of historical statistical data from January 1, 2021 to April 10, 2021 are selected to train the regression model, of which 90 groups constitute a training data set, which is used to train the network to obtain the calculation model; the remaining 10 groups are simulation Data used to verify the computational accuracy of the model.

Step 3. Establishment of multi-objective open-pit mine driverless truck scheduling optimization model:

Based on the actual production and transportation process of the open-pit mine, with the goals of minimizing truck waiting time, minimizing truck waiting time, minimizing remaining production, and minimizing ore grade deviation, an open-pit mine unmanned truck scheduling optimization model is established; the specific expression of the model is as follows :

(1): Objective function:

Minimize trucking costs:

The transportation cost of a truck consists of the transportation cost of a heavy truck from the loading point to the unloading point, the transportation cost of an empty truck from the unloading point to the loading point, and the transportation cost of an empty truck to and from the refueling point;

Minimize truck waiting time:

The waiting time of a truck is composed of the truck shift working time minus the actual running time of the truck. The actual running time includes the running time between loading and unloading points, loading and unloading time, time to and from the gas station, and refueling time;

Minimize remaining production:

In order to ensure that the ore produced in one shift at the mining site can be transported as much as possible, that is, the arranged transportation capacity is as close as possible to the production capacity, the absolute value of the output minus the transportation volume is required to be minimum;

Minimize ore grade deviation:

The sum of the ore grade deviations at each unloading point is the smallest;

(2) Constraints:

Production capacity constraints at the unloading point:

The total amount of ore shipped to each unloading point shall not exceed the maximum production capacity of that unloading point;

Production planning requirements for unloading points:

The total amount of ore unloaded at each unloading point must at least meet the production demand of that unloading point;

Production capacity constraints for load points:

The total amount of ore loaded at each loading point shall not exceed the maximum production capacity of that loading point;

Ore grade requirements at the unloading point:

The grade deviation of ore unloaded at each unloading point shall not exceed the maximum grade deviation;

Traffic flow continuity constraints ensure that the incoming and outgoing traffic flow of each loading/unloading point is equal;

Remaining fuel constraints:

Self-driving trucks are no longer like traditional trucks. There is a driver who can keep an eye on the remaining fuel level at any time while driving. Therefore, when planning and dispatching driverless trucks for traffic flow, it is necessary to monitor their remaining fuel levels. Its specific manifestation is that the remaining fuel quantity of each truck must not be less than K in order to return to the refueling point for refueling. Among them, the minimum remaining oil amount K = the maximum oil consumption from the unloading point to the loading point + the maximum oil consumption from the loading point to the unloading point + the maximum oil consumption from the unloading point to the refueling point O;

The number of truck transportation requirements, the number of truck transportation must be a positive integer;

X _ijkl ,Y _ijkl ∈{0,1,2,3...}

Constraints in conflict avoidance situations;

When two trucks are at a certain loading or unloading point at the same time, the truck routes will conflict, which will have a greater impact on the truck's operating costs and queuing time. In order to reduce the impact caused by truck task conflicts as much as possible, it is necessary to set priorities for trucks under the constraints of the initial state. Given that fully loaded trucks are more expensive to operate than empty trucks, priority is given to fully loaded trucks. In addition, the priority determination also applies to when two trucks meet at the same intersection, and the unloaded driverless truck should give way.

Constraints in truck failure conditions;

Once a truck fails and cannot move, the truck should exit the traffic flow planning and dispatching system in a timely manner. Therefore, when a driverless truck alerts the general dispatch center, the truck should be immediately excluded from the system, and the priority of the path, loading point or unloading point where the truck is located should be reduced so that the planned dispatch plan can be adjusted in a timely manner. Achieve the company’s established goals;

The parameter symbols used in the above formula are defined as follows:

i: The index number of the loading point, indicating the i-th loading point (i.e., the excavator position), i = 1, 2,...,I,; j: The index number of the unloading point, indicating the j-th unloading point (i.e., crushing station location), j=1,2,...,J,; k: truck model index number, indicating the k-th model of truck, k=1,2,...,K, type; l: truck number index number, Represents the l-th k-type truck, l= _{1,2 ,} ...,L,; The number of times a k-type truck of l is empty from unloading point j to loading point i, times; d _ij : the distance from loading point i to unloading point j, km; d _jo : the distance from unloading point j to refueling point O, km ; d _oi , the optimal distance from refueling point O to loading point i, km; C _k : loading capacity of k-type truck, t; CE _kl : unit distance cost of k-type truck numbered l when overloaded, yuan /km; CL _kl : The unit distance cost of the k-type truck numbered l when empty, yuan/km; E _kl : The fuel tank capacity of the k-type truck numbered l, L; EE _kl : The k-type truck numbered l Fuel consumption per unit distance when the truck is heavily loaded, L/km; EL _kl , fuel consumption per unit distance when the k-type truck numbered l is unloaded, L/km; K: minimum remaining fuel volume, L; g _i : loading point i The maximum production capacity, t; f _j : the minimum production demand of unloading point j (within one shift), t; q _j : the maximum production capacity of unloading point j, t; e: the limit of ore grade; α _i : loading point The ore grade of i; β: the allowable error of the ore grade; G _j : the target grade of the unloading point j; T _lim : the shift working time; SE _kl : the average speed of the K-type truck numbered l when heavy load, km/h; SL _kl : the average speed of the K-type truck numbered l when empty, km/h; T _O , the average time for truck refueling, min; T _z : the average time for truck loading, min; T _q : the average time for truck unloading Time, min; K _jOkl : The number of runs of the k-type truck numbered l from the unloading point j to the refueling point O when empty, times; K _Oikl : The k-type truck numbered l from the refueling point O to the loading point when empty The number of runs of point i, times.

Step 4. Use the multi-objective particle swarm algorithm to solve the scheduling model: used to optimize the final scheduling plan.

Set the basic parameters of the multi-objective particle swarm algorithm such as inertia weight ω and flight speed c1, c2, and solve the open-pit mine driverless truck scheduling optimization model through the multi-objective particle swarm algorithm;

Determine whether the current solution (scheme) meets the three constraints of production capacity, grade requirements, and traffic flow control, and then update the position and speed of the particles; the above entire process is completed as one iteration. When the algorithm iteration is completed, the result is output and the calculation process ends. Otherwise, continue with the multi-objective particle swarm algorithm solution; the specific steps are as follows:

Step1: Initialize the particle swarm P; for each particle, determine its initial position and speed and calculate the target vector of each particle in the particle swarm P;

Step2: Save some particles in the particle swarm P in the external particle swarm NP, and the positions of these particles are non-inferior solutions;

Step3: Determine the best initial position of each particle, that is, the initial position of each particle itself;

Step 4: Divide the target space into many grids (hypercubes), and determine the grid where each particle is located based on the target vector corresponding to the particle;

Step5: Define the fitness value (the ratio of a number equal to or greater than 1 to the number of NP members contained in the grid) for each grid that contains at least one external particle swarm individual, and then for each particle, according to the roulette method Select a grid and randomly select an individual from the external particle group as the gbest of the particle;

Step6: Update the position and velocity of all particles according to the basic formula of PSO;

Step7: Use the following measures to prevent particles from flying out of the search space: Once a particle flies out of the boundary of a certain decision variable, the particle stays on the boundary and changes its flight direction at the same time;

Step8: Calculate the target vector of each particle in the particle swarm P;

Step9: Use the adaptive grid method to update and maintain the external particle swarm NP;

Step10: Update the pbest of the particles. According to the comparison between the new solution obtained during the particle flight and its own best position pbest, if the new solution dominates pbest, the new solution is the new pbest; otherwise, pbest remains unchanged; if the new solution and pbest are not dominated by each other, then Randomly select one of the two as the new best position;

Step11: If the termination condition is established, stop the search and determine whether the plan meets the three constraints of production capacity, grade limit and traffic flow control in the model; when the above conditions are met at the same time, output the optimization plan; otherwise, go to Step6.

In summary, this invention aims at the problem of unmanned truck scheduling optimization in open-pit mines considering truck performance, with the goal of minimizing truck waiting time, minimizing truck waiting time, minimizing remaining production, and minimizing ore grade deviation, taking into account the different characteristics of trucks. To improve the transportation efficiency, a driverless truck scheduling optimization model for open-pit mines was constructed. The multi-objective particle swarm algorithm is used for optimization and solution, which effectively solves the optimization problem of driverless truck dispatching in open-pit mines and meets the needs of actual mine transportation.

Claims (6)

1. A method for dispatching unmanned trucks in open-pit mines considering truck performance, which is characterized by including the following steps; Step 1: Establishment of open-pit mine truck operating status database; Step 2: Establishment of truck operating efficiency prediction model; Step 3: Establish a multi-objective open-pit mine driverless truck scheduling optimization model based on the open-pit mine truck operating status database and truck operating efficiency prediction model; Step 4: Use the multi-objective particle swarm algorithm to solve the scheduling optimization model for optimizing the final scheduling solution. 2. A method for dispatching unmanned trucks in open-pit mines that considers truck performance according to claim 1, characterized in that step 1 specifically includes: collecting the truck load and completed kilometers of each truck in open-pit mine production. It collects historical data on engine speed, engine condition, tire wear, repair and maintenance status, and the average speed of each shift, and establishes a database on the operating status of open-pit mine trucks; develops a statistics table on the operating status of open-pit mine trucks, and collects historical data on the operating status of open-pit mine trucks. 3. A method for dispatching unmanned trucks in open-pit mines that considers truck performance according to claim 1, characterized in that step 2 is specifically: through the analysis of factors affecting truck transportation performance, establishing a method based on truck load It is a prediction model that uses five influencing factors as input parameters, the number of kilometers traveled, the number of engine failures, the number of kilometers used by tires, and the number of daily repairs, and uses the average speed of each vehicle shift as the output parameter. 4. A method for dispatching unmanned trucks in open-pit mines that considers truck performance according to claim 1, characterized in that in step 3, the method is based on the actual production and transportation process of open-pit mines to minimize truck waiting time, minimum With the goals of minimizing truck waiting time, minimizing remaining production, and minimizing ore grade deviation, an open-pit mine driverless truck scheduling optimization model is established; the open-pit mine driverless truck scheduling optimization model includes an objective function and constraints. 5. A method for dispatching unmanned trucks in open pit mines considering truck performance according to claim 4, characterized in that the objective function is specifically: Minimize trucking costs: The transportation cost of a truck consists of the transportation cost of a heavy truck from the loading point to the unloading point, the transportation cost of an empty truck from the unloading point to the loading point, and the transportation cost of an empty truck to and from the refueling point; Minimize truck waiting time: The waiting time of a truck is composed of the truck shift working time minus the actual running time of the truck. The actual running time includes the running time between loading and unloading points, loading and unloading time, time to and from the gas station, and refueling time; Minimize remaining production: In order to ensure that the ore produced in one shift at the mining site can be transported as much as possible, that is, the arranged transportation capacity is as close as possible to the production capacity, the absolute value of the output minus the transportation volume is required to be minimum; Minimize ore grade deviation: The sum of the ore grade deviations at each unloading point is the smallest; The constraints are specifically: Production capacity constraints at the unloading point: The total amount of ore transported to each unloading point shall not exceed the maximum production capacity of that unloading point; Production planning requirements for unloading points: The total amount of ore unloaded at each unloading point must at least meet the production demand of that unloading point; Production capacity constraints for load points: The total amount of ore loaded at each loading point shall not exceed the maximum production capacity of that loading point; Ore grade requirements at the unloading point: The grade deviation of ore unloaded at each unloading point shall not exceed the maximum grade deviation; Traffic flow continuity constraints ensure that the incoming and outgoing traffic flow of each loading/unloading point is equal; Remaining fuel constraints: Minimum remaining oil amount K = maximum oil consumption from unloading point to loading point + maximum oil consumption from loading point to unloading point + maximum oil consumption from unloading point to refueling point O; The number of truck transportation requirements, the number of truck transportation must be a positive integer; X _ijkl ,Y _ijkl ∈{0,1,2,3...} Constraints in conflict avoidance situations; When two trucks are at a loading or unloading point at the same time, the fully loaded truck will go first. In addition, the priority determination also applies to when two trucks meet at the same intersection, and the empty driverless truck should give way; Constraints in truck failure conditions; When a driverless truck alerts the general dispatch center, the truck should be immediately excluded from the system and the priority of the truck's path, loading point or unloading point should be reduced so that the planning and dispatching plan can be adjusted in a timely manner to complete the enterprise's mission. stated goals; The parameter symbols used in the above formula are defined as follows: i: The index number of the loading point, indicating the i-th loading point (i.e., the excavator position), i = 1, 2,...,I,; j: The index number of the unloading point, indicating the j-th unloading point (i.e., crushing station location), j=1,2,...,J,; k: truck model index number, indicating the k-th model of truck, k=1,2,...,K, type; l: truck number index number, Represents the l-th k-type truck, l= _{1,2 ,} ...,L,; The number of times a k-type truck of l is empty from unloading point j to loading point i, times; d _ij : the distance from loading point i to unloading point j, km; d _jo : the distance from unloading point j to refueling point O, km ; d _oi , the optimal distance from refueling point O to loading point i, km; C _k : loading capacity of k-type truck, t; CE _kl : unit distance cost of k-type truck numbered l when overloaded, yuan /km; CL _kl : The unit distance cost of the k-type truck numbered l when empty, yuan/km; E _kl : The fuel tank capacity of the k-type truck numbered l, L; EE _kl : The k-type truck numbered l Fuel consumption per unit distance when the truck is heavily loaded, L/km; EL _kl , fuel consumption per unit distance when the k-type truck numbered l is unloaded, L/km; K: minimum remaining fuel volume, L; gi: loading point i Maximum production capacity, t; f _j : minimum production demand (within one shift) of unloading point j, t; q _j : maximum production capacity of unloading point j, t; e: limit of ore grade; α _i : loading point i ore grade; β: ore grade allowable error; G _j : target grade of unloading point j; T _lim : shift working time; SE _kl : average speed of K-type truck numbered l when heavy load, km/h; SL _kl : the average speed of the K-type truck numbered l when empty, km/h; T _O , the average time it takes for the truck to refuel, min; T _z : the average time it takes for the truck to load, min; T _q : the average time it takes for the truck to unload , min; K _jOkl : The number of runs of the k-type truck numbered l from the unloading point j to the refueling point O when empty, times; K _Oikl : The k-type truck numbered l from the refueling point O to the loading point when empty The number of runs of i, times. 6. A method for dispatching unmanned trucks in open-pit mines that considers truck performance according to claim 5, characterized in that step 4 is specifically: Set the basic parameters of the multi-objective particle swarm algorithm such as inertia weight ω and flight speed c1, c2, and solve the open-pit mine driverless truck scheduling optimization model through the multi-objective particle swarm algorithm; Determine whether the current solution satisfies the three constraints of production capacity, grade requirements, and traffic flow control, and then update the position and speed of the particles; the above entire process is completed as one iteration. When the algorithm iteration is completed, the result is output and the calculation process ends. Otherwise, continue Multi-objective particle swarm algorithm is used to solve the problem; the specific steps are as follows: Step1: Initialize the particle swarm P; for each particle, determine its initial position and speed and calculate the target vector of each particle in the particle swarm P; Step2: Save some particles in the particle swarm P in the external particle swarm NP, and the positions of these particles are non-inferior solutions; Step3: Determine the best initial position of each particle, that is, the initial position of each particle itself; Step 4: Divide the target space into many grids (hypercubes), and determine the grid where each particle is located based on the target vector corresponding to the particle; Step5: Define the fitness value (the ratio of a number equal to or greater than 1 to the number of NP members contained in the grid) for each grid that contains at least one external particle swarm individual, and then for each particle, according to the roulette method Select a grid and randomly select an individual from the external particle group as the gbest of the particle; Step6: Update the position and velocity of all particles according to the basic formula of PSO; Step7: Use the following measures to prevent particles from flying out of the search space: Once a particle flies out of the boundary of a certain decision variable, the particle stays on the boundary and changes its flight direction at the same time; Step8: Calculate the target vector of each particle in the particle swarm P; Step9: Use the adaptive grid method to update and maintain the external particle swarm NP; Step10: Update the pbest of the particles. According to the comparison between the new solution obtained during the particle flight and its own best position pbest, if the new solution dominates pbest, the new solution is the new pbest; otherwise, pbest remains unchanged; if the new solution and pbest are not dominated by each other, then Randomly select one of the two as the new best position; Step11: If the termination condition is established, stop the search and determine whether the plan meets the three constraints of production capacity, grade limit and traffic flow control in the model; when the above conditions are met at the same time, output the optimization plan; Otherwise, go to Step6.