Mine car optimization scheduling method and system based on improved PSO-ACO
The improved PSO-ACO method optimizes mine car scheduling by integrating environmental factors and constraints, enhancing dispatcher adaptability and algorithm convergence, thus improving efficiency and reducing carbon emissions.
Patent Information
- Application Number
- CN202411260113.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-10
- Publication Date
- 2024-12-03
AI Technical Summary
Existing mine car optimization scheduling methods fail to consider environmental factors such as time windows and carbon emission costs, are limited by dispatcher experience, struggle with complex environments, and face high complexity due to reliance on accurate data and slow convergence speeds of algorithms.
A method and system using an improved PSO-ACO combination to optimize mine car scheduling, incorporating environmental factors, dispatcher adaptability, and multiple constraints, with enhanced local search ability and convergence speed through a mathematical model and intelligent algorithms.
Improves scheduling efficiency, accuracy, and reduces carbon emissions by optimizing mine car operations, addressing complex environments and data complexity, and enhancing algorithm convergence.
Smart Images

Figure CN119067281A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of mine operation optimization scheduling and the field of smart mines. More specifically, it relates to a mine car optimization scheduling method and system based on improved particle swarm optimization (PSO)-improved ant colony optimization (ACO). Background Art
[0002] As the core transportation tool for minerals, the operating efficiency and energy consumption level of mine cars directly affect the overall operating cost and carbon emission reduction capacity of mines. During the entire transportation process, the transportation cost of mine cars accounts for about 30% to 40% of the total cost of mining ore. Its transportation energy consumption accounts for about 60% of the total energy consumption. Taking the Pingshuo open-pit mine as an example, the annual fuel consumption of a single mine car exceeds 500 tons, and the fuel consumption of the 930E model is as high as 800 tons. Considering that there are 200 such mine cars in the Pingshuo mine, the annual fuel consumption exceeds 100,000 tons. If calculated at a price of 8,000 yuan per ton of oil, the annual fuel consumption cost of a mining area will be as high as 800 million yuan. Taking the optimization of fuel consumption cost as an example, through effective scheduling strategies, the transportation cost can be reduced by 5%-10%, thereby bringing economic benefits of 40-80 million yuan. At present, there are widespread scheduling inefficiencies such as serious energy consumption of mine cars and unreasonable mine car transportation time arrangements, which not only increase transportation costs, but also lead to a significant increase in carbon emissions. Therefore, finding an efficient and accurate mine cart scheduling strategy is the key to reducing mine cart transportation costs and carbon emissions.
[0003] At present, the optimization scheduling methods of mine cars are mainly divided into the following three types, namely traditional manual scheduling, operations research scheduling, and intelligent optimization algorithm scheduling; the traditional manual scheduling method judges the current situation based on the dispatcher's experience and arranges the transportation tasks, routes and time of the mine cars based on this; operations research scheduling can fully consider the global state through mathematical models and optimization algorithms, so as to allocate transportation tasks on a global scale; the intelligent optimization algorithm adopts particle swarm algorithm or ant colony algorithm, and with its ability to handle complex problems and adaptability, it can dynamically adjust strategies to adapt to different parameter changes.
[0004] However, the above three existing mine car optimization scheduling technologies all have some problems that cannot be ignored: First, since these three mine car optimization scheduling methods do not fully consider environmental factors such as the time window of the mine car unloading point and the carbon emission cost, the mine car scheduling efficiency will be low; Second, for the above traditional manual scheduling method, it is severely limited by the differences in dispatcher experience and environmental adaptability, and in the face of complex environments such as parallel execution of multiple transportation tasks and time window constraints, it is difficult to meet the operations in complex mine environments. The requirements for accuracy, efficiency and low carbonization; Third, the above operations research scheduling method relies on accurate data input, and when dealing with large-scale data and multiple constraints, the complexity of this scheduling method is relatively high; Fourth, for the scheduling method of the above intelligent optimization algorithm, although the particle swarm algorithm can quickly approach the global optimal solution, its local search ability is insufficient; although the ant colony algorithm has a good positive feedback mechanism, its convergence speed is slow and it is easy to fall into the local optimal solution. Summary of the invention
[0005] In view of the above defects or improvement needs of the prior art, the present invention provides a method and system for optimizing the scheduling of mine cars based on an improved PSO-ACO combination. The purpose is to solve the technical problem that the existing optimization scheduling method does not fully consider environmental factors such as the time window of the unloading point of the mine car and the carbon emission cost, thereby resulting in low efficiency in the scheduling of mine cars, and the existing optimization scheduling method is severely limited by the differences in dispatcher experience and environmental adaptability, and in the face of complex environments such as parallel execution of multiple transportation tasks and time window constraints, it is difficult to meet the requirements of accuracy, efficiency and low carbonization in the complex environment of the mine, and the existing optimization scheduling method relies on accurate data input, and when processing large-scale data and multiple constraints, it leads to a high complexity. The technical problem that the particle swarm algorithm can quickly approach the global optimal solution, but its local search ability is insufficient, and the technical problem that the ant colony algorithm has a good positive feedback mechanism, but its convergence speed is slow and it is easy to fall into the local optimal solution.
[0006] To achieve the above object, according to one aspect of the present invention, a method for optimizing the scheduling of mine vehicles based on an improved PSO-ACO is provided, comprising the following steps:
[0007] (1) Obtaining basic operation scheduling data and basic working data of a mine car unloading point, and preprocessing the basic operation scheduling data and basic working data respectively to obtain preprocessed basic operation scheduling data and preprocessed basic working data respectively.
[0008] (2) Obtaining comprehensive parameters of mine car scheduling and operation, and obtaining the time parameters of the mine car arriving at the unloading point based on the basic operation scheduling data preprocessed in step (1) and the basic working data preprocessed.
[0009] (3) constructing a mathematical model for minimizing the total transportation cost of all mine cars based on the comprehensive parameters of mine car scheduling and operation obtained in step (2) and the time parameters of the mine cars arriving at the unloading point;
[0010] (4) Add multiple constraints to the mathematical model for minimizing the total transportation cost of all mine carts obtained in step (3) to obtain an updated mathematical model for minimizing the total transportation cost of all mine carts.
[0011] (5) Using the improved PSO algorithm and the updated mathematical model of minimizing the total transportation cost of all mine carts obtained in step (4), the mathematical model of minimizing the total transportation cost of all mine carts obtained in step (3) is converted into a fitness evaluation function of the improved PSO algorithm, and the fitness value of each particle in the particle swarm is obtained according to the fitness evaluation function to obtain the minimum total transportation cost of all mine carts.
[0012] (6) Convert the minimum total transportation cost of all mine carts obtained in step (5) into the initial pheromone enhancement amount of the improved ACO algorithm.
[0013] (7) Using the improved ACO algorithm and the initial pheromone increment of the improved ACO algorithm obtained in step (6), the minimum total transportation cost of all mine carts is obtained.
[0014] Preferably, the basic operation scheduling data of the unloading point of the mine car includes: the unloading point location information data of the mine car and the unloading point time window constraint data of the mine car;
[0015] The basic working data of the mine car include: the travel speed of the nth mine car and the time when the nth mine car arrives at the unloading point u.
[0016] In step (1), the basic operation scheduling data of the unloading point of the mine car is preprocessed to obtain the preprocessed basic operation scheduling data. In this process, the position information data of the unloading point and the departure point of the mine car and the time window constraint data are first obtained; then, the position information data of the unloading point and the departure point of the mine car are converted into accurate coordinate data, and the obtained coordinate data is expressed in the two-dimensional plane as coordinates (x u ,y u ) form, and then, according to the coordinates (x u ,y u), and use the Euclidean distance to calculate the Euclidean distance between the departure point of the nth mine car and each unloading point, as well as the Euclidean distance between every two unloading points of the nth mine car. Then, the time window constraint data is converted into numerical time window constraint data in the form of hours and minutes. Then, the numerical time window constraint data is normalized. Finally, all time windows in the normalized time window constraint data are sorted in chronological order to obtain sorted time window constraint data. All Euclidean distances and sorted time window constraint data together constitute the preprocessed basic operation scheduling data, where x u Represents the horizontal coordinate of the u-th unloading point among all the unloading points of the mine car, y u Represents the ordinate of the u-th unloading point among all the unloading points of the mine car;
[0017] In step (1), the basic working data of the mine car is preprocessed to obtain the preprocessed basic working data of the mine car. Specifically, according to the travel speed of the nth mine car and the obtained Euclidean distance between the departure point of the nth mine car and each unloading point, and the Euclidean distance between every two unloading points of the nth mine car, the transportation time t between the departure point of the nth mine car and each unloading point is calculated. 1u =d 1u / v c , and the transportation time t between each two unloading points of the nth mine car uv =d uv / v c , all the transportation time obtained is the basic working data after preprocessing, among which, t 1u represents the transportation time between the departure point of the mine car and each unloading point u, d 1u represents the Euclidean distance between the departure point and the unloading point u, v c Indicates the current speed of the minecart, d uv represents the Euclidean distance between unloading point u and unloading point v, t uv Indicates the transportation time between every two unloading points of the minecart.
[0018] Preferably, the comprehensive parameters of mine car scheduling and operation obtained in step (2) include: the set of all mine cars located at the departure point, the set of unloading points, the departure point, the number of mine cars used, the opportunity cost of the nth mine car arriving earlier than the time window, the penalty cost of the nth mine car arriving later than the time window, the time the nth mine car waits for loading at the unloading point u, the unloading time of the nth mine car at the unloading point v, the time when the nth mine car arrives at the unloading point u, and the fixed transportation cost C of the nth mine car s, the fuel consumption price per unit distance of the nth mine car, the fuel consumption per unit distance of the nth mine car in an empty state, the fuel consumption per unit distance of the nth mine car in a fully loaded state, the real-time loading capacity of the nth mine car, the maximum load capacity of the nth mine car, the carbon emission coefficient of the nth mine car, the carbon emission cost per unit distance of the nth mine car, the number of unloading points served by the nth mine car, the demand for ore at the unloading points, the load of the nth mine car from the unloading point u to the unloading point v, and the usage status parameter S of the nth mine car n , the driving parameters of the nth vehicle on the road section from unloading point u to unloading point v , the driving parameters of the nth mine car on the road section from the departure point to the unloading point u , and n∈[1, the total number of minecarts N].
[0019] The time when the nth mine car arrives at the unloading point v obtained in step (2) is expressed by the following formula:
[0020] T v =T u +t uv +w u +w v ,v∈M
[0021] Where T v represents the time when the nth mine car arrives at the unloading point v; T u represents the time when the nth mine car arrives at the unloading point u; w u represents the time that the nth mine car waits and loads at the unloading point u; w v represents the unloading time of the nth mine car at the unloading point v, and M represents the set of all unloading points.
[0022] Preferably, the total transportation cost of the mine carts in step (3) includes: the fixed cost of all mine carts, the opportunity cost and penalty cost of all mine cart delivery time windows, the fuel consumption cost of all mine carts, and the carbon emission cost of all mine carts;
[0023] The process of constructing the mathematical model for minimizing the total transportation cost of the mine cart in step (3) includes the following sub-steps:
[0024] (3-1) According to the fixed transportation cost C of the nth mine car obtained in step (2) s And the usage status parameter S of the nth mine car n , obtain the various fixed costs of the mine carts, and systematically integrate the obtained costs of the mine carts to obtain the fixed costs of all mine carts during the transportation process.
[0025] Specifically, this step includes firstly calculating the fixed transportation cost of the nth mine car and the usage state parameter S of the nth mine car obtained in step (2). n, obtain the various fixed transportation costs of the nth mine car when performing the transportation task, including the depreciation cost of the mine car, daily loss cost, fixed loading and unloading cost, and maintenance cost. Then, all fixed transportation costs are systematically integrated to obtain the fixed cost of each mine car. Finally, the fixed cost C1 of all mine cars in the transportation process is constructed according to the fixed cost of the mine car. The specific expression is shown in the following formula:
[0026]
[0027] (3-2) According to the time when the nth mine car arrives at the unloading point u obtained in step (2), the opportunity cost and penalty cost of all mine car delivery time windows are obtained.
[0028] Specifically, this step includes: first, comparing the time when the nth mine car arrives at the unloading point u obtained in step (2) with the time window of the unloading point obtained in step (1); then, based on the comparison result and the opportunity cost of the nth mine car arriving earlier than the time window and the penalty cost of the nth mine car arriving later than the time window obtained in step (2), the opportunity cost and penalty cost distribution C2(u) of the nth mine car arriving at the unloading point u are obtained, as shown in the following formula.
[0029]
[0030] Among them B u E represents the earliest acceptable transportation time for the unloading point u; u represents the latest acceptable transportation time for the unloading point u; A represents the opportunity cost of the nth mine car arriving earlier than the time window; B represents the penalty cost of the nth mine car arriving later than the time window (the opportunity cost and penalty cost of all mine cars are the same).
[0031] Finally, according to the opportunity cost and penalty cost distribution of the nth mine car arriving at the unloading point u, the opportunity cost and penalty cost C2 of all mine car delivery time windows are obtained as follows:
[0032]
[0033] (3-3) using the load estimation method to process the unit distance fuel consumption of the nth mine car in the empty state, the unit distance fuel consumption of the nth mine car in the fully loaded state, the real-time load of the nth mine car, and the maximum load of the nth mine car obtained in step (2) to obtain the fuel consumption rate of the nth mine car when the load is q; and according to the obtained fuel consumption rate, the Euclidean distance between the departure point of the nth mine car and each unloading point and the Euclidean distance between every two unloading points obtained by preprocessing in step (1), and the fuel consumption price per unit distance of the nth mine car obtained in step (2), obtain the fuel consumption cost of all mine cars;
[0034] Specifically, this step is as follows: first, according to the unit distance fuel consumption of the nth mine car in an empty state, the unit distance fuel consumption of the nth mine car in a fully loaded state, the real-time load of the nth mine car, and the maximum load of the nth mine car obtained in step (2), the unit distance fuel consumption of the nth mine car under the current load is calculated using the load estimation method, and the fuel consumption rate R(q) of the nth mine car when the load is q is obtained. The specific calculation formula is shown as follows:
[0035]
[0036] Where Q L It represents the maximum load of the nth minecart (the maximum load of all minecarts is the same); R1 represents the fuel consumption per unit distance of the nth minecart in the empty state; R2 represents the fuel consumption per unit distance of the nth minecart in the heavy-loaded state (the fuel consumption per unit distance of all minecarts in the empty state and the heavy-loaded state are the same).
[0037] Then, according to the calculated fuel consumption rate of the nth mine car when the load is q, and the Euclidean distance d between the departure point of the nth mine car and each unloading point obtained by preprocessing in step (1), 1u and the Euclidean distance d between every two unloading points uv , obtain the fuel consumption F of the nth mine car in transporting between the departure point and each unloading point when the load is q 1v (q), and the fuel consumption F of the nth ore car in transporting between every two unloading points uv (q), and its specific calculation formula is shown below.
[0038]
[0039] Then, based on the obtained fuel consumption of the nth mine car between the departure point and each unloading point and the fuel consumption between every two unloading points when the load is q, and the fuel consumption price per unit distance of the nth mine car obtained in step (2), f Get the fuel consumption cost C3(q) of all minecarts when the load is q, as shown in the following formula:
[0040]
[0041] Then, replace q / Q in the above formula L Converted to the load factor of a minecart To obtain the simplified fuel consumption cost of all mine cars;
[0042] Then, based on the simplified fuel consumption cost of all mining vehicles and the driving parameters of the nth mining vehicle on the road section from the departure point to the unloading point v obtained in step (2), and the driving parameters of the nth vehicle on the road section from unloading point u to unloading point v Get the fuel consumption cost C3 of all minecarts as shown below:
[0043]
[0044] (3-4) Based on the fuel consumption of the nth mine car in transporting between the departure point and each unloading point and the fuel consumption between every two unloading points when the load is q obtained in step (3-3), and the carbon emission coefficient and unit distance carbon emission cost of the nth mine car obtained in step (2), the carbon emission cost of all mine cars is obtained.
[0045] Specifically, this step includes: first, according to the fuel consumption of the nth mine car in transporting between the departure point and each unloading point and the fuel consumption between every two unloading points when the load is q obtained in step (3-3) and the carbon emission coefficient of the nth mine car obtained in step (2), the carbon emissions of the nth mine car during transportation are obtained; then, according to the obtained carbon emissions of the nth mine car during transportation and the carbon emission cost per unit distance of the nth mine car obtained in step (2), the carbon emission cost C4 of all mine cars is obtained, as shown in the following formula.
[0046]
[0047] Where γ represents the carbon emission cost per unit distance of the nth mine car; ε represents the carbon emission coefficient of the nth mine car (the carbon emission cost per unit distance and the carbon emission coefficient of all mine cars are the same).
[0048] Finally, according to the mine car load rate And the carbon emission costs of all mine carts, to obtain the final carbon emission cost C4 of all mine carts, the specific form of which is shown in the following formula:
[0049]
[0050] (3-5) The fixed costs of all mine carts obtained in step (3-1), the opportunity costs and penalty costs of all mine cart delivery time windows obtained in step (3-2), the fuel consumption costs of all mine carts obtained in step (3-3), and the carbon emission costs of all mine carts obtained in step (3-4) are integrated to obtain a mathematical model for minimizing the total transportation cost of all mine carts, as shown in the following formula.
[0051] minC=C1+C2+C3+C4
[0052] The fixed cost of all mine cars obtained in step (3-1), the opportunity cost and penalty cost of all mine car delivery time windows obtained in step (3-2), the fuel consumption cost of all mine cars obtained in step (3-3), and the carbon emission cost of all mine cars obtained in step (3-4) are substituted into the above-mentioned total transportation cost minimization mathematical model to obtain the following expression:
[0053]
[0054] Preferably, the multiple constraints added in step (4) are:
[0055] A. The number of minecarts used cannot exceed the total number of minecarts N, as shown in the following formula.
[0056]
[0057] B. Each mining car needs to start from the departure point and return to the departure point after completing the transportation task, as shown in the following formula.
[0058]
[0059] C. Each unloading point requires and only requires one mine car to provide transportation services, as shown in the following formula.
[0060]
[0061] D. The number of mine cars arriving at the unloading point is the same as the number of mine cars leaving the unloading point, as shown in the following formula.
[0062]
[0063] E. The load of each mine car cannot exceed its maximum load capacity, as shown in the following formula.
[0064]
[0065] Among them, q u Represents the demand for ore at unloading point u.
[0066] F. The total loading capacity of all mine cars used must meet (that is, be greater than or equal to) the total amount of ore that needs to be transported at all unloading points served by all mine cars, as shown in the following formula.
[0067]
[0068] in, Represents the load of the nth mine car from the departure point to the unloading point u.
[0069] G. The time window constraint for the mine car to arrive at the unloading point it serves is shown in the following formula.
[0070]
[0071] H. Ensure that the transportation task of each mine car is a continuous process, as shown in the following formula.
[0072]
[0073] I. The sub-circuit of the minecart eliminates constraints as shown below.
[0074]
[0075] Among them, |K n | represents the number of unloading points served by the nth minecart.
[0076] J. Parameter S n , The constraints on the value range of are shown in the following formula.
[0077]
[0078] Preferably, step (5) comprises the following sub-steps:
[0079] (5-1) Initialize the particle swarm parameters of the improved PSO algorithm to obtain the initialized particle swarm parameters.
[0080] (5-2) Set counter k=0.
[0081] (5-3) Determine whether k is greater than the preset maximum number of iterations Num, if so, proceed to step (5-13), otherwise proceed to step (5-4);
[0082] (5-4) Based on the initialized particle swarm parameters obtained in step (5-1) and the multiple constraints added in step (4), the mathematical model of minimum total transportation cost of all mine cars constructed in step (3) is used as the fitness evaluation function of the improved PSO algorithm, where the fitness evaluation function F(C) is shown in the following formula.
[0083] F(C)=minC
[0084] (5-5) According to the fitness evaluation function of the improved PSO algorithm obtained in step (5-4) and the fitness value of each particle i in the particle swarm at its initial position, obtain the individual historical optimal fitness value of particle i at the kth iteration The number of particles in the particle swarm NP].
[0085] (5-6) Select the minimum value from the individual historical optimal fitness values of all particles obtained in step (5-5) as the global historical optimal fitness value Gbest at the kth iteration k(At this time, the number of iterations k = 0).
[0086] (5-7) Obtain the velocity vector and position vector of each particle in the particle swarm at the kth iteration according to the initialized particle swarm parameters obtained in step (5-1), and calculate the velocity vector and position vector of the particle at the k+1th iteration according to the velocity vector and position vector of the particle at the kth iteration.
[0087] (5-8) According to the speed limit and position limit of each particle in the particle swarm preset in step (5-1), the speed vector and position vector of the particle obtained in step (5-7) at the k+1th iteration are screened to obtain a plurality of screened particles.
[0088] (5-9) The fitness evaluation function obtained in step (5-4) is used to recalculate the fitness value of each particle screened in step (5-8) to obtain the fitness value of the particle at the k+1th iteration.
[0089] (5-10) The individual historical optimal fitness value of the particle at the kth iteration calculated according to step (5-5) The fitness value of each particle at the k+1th iteration calculated by step (5-9) is used to obtain the individual historical optimal fitness value of each particle in the particle swarm at the k+1th iteration after adjustment.
[0090] (5-11) The global historical optimal fitness value Gbest at the kth iteration calculated according to step (5-6) k And the individual historical optimal fitness value of each particle at the k+1th iteration obtained in step (5-10) Get the global historical optimal fitness value Gbest of each particle in the particle swarm at the k+1th iteration after adjustment k+1 :
[0091] (5-12) Set k=k+1 and return to step (5-3).
[0092] (5-13) Output the global historical optimal fitness value as the minimum total transportation cost of all mine carts.
[0093] Preferably, in step (5-1), the particle swarm parameters of the improved PSO algorithm are initialized as follows: setting the initial velocity vector and initial position vector of the particles in the particle swarm to random values; setting the number of particles in the particle swarm NP = 50; setting the speed limit of the particles in the particle swarm in the search space [v b ,v s]=[-1,1]; Set the position limit of particles in the particle swarm in the search space [x b ,x s ]=[-1,1], where x b represents the maximum limit position of particles in the particle swarm, x s represents the minimum restricted position of particles in the particle swarm, v b represents the maximum speed limit of particles in the particle swarm, v s Indicates the minimum speed limit of particles in the particle swarm; sets the inertia weight, individual learning factor and group learning factor of particles in the particle swarm as random perturbation adaptive adjustment mode.
[0094] The velocity update formula and position update formula of each particle i in the particle swarm at the k+1th iteration in step (5-7) are as follows:
[0095]
[0096] Where ω represents the inertia weight, and ω=-0.4k / K+(1 / n)·randnd(), randnd() represents a random value that obeys the normal distribution; c1, c2 represent the individual learning factor and the group learning factor, c1=-2k / K+(1 / n)·randnd(), c2=2k / K+n·randnd(); r1 and r2 represent random numbers between the interval [0,1]; represents the velocity vector of particle i at the k+1th iteration; represents the velocity vector of particle i at the kth iteration; Gbest represents the individual historical optimal fitness value of particle i at the kth iteration; k Represents the global historical optimal fitness value at the kth iteration; represents the position vector of particle i at the k+1th iteration; represents the position vector of particle i at the kth iteration.
[0097] Specifically, step (5-8) firstly screens the velocity vector of each particle obtained in step (5-7) at the k+1th iteration according to the velocity limit of the particles in the particle group preset in step (5-1), thereby obtaining a first group of screened particles.
[0098]
[0099] Then, the position vector of each particle obtained in step (5-7) at the k+1th iteration is screened according to the position restriction of the particles in the particle group preset in step (5-1) (the specific expression is shown in the following formula), thereby obtaining a second group of screened particles.
[0100]
[0101] Finally, the first group of particles and the second group of particles obtained above are combined to obtain the screened particles.
[0102] Preferably, in step (6), the specific calculation formula for converting the minimum total transportation cost of all mine carts obtained in step (5) into the initial pheromone enhancement amount of the improved ACO algorithm is as follows:
[0103] τ(0)=τ Β (0)+Δτ P
[0104] Among them, τ(0) represents the initial pheromone distribution value; τ Β (0) represents the initial value of pheromone, τ Β (0) = 0; Δτ P Indicates the conversion of the minimum total transportation cost of all mine carts obtained in the improved PSO algorithm into the increment of the pheromone intensity of the ACO algorithm, Δτ p =Gbest k+1 , Gbest k+1 is the global historical optimal fitness value obtained at the k+1th iteration obtained in step (5), and k+1=Num.
[0105] Preferably, step (7) specifically includes the following sub-steps:
[0106] (7-1) Initialize the ant colony related parameters in the improved ACO algorithm, and use the improved ACO algorithm initial pheromone increment obtained in step (6) as the initial value of the pheromone in the improved ACO algorithm to obtain the initialized ant colony parameters.
[0107] Specifically, the step is as follows: first, set the number of ants in the ant colony R = 50; set the time window heuristic factor λ = 1; set the pheromone heuristic factor α∈[0,5], preferably 1; set the expected heuristic factor β∈[0,10], preferably 7; set the pheromone volatility coefficient The constant C0∈[0,1] is preferably 0.5. Then, the initial pheromone increment of the improved ACO algorithm obtained in step (6) is used as the initial value of the pheromone in the improved ACO algorithm.
[0108] (7-2) Set counter NC=1.
[0109] (7-3) Determine whether NC is greater than the preset maximum number of iterations NC max (The maximum number of iterations NC max The value range of is 100 to 500, preferably 200), if yes, go to step (7-15), otherwise, go to step (7-4).
[0110] (7-4) Set the current number of ants r = 1.
[0111] (7-5) Determine whether r is greater than the number of ants R in the ant colony preset in the ant colony parameters after initialization in step (7-1). If so, proceed to step (7-11); otherwise, proceed to step (7-6).
[0112] (7-6) Place the rth ant in the ant colony at the departure point of the mine cart, and create a taboo table to record each unloading point reached by the rth ant in this cycle, add the departure point of the mine cart to the taboo table, and add multiple unloading points of the mine cart that the rth ant has not yet reached to the preset unloading point set m r middle.
[0113] (7-7) Determine whether the rth ant has reached the preset unloading point set m r If yes, go to step (7-10), otherwise go to step (7-8).
[0114] (7-8) Combine the time window constraint data of the unloading point of the mine car in the preprocessed basic operation scheduling data obtained in step (1) with the state transition probability formula of the traditional ACO algorithm to obtain an improved state transition probability formula, and obtain the next unloading point that the rth ant will reach according to the improved state transition probability formula, and add the next unloading point to the taboo table, and then enter step (7-9).
[0115] In this step, the time window constraint data of the unloading point of the mine car obtained in step (1) is combined with the state transition probability formula of the traditional ACO algorithm to obtain an improved state transition probability formula, the specific expression of which is shown in the following formula.
[0116]
[0117] in represents the probability selection function of the rth ant at time t choosing the unloading point v of the mine car at the unloading point u of the mine car as the next node; τ uv (t) represents the pheromone concentration on the transport path (u, v) at time t; η uv (t) represents the expected degree of the ant from the unloading point u of the mine car to the unloading point v of the target mine car at time t, η uv =1 / c uv ;c uv represents the transportation cost required for the ant to travel from the unloading point u of the mine cart to the unloading point v of the mine cart; φ uv (t) represents the relative coincidence value of the time window of the unloading point of the mine car at time t; p represents a random number between the interval [0,1];
[0118] (7-9) Obtain the current position information of the rth ant according to the next unloading point that the rth ant will reach in step (7-8), and update the unloading point set according to the current position information of the rth ant to obtain an updated unloading point set, and return to step (7-7).
[0119] (7-10) Set r=r+1 and return to step (7-5).
[0120] (7-11) Obtain the path formed by all the unloading points reached by each ant, and calculate the transportation cost of the corresponding path based on the mathematical model of minimizing the total transportation cost of all mine carts obtained in step (3), and then select the minimum value from the transportation costs corresponding to the paths generated by all ants as the transportation cost pbest(t) corresponding to the path generated by the optimal ant at the NC-th iteration at time t, where if the NC-th iteration is the first iteration, then set the transportation cost corresponding to the path generated by the optimal ant at the NC-th iteration at time t to the transportation cost gbest(t) corresponding to the path generated by the global optimal ant at the NC-th iteration at time t, and then enter step (7-12).
[0121] (7-12) Based on the transportation cost pbest(t) corresponding to the path generated by the optimal ant at the NC-th iteration at time t obtained in step (7-11), the transportation cost gbest(t-1) corresponding to the path generated by the global optimal ant at the NC-th iteration at time t-1 is updated to obtain the transportation cost gbest(t) corresponding to the path generated by the global optimal ant at the NC-th iteration at time t, and then enter step (7-13).
[0122] The transportation cost update formula corresponding to the path generated by the global optimal ant at the NC-th iteration at time t is as follows:
[0123]
[0124] (7-13) Determine whether NC is less than or equal to 20. If so, update the pheromone concentration on the path at the NC-th iteration at time t according to the transportation cost pbest(t) corresponding to the path generated by the optimal ant at the NC-th iteration at time t obtained in step (7-11) to obtain the pheromone concentration on the path at the NC-th iteration at time t+1. Otherwise, update the pheromone concentration on the path at the NC-th iteration at time t according to the transportation cost gbest(t) corresponding to the path generated by the global optimal ant at the NC-th iteration at time t obtained in step (7-12) to obtain the pheromone concentration on the path at the NC-th iteration at time t+1, and then enter step (7-14).
[0125] The formula for updating the pheromone concentration on the path at the NC-th iteration at time t+1 is as follows:
[0126]
[0127] where Δτ uv (t) represents the pheromone increment on the path from unloading point u to unloading point v at time t; Δτ uv (t) = 1 / b t , where b t The values are as follows:
[0128]
[0129] (7-14) Set NC=NC+1, clear the taboo table, and return to step (7-3).
[0130] (7-15) Output the NC at time t+1 max The transportation cost gbest(t+1) corresponding to the path generated by the global optimal ant at the iteration is taken as the minimum total transportation cost of all mine carts.
[0131] According to another aspect of the present invention, there is provided a mine car optimization scheduling system based on improved PSO-ACO, comprising:
[0132] The first module is used to obtain basic operation scheduling data and basic work data of the unloading point of the mine car, and preprocess the basic operation scheduling data and basic work data respectively to obtain preprocessed basic operation scheduling data and preprocessed basic work data respectively.
[0133] The second module is used to obtain the comprehensive parameters of mine car scheduling and operation, and obtain the time parameters of the mine car arriving at the unloading point based on the basic operation scheduling data and the basic work data preprocessed by the first module.
[0134] The third module is used to construct a mathematical model for minimizing the total transportation cost of all mine cars based on the comprehensive parameters of mine car scheduling and operation obtained in the second module and the parameters of the time when the mine cars arrive at the unloading point;
[0135] The fourth module is used to add multiple constraints to the mathematical model for minimizing the total transportation cost of all mine carts obtained in the third module to obtain an updated mathematical model for minimizing the total transportation cost of all mine carts.
[0136] The fifth module is used to use the improved PSO algorithm and the updated mathematical model of minimizing the total transportation cost of all mine carts obtained in the fourth module to convert the mathematical model of minimizing the total transportation cost of all mine carts obtained in the third module into a fitness evaluation function of the improved PSO algorithm, and obtain the fitness value of each particle in the particle swarm according to the fitness evaluation function to obtain the minimum total transportation cost of all mine carts.
[0137] The sixth module is used to convert the minimum total transportation cost of all mine carts obtained in the fifth module into the initial pheromone enhancement amount of the improved ACO algorithm.
[0138] The seventh module is used to obtain the minimum total transportation cost of all mine carts using the improved ACO algorithm and the initial pheromone increment of the improved ACO algorithm obtained in the sixth module.
[0139] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:
[0140] (1) Since the present invention adopts steps (1) to (4), a mathematical model for minimizing the total transportation cost of mine cars is established, which includes the fixed costs of all mine cars, the opportunity costs and penalty costs of all mine car delivery time windows, the fuel consumption costs of all mine cars, and the carbon emission costs of all mine cars, and multiple constraints are added to the mathematical model for minimizing the total transportation cost of all mine cars obtained in step (3), thereby obtaining an updated mathematical model for minimizing the total transportation cost of all mine cars, and realizing a comprehensive evaluation of the total transportation cost of mine cars. Therefore, it is possible to solve the technical problem that the existing optimization scheduling method does not fully consider environmental factors such as the time window of the mine car unloading point and the carbon emission cost, thereby resulting in low efficiency in scheduling mine cars.
[0141] (2) Since the present invention adopts steps (3) to (7), by incorporating the opportunity cost and penalty cost of all mine car delivery time windows and the carbon emission cost of all mine cars into the calculation of the total transportation cost of the mine cars, and using the improved PSO algorithm and ACO algorithm in the intelligent optimization algorithm to solve the mathematical model of minimizing the total transportation cost of all mine cars obtained in step (3), it is able to solve the technical problems that the traditional manual scheduling method is severely limited by the differences in dispatcher experience and environmental adaptability, and is difficult to meet the requirements of accuracy, efficiency and low carbonization for operations in complex mine environments under complex environments such as parallel execution of multiple transportation tasks and time window constraints.
[0142] (3) Since the present invention adopts step (4), multiple constraints are added to the mathematical model for minimizing the total transportation cost of all mine cars obtained in step (3), thereby obtaining an updated mathematical model for minimizing the total transportation cost of all mine cars. Therefore, it is possible to solve the technical problem that the existing optimization scheduling method relies on accurate data input and has a high complexity when processing large-scale data and multiple constraints.
[0143] (4) Since the present invention adopts steps (5) to (7), the improved PSO algorithm and the improved ACO algorithm are used to achieve the optimization of minimizing the total transportation cost of the mine carts. By converting the minimum total transportation cost of all mine carts obtained by the improved PSO into the initial pheromone enhancement amount of the improved ACO algorithm, the convergence speed of the traditional ACO algorithm is improved and it is avoided that it is easy to fall into the local optimal solution. At the same time, the local search ability of the PSO algorithm is enhanced by adopting the positive feedback mechanism of the improved ACO algorithm. Therefore, it is possible to solve the technical problem that the time complexity of the traditional ACO algorithm is too high, resulting in a slow convergence speed of the algorithm, and the technical problem that the initial pheromone distribution is insufficient, resulting in a long initial convergence time. BRIEF DESCRIPTION OF THE DRAWINGS
[0144] Figure 1 It is a flow chart of the present invention based on the improved PSO-ACO mine car optimization scheduling method;
[0145] Figure 2 is a schematic diagram of a mathematical model for minimizing the total transportation cost of all mining vehicles constructed in step (3) of the method of the present invention;
[0146] Figure 3 is a detailed flow chart of step (5) in the method of the present invention;
[0147] Figure 4 It is a detailed flow chart of step (7) in the method of the present invention. DETAILED DESCRIPTION
[0148] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.
[0149] The present invention proposes a basic improved PSO-ACO mine car optimization scheduling method. The method completes the optimization scheduling of mine cars from three aspects. First, based on the actual transportation situation of the mine, a mathematical model for minimizing the total transportation cost of the mine car including vehicle fixed cost, time window cost, fuel consumption cost, carbon emission cost, and various constraints is established. Subsequently, the above model is used as the fitness evaluation function of the PSO algorithm, and the group learning factor, individual learning factor, and inertia weight of the PSO algorithm are randomly disturbed and adaptively adjusted according to the number of iterations and the vehicle scale. Then, the improved PSO algorithm is used to solve the mathematical model for minimizing the total transportation cost of the mine car. Finally, in order to address the problems of slow convergence speed of the ACO algorithm, easy to fall into local optimal solutions and insufficient initial pheromone distribution, the optimal solution obtained by the improved PSO algorithm is converted into the enhancement of the initial information of the ACO algorithm, and the state transition rules and pheromone update methods of the ACO algorithm are improved in combination with the actual transportation conditions of the mine. The improved ACO algorithm is then used to solve the mathematical model of minimizing the total transportation cost of mine cars, and the improved PSO algorithm and the improved ACO algorithm are jointly solved, thereby enhancing the stability and robustness of the algorithm and obtaining the transportation plan with the lowest transportation cost, thereby realizing the optimal scheduling of mine cars in the mining area.
[0150] like Figure 1 As shown, the present invention provides a method for optimizing the scheduling of mine vehicles based on an improved PSO-ACO, comprising the following steps:
[0151] (1) Obtaining basic operation scheduling data and basic working data of a mine car unloading point, and preprocessing the basic operation scheduling data and basic working data respectively to obtain preprocessed basic operation scheduling data and preprocessed basic working data respectively.
[0152] Among them, the basic operation scheduling data of the mine car's unloading point in this step includes: the mine car's unloading point location information data and the mine car's unloading point time window constraint data; the mine car's basic working data includes: the driving speed of the nth mine car and the time when the nth mine car arrives at the unloading point u.
[0153] Specifically, in this step, the basic operation scheduling data of the unloading point of the mine car is preprocessed to obtain the preprocessed basic operation scheduling data. First, the position information data of the unloading point and the departure point of the mine car and the time window constraint data are obtained; then, the position information data of the unloading point and the departure point of the mine car are converted into accurate coordinate data, and the obtained coordinate data is expressed in the two-dimensional plane as coordinates (x u ,y u ) in the form of visual marking (where x u Represents the horizontal coordinate of the u-th unloading point among all the unloading points of the mine car, y urepresents the ordinate of the uth unloading point among all the unloading points of the mine car), then, according to the coordinate (x u ,y u ), and use Euclidean distance to calculate the Euclidean distance between the departure point of the nth mine car and each unloading point, as well as the Euclidean distance between every two unloading points of the nth mine car. Then, the time window constraint data is converted into numerical time window constraint data in the form of hours and minutes. Subsequently, the numerical time window constraint data is normalized (the purpose is to eliminate the difference in the value range of the time window). Finally, all time windows in the normalized time window constraint data are sorted in chronological order to obtain the sorted time window constraint data. All Euclidean distances and sorted time window constraint data together constitute the pre-processed basic operation scheduling data.
[0154] In this step, the basic working data of the mine car is preprocessed to obtain the basic working data of the mine car after preprocessing. This process is specifically as follows: according to the travel speed of the nth mine car and the obtained Euclidean distance between the departure point of the nth mine car and each unloading point, and the Euclidean distance between every two unloading points of the nth mine car, the transportation time t between the departure point of the nth mine car and each unloading point is calculated. 1u =d 1u / v c (where t 1u represents the transportation time between the departure point of the mine car and each unloading point u, d 1u represents the Euclidean distance between the departure point and the unloading point u, v c represents the current speed of the mine car), and the transportation time t between every two unloading points of the nth mine car uv =d uv / v c (where d uv represents the Euclidean distance between unloading point u and unloading point v, t uv Represents the transportation time between every two unloading points of the mine car), and all the transportation times obtained are the basic working data after preprocessing.
[0155] (2) Obtaining comprehensive parameters of mine car scheduling and operation, and obtaining the time parameters of the mine car arriving at the unloading point based on the basic operation scheduling data preprocessed in step (1) and the basic working data preprocessed.
[0156] Specifically, the comprehensive parameters of mine car scheduling and operation obtained in this step include: the set of all mine cars at the departure point, the set of unloading points, the departure point, the number of mine cars used, the opportunity cost of the nth mine car arriving earlier than the time window, the penalty cost of the nth mine car arriving later than the time window, the time the nth mine car waits for loading at the unloading point u, the unloading time of the nth mine car at the unloading point v, the time when the nth mine car arrives at the unloading point u, and the fixed transportation cost C of the nth mine car s , the fuel consumption price per unit distance of the nth mine car, the fuel consumption per unit distance of the nth mine car in an empty state, the fuel consumption per unit distance of the nth mine car in a fully loaded state, the real-time loading capacity of the nth mine car, the maximum load capacity of the nth mine car, the carbon emission coefficient of the nth mine car, the carbon emission cost per unit distance of the nth mine car, the number of unloading points served by the nth mine car, the demand for ore at the unloading points, the load of the nth mine car from the unloading point u to the unloading point v, and the usage status parameter S of the nth mine car n (where 1 means the nth mine car is used, and 0 means the nth mine car is not used), the driving parameters of the nth car on the road section from unloading point u to unloading point v (It is equal to 1, indicating that the nth mine car is traveling on the road section from unloading point u to unloading point v, and it is equal to 0, indicating that the nth mine car is not traveling on the road section from unloading point u to unloading point v), the driving parameters of the nth mine car on the road section from the departure point to the unloading point u (Its equal to 1 means that the nth mine car is traveling on the section from the departure point to the unloading point u, and it is equal to 0 means that the nth mine car is not traveling on the section from the departure point to the unloading point v), and n∈[1, the total number of mine cars N].
[0157] Specifically, the time when the nth mine car arrives at the unloading point v obtained in this step is expressed by the following formula:
[0158] T v =T u +t uv +w u +w v ,v∈M
[0159] Where T v represents the time when the nth mine car arrives at the unloading point v; T u represents the time when the nth mine car arrives at the unloading point u; w u represents the time that the nth mine car waits and loads at the unloading point u; w v represents the unloading time of the nth mine car at the unloading point v, and M represents the set of all unloading points.
[0160] (3) Based on the comprehensive parameters of mine car scheduling and operation obtained in step (2) and the parameters of the time when the mine car arrives at the unloading point, a mathematical model for minimizing the total transportation cost of all mine cars is constructed, such as Figure 2 As shown;
[0161] The total transportation cost of the minecarts in this step includes: the fixed cost of all minecarts, the opportunity cost and penalty cost of all minecart delivery time windows, the fuel consumption cost of all minecarts, and the carbon emission cost of all minecarts.
[0162] The process of constructing a mathematical model to minimize the total transportation cost of mine carts in this step includes the following sub-steps:
[0163] (3-1) According to the fixed transportation cost C of the nth mine car obtained in step (2) s And the usage status parameter S of the nth mine car n , obtain the various fixed costs of the mine carts, and systematically integrate the obtained costs of the mine carts to obtain the fixed costs of all mine carts during the transportation process.
[0164] Specifically, this step includes firstly calculating the fixed transportation cost of the nth mine car and the usage state parameter S of the nth mine car obtained in step (2). n , obtain the various fixed transportation costs of the nth mine car when performing the transportation task, including the depreciation cost of the mine car, daily loss cost, fixed loading and unloading cost, and maintenance cost. Then, all fixed transportation costs are systematically integrated to obtain the fixed cost of each mine car. Finally, the fixed cost C1 of all mine cars in the transportation process is constructed according to the fixed cost of the mine car. The specific expression is shown in the following formula:
[0165]
[0166] (3-2) According to the time when the nth mine car arrives at the unloading point u obtained in step (2), the opportunity cost and penalty cost of all mine car delivery time windows are obtained.
[0167] Specifically, this step includes: first, comparing the time when the nth mine car arrives at the unloading point u obtained in step (2) with the time window of the unloading point obtained in step (1); then, based on the comparison result and the opportunity cost of the nth mine car arriving earlier than the time window and the penalty cost of the nth mine car arriving later than the time window obtained in step (2), the opportunity cost and penalty cost distribution C2(u) of the nth mine car arriving at the unloading point u are obtained, as shown in the following formula.
[0168]
[0169] Among them B u E represents the earliest acceptable transportation time for the unloading point u;u represents the latest acceptable transportation time for the unloading point u; A represents the opportunity cost of the nth mine car arriving earlier than the time window; B represents the penalty cost of the nth mine car arriving later than the time window (the opportunity cost and penalty cost of all mine cars are the same).
[0170] Finally, according to the opportunity cost and penalty cost distribution of the nth mine car arriving at the unloading point u, the opportunity cost and penalty cost C2 of all mine car delivery time windows are obtained as follows:
[0171]
[0172] (3-3) using the load estimation method to process the unit distance fuel consumption of the nth mine car in the empty state, the unit distance fuel consumption of the nth mine car in the fully loaded state, the real-time load of the nth mine car, and the maximum load of the nth mine car obtained in step (2) to obtain the fuel consumption rate of the nth mine car when the load is q; and according to the obtained fuel consumption rate, the Euclidean distance between the departure point of the nth mine car and each unloading point and the Euclidean distance between every two unloading points obtained by preprocessing in step (1), and the fuel consumption price per unit distance of the nth mine car obtained in step (2), obtain the fuel consumption cost of all mine cars;
[0173] Specifically, this step is as follows: first, according to the unit distance fuel consumption of the nth mine car in an empty state, the unit distance fuel consumption of the nth mine car in a fully loaded state, the real-time load of the nth mine car, and the maximum load of the nth mine car obtained in step (2), the unit distance fuel consumption of the nth mine car under the current load is calculated using the load estimation method, and the fuel consumption rate R(q) of the nth mine car when the load is q is obtained. The specific calculation formula is shown as follows:
[0174]
[0175] Where Q L It represents the maximum load of the nth minecart (the maximum load of all minecarts is the same); R1 represents the fuel consumption per unit distance of the nth minecart in the empty state; R2 represents the fuel consumption per unit distance of the nth minecart in the heavy-loaded state (the fuel consumption per unit distance of all minecarts in the empty state and the heavy-loaded state are the same).
[0176] Then, according to the calculated fuel consumption rate of the nth mine car when the load is q, and the Euclidean distance d between the departure point of the nth mine car and each unloading point obtained by preprocessing in step (1), 1u and the Euclidean distance d between every two unloading points uv , obtain the fuel consumption F of the nth mine car in transporting between the departure point and each unloading point when the load is q 1v(q), and the fuel consumption F of the nth ore car in transporting between every two unloading points uv (q), and its specific calculation formula is shown below.
[0177]
[0178] Then, based on the obtained fuel consumption of the nth mine car between the departure point and each unloading point and the fuel consumption between every two unloading points when the load is q, and the fuel consumption price per unit distance of the nth mine car obtained in step (2), f Get the fuel consumption cost C3(q) of all minecarts when the load is q, as shown in the following formula:
[0179]
[0180] Then, replace q / Q in the above formula L Converted to the load factor of a minecart To obtain the simplified fuel consumption cost of all mine cars;
[0181] Then, based on the simplified fuel consumption cost of all mining vehicles and the driving parameters of the nth mining vehicle on the road section from the departure point to the unloading point v obtained in step (2), and the driving parameters of the nth vehicle on the road section from unloading point u to unloading point v Get the fuel consumption cost C3 of all minecarts as shown below:
[0182]
[0183] The load estimation method in the present invention has been disclosed in detail in "Research on Dynamic Distribution Vehicle Route Optimization under the Influence of Carbon Emissions" by Zhang Jinliang and Li Chao, Chinese Journal of Management Science, 2022, 30(09): 184-194, and will not be repeated here.
[0184] (3-4) Based on the fuel consumption of the nth mine car in transporting between the departure point and each unloading point and the fuel consumption between every two unloading points when the load is q obtained in step (3-3), and the carbon emission coefficient and unit distance carbon emission cost of the nth mine car obtained in step (2), the carbon emission cost of all mine cars is obtained.
[0185] Specifically, this step includes: first, according to the fuel consumption of the nth mine car in transporting between the departure point and each unloading point and the fuel consumption between every two unloading points when the load is q obtained in step (3-3) and the carbon emission coefficient of the nth mine car obtained in step (2), the carbon emissions of the nth mine car during transportation are obtained; then, according to the obtained carbon emissions of the nth mine car during transportation and the carbon emission cost per unit distance of the nth mine car obtained in step (2), the carbon emission cost C4 of all mine cars is obtained, as shown in the following formula.
[0186]
[0187] Where γ represents the carbon emission cost per unit distance of the nth mine car; ε represents the carbon emission coefficient of the nth mine car (the carbon emission cost per unit distance and the carbon emission coefficient of all mine cars are the same).
[0188] Finally, according to the mine car load rate And the carbon emission costs of all minecarts, to obtain the final carbon emission cost C4 of all minecarts, the specific form of which is shown in the following formula:
[0189]
[0190] (3-5) The fixed costs of all mine carts obtained in step (3-1), the opportunity costs and penalty costs of all mine cart delivery time windows obtained in step (3-2), the fuel consumption costs of all mine carts obtained in step (3-3), and the carbon emission costs of all mine carts obtained in step (3-4) are integrated to obtain a mathematical model for minimizing the total transportation cost of all mine carts, as shown in the following formula.
[0191] minC=C1+C2+C3+C4
[0192] Furthermore, the calculation formula of the fixed cost of all mine cars obtained in step (3-1), the opportunity cost and penalty cost of all mine car delivery time windows obtained in step (3-2), the fuel consumption cost of all mine cars obtained in step (3-3), and the carbon emission cost of all mine cars obtained in step (3-4) are substituted into the above-mentioned total transportation cost minimization mathematical model to obtain the following expression:
[0193]
[0194] (4) Add multiple constraints to the mathematical model for minimizing the total transportation cost of all mine carts obtained in step (3) to obtain an updated mathematical model for minimizing the total transportation cost of all mine carts.
[0195] Specifically, the multiple constraints added in this step are as follows:
[0196] ① The number of minecarts used cannot exceed the total number of minecarts N, as shown in the following formula.
[0197]
[0198] ② Each mine car needs to start from the departure point and return to the departure point after completing the transportation task, as shown in the following formula.
[0199]
[0200] ③ Each unloading point requires and only requires one mine car to carry out transportation services, as shown in the following formula.
[0201]
[0202] ④ The number of mine cars arriving at the unloading point is the same as the number of mine cars leaving the unloading point, as shown in the following formula.
[0203]
[0204] ⑤The load of each mine car cannot exceed its maximum load capacity, as shown in the following formula.
[0205]
[0206] Among them, q u Represents the demand for ore at unloading point u.
[0207] ⑥ The total loading capacity of all mine cars used must meet (that is, be greater than or equal to) the total amount of ore that needs to be transported by all unloading points served by all mine cars, as shown in the following formula.
[0208]
[0209] in, Represents the load of the nth mine car from the departure point to the unloading point u.
[0210] ⑦ The time window constraint for the mine car to arrive at the unloading point it serves is shown in the following formula.
[0211]
[0212] ⑧ Ensure that the transportation task of each mine car is a continuous process, as shown in the following formula.
[0213]
[0214] ⑨The sub-circuit of the minecart eliminates constraints, as shown in the following formula.
[0215]
[0216] Among them, |K n| represents the number of unloading points served by the nth minecart.
[0217] ⑩Parameter S n , The constraints on the value range of are shown in the following formula.
[0218]
[0219] The advantage of this step is that by adding multiple constraints to the mathematical model for minimizing the total transportation cost of all mine carts obtained in step (3), an updated mathematical model for minimizing the total transportation cost of all mine carts is obtained, so that the model reduces its dependence on data and enhances its ability to adapt to multiple constraints.
[0220] The advantage of the above steps (1) to (4) is that a mathematical model for minimizing the total transportation cost of mine carts is established, which includes the fixed costs of all mine carts, the opportunity costs and penalty costs of all mine cart delivery time windows, the fuel consumption costs of all mine carts, and the carbon emission costs of all mine carts. Multiple constraints are added to the mathematical model for minimizing the total transportation cost of all mine carts obtained in step (3), thereby obtaining an updated mathematical model for minimizing the total transportation cost of all mine carts, thereby achieving a comprehensive evaluation of the total transportation cost of mine carts.
[0221] (5) Using the improved PSO algorithm and the updated mathematical model of minimizing the total transportation cost of all mine carts obtained in step (4), the mathematical model of minimizing the total transportation cost of all mine carts obtained in step (3) is converted into a fitness evaluation function of the improved PSO algorithm, and the fitness value of each particle in the particle swarm is obtained according to the fitness evaluation function to obtain the minimum total transportation cost of all mine carts.
[0222] like Figure 3 As shown, this step includes the following sub-steps:
[0223] (5-1) Initialize the particle swarm parameters of the improved PSO algorithm to obtain the initialized particle swarm parameters.
[0224] Specifically, in this step, the particle swarm parameters of the improved PSO algorithm are initialized as follows: the initial velocity vector and initial position vector of the particles in the particle swarm are set to random values; the number of particles in the particle swarm is set to NP = 50; the speed limit of the particles in the particle swarm in the search space is set to [v b ,v s ]=[-1,1](where v b represents the maximum speed limit of particles in the particle swarm, v s represents the minimum speed limit of particles in the particle swarm); sets the position limit of particles in the particle swarm in the search space [x b ,x s]=[-1,1](where x b represents the maximum limit position of particles in the particle swarm, x s represents the minimum restricted position of particles in the particle swarm); the inertia weight, individual learning factor and group learning factor of particles in the particle swarm are set to the random perturbation adaptive adjustment mode.
[0225] (5-2) Set counter k=0.
[0226] (5-3) Determine whether k is greater than a preset maximum number of iterations Num (the maximum number of iterations Num ranges from 50 to 500, preferably 200). If so, proceed to step (5-13); otherwise, proceed to step (5-4);
[0227] (5-4) Based on the initialized particle swarm parameters obtained in step (5-1) and the multiple constraints added in step (4), the mathematical model for minimizing the total transportation cost of all mine cars constructed in step (3) is used as the fitness evaluation function of the improved PSO algorithm.
[0228] Specifically, the fitness evaluation function F(C) in this step is as shown below.
[0229] F(C)=minC
[0230] (5-5) According to the fitness evaluation function of the improved PSO algorithm obtained in step (5-4) and the fitness value of each particle i in the particle swarm at its initial position, obtain the individual historical optimal fitness value of particle i at the kth iteration The number of particles in the particle swarm NP].
[0231] (5-6) Select the minimum value from the individual historical optimal fitness values of all particles obtained in step (5-5) as the global historical optimal fitness value Gbest at the kth iteration k (At this time, the number of iterations k = 0).
[0232] (5-7) Obtain the velocity vector and position vector of each particle in the particle swarm at the kth iteration according to the initialized particle swarm parameters obtained in step (5-1), and calculate the velocity vector and position vector of the particle at the k+1th iteration according to the velocity vector and position vector of the particle at the kth iteration.
[0233] Specifically, in this step, the velocity update formula and position update formula of each particle i in the particle swarm at the k+1th iteration are as follows:
[0234]
[0235] Where ω represents the inertia weight, and ω=-0.4k / K+(1 / n)·randnd(), randnd() represents a random value that obeys the normal distribution; c1, c2 represent the individual learning factor and the group learning factor, c1=-2k / K+(1 / n)·randnd(), c2=2k / K+n·randnd(); r1 and r2 represent random numbers between the interval [0,1]; represents the velocity vector of particle i at the k+1th iteration; represents the velocity vector of particle i at the kth iteration; Gbest represents the individual historical optimal fitness value of particle i at the kth iteration; k Represents the global historical optimal fitness value at the kth iteration; represents the position vector of particle i at the k+1th iteration; represents the position vector of particle i at the kth iteration.
[0236] (5-8) According to the speed limit and position limit of each particle in the particle swarm preset in step (5-1), the speed vector and position vector of the particle obtained in step (5-7) at the k+1th iteration are screened to obtain a plurality of screened particles.
[0237] Specifically, this step includes firstly screening the velocity vector of each particle obtained in step (5-7) at the k+1th iteration according to the velocity limit of the particles in the particle group preset in step (5-1) (the specific expression is shown in the following formula), thereby obtaining a first group of screened particles;
[0238]
[0239] Then, the position vector of each particle obtained in step (5-7) at the k+1th iteration is screened according to the position restriction of the particles in the particle group preset in step (5-1) (the specific expression is shown in the following formula), thereby obtaining a second group of screened particles.
[0240]
[0241] Finally, the first group of particles and the second group of particles obtained above are combined to obtain the screened particles.
[0242] (5-9) The fitness evaluation function obtained in step (5-4) is used to recalculate the fitness value of each particle screened in step (5-8) to obtain the fitness value of the particle at the k+1th iteration.
[0243] (5-10) The individual historical optimal fitness value of the particle at the kth iteration calculated according to step (5-5) The fitness value of each particle at the k+1th iteration calculated by step (5-9) is used to obtain the individual historical optimal fitness value of each particle in the particle swarm at the k+1th iteration after adjustment.
[0244] The specific expression of this step is as follows:
[0245]
[0246] (5-11) The global historical optimal fitness value Gbest at the kth iteration calculated according to step (5-6) k And the individual historical optimal fitness value of each particle at the k+1th iteration obtained in step (5-10) Get the global historical optimal fitness value Gbest of each particle in the particle swarm at the k+1th iteration after adjustment k+1 .
[0247] The specific expression of this step is as follows:
[0248]
[0249] (5-12) Set k=k+1 and return to step (5-3).
[0250] (5-13) Output the global historical optimal fitness value as the minimum total transportation cost of all mine carts.
[0251] The advantage of this step is that, by adaptively improving the inertia weight, individual learning factor and group learning factor of the PSO algorithm, the dynamic adaptability of the algorithm in solving the problem of minimizing the total transportation cost of mine carts is enhanced.
[0252] (6) Convert the minimum total transportation cost of all mine carts obtained in step (5) (global suboptimal solution) into the initial pheromone enhancement amount of the improved ACO algorithm.
[0253] In this step, the specific calculation formula for converting the minimum total transportation cost of all mine carts obtained in step (5) into the initial pheromone enhancement amount of the improved ACO algorithm is as follows:
[0254] τ(0)=τ Β (0)+Δτ P
[0255] Among them, τ(0) represents the initial pheromone distribution value; τ Β (0) represents the initial value of pheromone, τ Β (0) = 0; Δτ PIndicates the conversion of the minimum total transportation cost of all mine carts obtained in the improved PSO algorithm into the increment of the pheromone intensity of the ACO algorithm, Δτ p =Gbest k+1 , Gbest k+1 is the global historical optimal fitness value obtained at the k+1th iteration obtained in step (5), and k+1=Num.
[0256] (7) Use the improved ACO algorithm and the initial pheromone increment of the improved ACO algorithm obtained in step (6) to obtain the minimum total transportation cost of all mine carts
[0257] like Figure 4 As shown, this step specifically includes the following sub-steps:
[0258] (7-1) Initialize the ant colony related parameters in the improved ACO algorithm, and use the improved ACO algorithm initial pheromone increment obtained in step (6) as the initial value of the pheromone in the improved ACO algorithm to obtain the initialized ant colony parameters.
[0259] Specifically, the step is as follows: first, set the number of ants in the ant colony R = 50; set the time window heuristic factor λ = 1; set the pheromone heuristic factor α∈[0,5], preferably 1; set the expected heuristic factor β∈[0,10], preferably 7; set the pheromone volatility coefficient The constant C0∈[0,1] is preferably 0.5. Then, the initial pheromone increment of the improved ACO algorithm obtained in step (6) is used as the initial value of the pheromone in the improved ACO algorithm.
[0260] (7-2) Set counter NC=1.
[0261] (7-3) Determine whether NC is greater than the preset maximum number of iterations NC max (The maximum number of iterations NC max The value range of is 100 to 500, preferably 200), if yes, go to step (7-15), otherwise, go to step (7-4).
[0262] (7-4) Set the current number of ants r = 1.
[0263] (7-5) Determine whether r is greater than the number of ants R in the ant colony preset in the ant colony parameters after initialization in step (7-1). If so, proceed to step (7-11); otherwise, proceed to step (7-6).
[0264] (7-6) Place the rth ant in the ant colony at the departure point of the mine cart, and create a taboo list to record each unloading point reached by the rth ant in this cycle, add the departure point of the mine cart to the taboo list, and add multiple unloading points of the mine cart that the rth ant has not yet reached to the preset unloading point set m r middle.
[0265] (7-7) Determine whether the rth ant has reached the preset unloading point set m r If yes, go to step (7-10), otherwise go to step (7-8).
[0266] (7-8) In combination with the actual transportation situation of the mine, the time window constraint data of the unloading point of the mine car in the preprocessed basic operation scheduling data obtained in step (1) is combined with the state transition probability formula of the traditional ACO algorithm to obtain an improved state transition probability formula. According to the obtained improved state transition probability formula, the next unloading point that the rth ant will reach is obtained, and the next unloading point is added to the taboo table, and then step (7-9) is entered.
[0267] In this step, the time window constraint data of the unloading point of the mine car obtained in step (1) is combined with the state transition probability formula of the traditional ACO algorithm to obtain an improved state transition probability formula, the specific expression of which is shown in the following formula.
[0268]
[0269] in represents the probability selection function of the rth ant at time t choosing the unloading point v of the mine car at the unloading point u of the mine car as the next node; τ uv (t) represents the pheromone concentration on the transport path (u, v) at time t; η uv (t) The expected degree of the ant from the unloading point u of the mine car to the unloading point v of the target mine car at time t, η uv =1 / c uv ;c uv represents the transportation cost required for the ant to travel from the unloading point u of the mine cart to the unloading point v of the mine cart; φ uv (t) represents the relative coincidence value of the time window of the unloading point of the mine car at time t; p represents a random number between the interval [0,1];
[0270] (7-9) Obtain the current position information of the rth ant according to the next unloading point that the rth ant will reach in step (7-8), and update the unloading point set according to the current position information of the rth ant to obtain an updated unloading point set, and return to step (7-7).
[0271] (7-10) Set r=r+1 and return to step (7-5).
[0272] (7-11) Obtain the path formed by all the unloading points reached by each ant, and calculate the transportation cost of the corresponding path according to the mathematical model for minimizing the total transportation cost of all mine carts obtained in step (3), and then select the minimum value from the transportation costs corresponding to the paths generated by all ants as the transportation cost pbest(t) corresponding to the path generated by the optimal ant at the NC-th iteration at time t, where if the NC-th iteration is the first iteration (i.e., NC=1), then the transportation cost corresponding to the path generated by the optimal ant at the NC-th iteration at time t is set to the transportation cost gbest(t) corresponding to the path generated by the global optimal ant at the NC-th iteration at time t, and then enter step (7-12).
[0273] (7-12) Based on the transportation cost pbest(t) corresponding to the path generated by the optimal ant at the NC-th iteration at time t obtained in step (7-11), the transportation cost gbest(t-1) corresponding to the path generated by the global optimal ant at the NC-th iteration at time t-1 is updated to obtain the transportation cost gbest(t) corresponding to the path generated by the global optimal ant at the NC-th iteration at time t, and then enter step (7-13).
[0274] The transportation cost update formula corresponding to the path generated by the global optimal ant at the NC-th iteration at time t is as follows:
[0275]
[0276] (7-13) Determine whether NC is less than or equal to 20. If so, update the pheromone concentration on the path at the NC-th iteration at time t according to the transportation cost pbest(t) corresponding to the path generated by the optimal ant at the NC-th iteration at time t obtained in step (7-11) to obtain the pheromone concentration on the path at the NC-th iteration at time t+1. Otherwise, update the pheromone concentration on the path at the NC-th iteration at time t according to the transportation cost gbest(t) corresponding to the path generated by the global optimal ant at the NC-th iteration at time t obtained in step (7-12) to obtain the pheromone concentration on the path at the NC-th iteration at time t+1, and then enter step (7-14).
[0277] The formula for updating the pheromone concentration on the path at the NC-th iteration at time t+1 is as follows.
[0278]
[0279] where Δτ uv(t) represents the pheromone increment on the path from unloading point u to unloading point v at time t; Δτ uv (t) = 1 / b t , where b t The values are as follows:
[0280]
[0281] (7-14) Set NC=NC+1, clear the taboo table, and return to step (7-3).
[0282] (7-15) Output the NC at time t+1 max The transportation cost gbest(t+1) corresponding to the path generated by the global optimal ant at the iteration is taken as the minimum total transportation cost of all mine carts.
[0283] The advantage of steps (6) to (7) is that by converting the minimum total transportation cost of all mine carts obtained in the improved PSO algorithm into the initial pheromone increment of the ACO algorithm, and improving the state transfer rules and pheromone update method of the ACO algorithm, the search ability and convergence speed of the ACO algorithm are improved, while the stability and robustness of the algorithm are enhanced.
[0284] The advantage of steps (5) to (7) is that the problem of minimizing the total transportation cost of the mine cart is solved by the joint optimization of the improved PSO-ACO algorithm. The improved PSO algorithm is used to make up for the problem of insufficient pheromone distribution and slow convergence speed of the ACO algorithm in the early stage, and the improved ACO algorithm is used to make up for the problem of poor local search ability of the improved PSO algorithm, so as to achieve complementary advantages of the two improved algorithms, thereby enhancing the stability and robustness of the algorithm.
[0285] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A mine car optimization scheduling method based on improved PSO-ACO, characterized in that: The following steps are involved: (1) Obtaining basic operation scheduling data and basic working data of a mine car unloading point, and preprocessing the basic operation scheduling data and basic working data respectively to obtain preprocessed basic operation scheduling data and preprocessed basic working data respectively. (2) Obtaining comprehensive parameters of mine car scheduling and operation, and obtaining the time parameters of the mine car arriving at the unloading point based on the basic operation scheduling data preprocessed in step (1) and the basic working data preprocessed. (3) constructing a mathematical model for minimizing the total transportation cost of all mine cars based on the comprehensive parameters of mine car scheduling and operation obtained in step (2) and the time parameters of the mine cars arriving at the unloading point; (4) Add multiple constraints to the mathematical model for minimizing the total transportation cost of all mine carts obtained in step (3) to obtain an updated mathematical model for minimizing the total transportation cost of all mine carts. (5) Using the improved PSO algorithm and the updated mathematical model of minimizing the total transportation cost of all mine carts obtained in step (4), the mathematical model of minimizing the total transportation cost of all mine carts obtained in step (3) is converted into a fitness evaluation function of the improved PSO algorithm, and the fitness value of each particle in the particle swarm is obtained according to the fitness evaluation function to obtain the minimum total transportation cost of all mine carts. (6) Convert the minimum total transportation cost of all mine carts obtained in step (5) into the initial pheromone enhancement amount of the improved ACO algorithm. (7) Using the improved ACO algorithm and the initial pheromone increment of the improved ACO algorithm obtained in step (6), the minimum total transportation cost of all mine carts is obtained.
2. The method for optimizing the dispatching of mine vehicles based on improved PSO-ACO according to claim 1 is characterized in that: The basic operation scheduling data of the unloading point of the mine car includes: the location information data of the unloading point of the mine car and the time window constraint data of the unloading point of the mine car; The basic working data of the mine car include: the travel speed of the nth mine car and the time when the nth mine car arrives at the unloading point u. In step (1), the basic operation scheduling data of the unloading point of the mine car is preprocessed to obtain the preprocessed basic operation scheduling data. In this process, the position information data of the unloading point and the departure point of the mine car and the time window constraint data are first obtained; then, the position information data of the unloading point and the departure point of the mine car are converted into accurate coordinate data, and the obtained coordinate data is expressed in the two-dimensional plane as coordinates (x u ,y u ) form, and then, according to the coordinates (x u ,y u ), and use the Euclidean distance to calculate the Euclidean distance between the departure point of the nth mine car and each unloading point, as well as the Euclidean distance between every two unloading points of the nth mine car. Then, the time window constraint data is converted into numerical time window constraint data in the form of hours and minutes. Then, the numerical time window constraint data is normalized. Finally, all time windows in the normalized time window constraint data are sorted in chronological order to obtain sorted time window constraint data. All Euclidean distances and sorted time window constraint data together constitute the preprocessed basic operation scheduling data, where x u Represents the horizontal coordinate of the u-th unloading point among all the unloading points of the mine car, y u Represents the ordinate of the u-th unloading point among all the unloading points of the mine car; In step (1), the basic working data of the mine car is preprocessed to obtain the preprocessed basic working data of the mine car. Specifically, according to the travel speed of the nth mine car and the obtained Euclidean distance between the departure point of the nth mine car and each unloading point, and the Euclidean distance between every two unloading points of the nth mine car, the transportation time t between the departure point of the nth mine car and each unloading point is calculated. 1u =d 1u / v c , and the transportation time t between each two unloading points of the nth mine car uv =d uv / v c All the transportation times obtained are the basic working data after preprocessing, among which, t 1u represents the transportation time between the departure point of the mine car and each unloading point u, d 1u represents the Euclidean distance between the departure point and the unloading point u, v c Indicates the current speed of the minecart, d uv represents the Euclidean distance between unloading point u and unloading point v, t uv Indicates the transportation time between every two unloading points of the minecart.
3. The method for optimizing the dispatching of mine vehicles based on improved PSO-ACO according to claim 1 or 2, characterized in that: The comprehensive parameters of mine car scheduling and operation obtained in step (2) include: the set of all mine cars at the departure point, the set of unloading points, the departure point, the number of mine cars used, the opportunity cost of the nth mine car arriving earlier than the time window, the penalty cost of the nth mine car arriving later than the time window, the time the nth mine car waits for loading at the unloading point u, the unloading time of the nth mine car at the unloading point v, the time when the nth mine car arrives at the unloading point u, and the fixed transportation cost C of the nth mine car s , the fuel consumption price per unit distance of the nth mine car, the fuel consumption per unit distance of the nth mine car in an empty state, the fuel consumption per unit distance of the nth mine car in a fully loaded state, the real-time loading capacity of the nth mine car, the maximum load capacity of the nth mine car, the carbon emission coefficient of the nth mine car, the carbon emission cost per unit distance of the nth mine car, the number of unloading points served by the nth mine car, the demand for ore at the unloading points, the load of the nth mine car from the unloading point u to the unloading point v, and the usage status parameter S of the nth mine car n , the driving parameters of the nth vehicle on the road section from unloading point u to unloading point v The driving parameters of the nth mine car on the road section from the departure point to the unloading point u And n∈[1, the total number of minecarts N]. The time when the nth mine car arrives at the unloading point v obtained in step (2) is expressed by the following formula: T v =T u +t uv +w u +w v ,v∈M Where T v represents the time when the nth mine car arrives at the unloading point v; T u represents the time when the nth mine car arrives at the unloading point u; w u represents the time that the nth mine car waits and loads at the unloading point u; w v represents the unloading time of the nth mine car at the unloading point v, and M represents the set of all unloading points.
4. The method for optimizing the dispatching of mine vehicles based on improved PSO-ACO according to any one of claims 1 to 3, characterized in that: The total transportation cost of the mine carts in step (3) includes: the fixed cost of all mine carts, the opportunity cost and penalty cost of all mine cart delivery time windows, the fuel consumption cost of all mine carts, and the carbon emission cost of all mine carts; The process of constructing the mathematical model for minimizing the total transportation cost of the mine cart in step (3) includes the following sub-steps: (3-1) According to the fixed transportation cost C of the nth mine car obtained in step (2) s And the usage status parameter S of the nth mine car n , obtain the various fixed costs of the mine carts, and systematically integrate the obtained costs of the mine carts to obtain the fixed costs of all mine carts during the transportation process. Specifically, this step includes firstly calculating the fixed transportation cost of the nth mine car and the usage state parameter S of the nth mine car obtained in step (2). n , obtain the various fixed transportation costs of the nth mine car when performing the transportation task, including the depreciation cost of the mine car, daily loss cost, fixed loading and unloading cost, and maintenance cost. Then, all fixed transportation costs are systematically integrated to obtain the fixed cost of each mine car. Finally, the fixed cost C1 of all mine cars in the transportation process is constructed according to the fixed cost of the mine car. The specific expression is shown in the following formula: (3-2) According to the time when the nth mine car arrives at the unloading point u obtained in step (2), the opportunity cost and penalty cost of all mine car delivery time windows are obtained. Specifically, this step includes: first, comparing the time when the nth mine car arrives at the unloading point u obtained in step (2) with the time window of the unloading point obtained in step (1); then, based on the comparison result and the opportunity cost of the nth mine car arriving earlier than the time window and the penalty cost of the nth mine car arriving later than the time window obtained in step (2), the opportunity cost and penalty cost distribution C2(u) of the nth mine car arriving at the unloading point u are obtained, as shown in the following formula. Among them B u E represents the earliest acceptable transportation time for the unloading point u; u represents the latest acceptable transportation time for the unloading point u; A represents the opportunity cost of the nth mine car arriving earlier than the time window; B represents the penalty cost of the nth mine car arriving later than the time window (the opportunity cost and penalty cost of all mine cars are the same). Finally, according to the opportunity cost and penalty cost distribution of the nth mine car arriving at the unloading point u, the opportunity cost and penalty cost C2 of all mine car delivery time windows are obtained as follows: (3-3) using the load estimation method to process the unit distance fuel consumption of the nth mine car in the empty state, the unit distance fuel consumption of the nth mine car in the fully loaded state, the real-time load of the nth mine car, and the maximum load of the nth mine car obtained in step (2) to obtain the fuel consumption rate of the nth mine car when the load is q; and according to the obtained fuel consumption rate, the Euclidean distance between the departure point of the nth mine car and each unloading point and the Euclidean distance between every two unloading points obtained by preprocessing in step (1), and the fuel consumption price per unit distance of the nth mine car obtained in step (2), obtain the fuel consumption cost of all mine cars; Specifically, this step is as follows: first, according to the unit distance fuel consumption of the nth mine car in an empty state, the unit distance fuel consumption of the nth mine car in a fully loaded state, the real-time load of the nth mine car, and the maximum load of the nth mine car obtained in step (2), the unit distance fuel consumption of the nth mine car under the current load is calculated using the load estimation method, and the fuel consumption rate R(q) of the nth mine car when the load is q is obtained. The specific calculation formula is shown as follows: Where Q L It represents the maximum load of the nth minecart (the maximum load of all minecarts is the same); R1 represents the fuel consumption per unit distance of the nth minecart in the empty state; R2 represents the fuel consumption per unit distance of the nth minecart in the heavy-loaded state (the fuel consumption per unit distance of all minecarts in the empty state and the heavy-loaded state are the same). Then, according to the calculated fuel consumption rate of the nth mine car when the load is q, and the Euclidean distance d between the departure point of the nth mine car and each unloading point obtained by preprocessing in step (1), 1u and the Euclidean distance d between every two unloading points uv , obtain the fuel consumption F of the nth mine car in transporting between the departure point and each unloading point when the load is q 1v (q), and the fuel consumption F of the nth ore car in transporting between every two unloading points uv (q), and its specific calculation formula is shown below. Then, based on the obtained fuel consumption of the nth mine car between the departure point and each unloading point and the fuel consumption between every two unloading points when the load is q, and the fuel consumption price per unit distance of the nth mine car obtained in step (2), f Get the fuel consumption cost C3(q) of all minecarts when the load is q, as shown in the following formula: Then, replace q / Q in the above formula L Converted to the load factor of a minecart To obtain the simplified fuel consumption cost of all mine cars; Then, based on the simplified fuel consumption cost of all mining vehicles and the driving parameters of the nth mining vehicle on the road section from the departure point to the unloading point v obtained in step (2), and the driving parameters of the nth vehicle on the road section from unloading point u to unloading point v Get the fuel consumption cost C3 of all minecarts as shown below: (3-4) Based on the fuel consumption of the nth mine car in transporting between the departure point and each unloading point and the fuel consumption between every two unloading points when the load is q obtained in step (3-3), and the carbon emission coefficient and unit distance carbon emission cost of the nth mine car obtained in step (2), the carbon emission cost of all mine cars is obtained. Specifically, this step includes: first, according to the fuel consumption of the nth mine car in transporting between the departure point and each unloading point and the fuel consumption between every two unloading points when the load is q obtained in step (3-3) and the carbon emission coefficient of the nth mine car obtained in step (2), the carbon emissions of the nth mine car during transportation are obtained; then, according to the obtained carbon emissions of the nth mine car during transportation and the carbon emission cost per unit distance of the nth mine car obtained in step (2), the carbon emission cost C4 of all mine cars is obtained, as shown in the following formula. Where γ represents the carbon emission cost per unit distance of the nth mine car; ε represents the carbon emission coefficient of the nth mine car. Finally, according to the mine car load rate And the carbon emission costs of all mine carts, to obtain the final carbon emission cost C4 of all mine carts, the specific form of which is shown in the following formula: (3-5) The fixed costs of all mine carts obtained in step (3-1), the opportunity costs and penalty costs of all mine cart delivery time windows obtained in step (3-2), the fuel consumption costs of all mine carts obtained in step (3-3), and the carbon emission costs of all mine carts obtained in step (3-4) are integrated to obtain a mathematical model for minimizing the total transportation cost of all mine carts, as shown in the following formula. minC=C1+C2+C3+C4 The fixed cost of all mine cars obtained in step (3-1), the opportunity cost and penalty cost of all mine car delivery time windows obtained in step (3-2), the fuel consumption cost of all mine cars obtained in step (3-3), and the carbon emission cost of all mine cars obtained in step (3-4) are substituted into the above-mentioned total transportation cost minimization mathematical model to obtain the following expression:
5. The method for optimizing the dispatching of mine vehicles based on improved PSO-ACO according to claim 4 is characterized in that: The multiple constraints added in step (4) are: A. The number of minecarts used cannot exceed the total number of minecarts N, as shown in the following formula. B. Each mining car needs to start from the departure point and return to the departure point after completing the transportation task, as shown in the following formula. C. Each unloading point requires and only requires one mine car to provide transportation services, as shown in the following formula. D. The number of mine cars arriving at the unloading point is the same as the number of mine cars leaving the unloading point, as shown in the following formula. E. The load of each mine car cannot exceed its maximum load capacity, as shown in the following formula. Among them, q u Represents the demand for ore at unloading point u. F. The total loading capacity of all mine cars used must meet (that is, be greater than or equal to) the total amount of ore that needs to be transported at all unloading points served by all mine cars, as shown in the following formula. in, Represents the load of the nth mine car from the departure point to the unloading point u. G. The time window constraint for the mine car to arrive at the unloading point it serves is shown in the following formula. H. Ensure that the transportation task of each mine car is a continuous process, as shown in the following formula. I. The sub-circuit of the minecart eliminates constraints as shown below. Among them, |K n | represents the number of unloading points served by the nth minecart. J. Parameter S n , The constraints on the value range of are shown in the following formula.
6. The method for optimizing the dispatching of mine vehicles based on improved PSO-ACO according to claim 5 is characterized in that: Step (5) includes the following sub-steps: (5-1) Initialize the particle swarm parameters of the improved PSO algorithm to obtain the initialized particle swarm parameters. (5-2) Set counter k=0. (5-3) Determine whether k is greater than the preset maximum number of iterations Num, if so, proceed to step (5-13), otherwise proceed to step (5-4); (5-4) Based on the initialized particle swarm parameters obtained in step (5-1) and the multiple constraints added in step (4), the mathematical model of minimum total transportation cost of all mine cars constructed in step (3) is used as the fitness evaluation function of the improved PSO algorithm, where the fitness evaluation function F(C) is shown in the following formula. F(C)=minC (5-5) According to the fitness evaluation function of the improved PSO algorithm obtained in step (5-4) and the fitness value of each particle i in the particle swarm at its initial position, obtain the individual historical optimal fitness value of particle i at the kth iteration i∈[1, the number of particles in the particle swarm NP]. (5-6) Select the minimum value from the individual historical optimal fitness values of all particles obtained in step (5-5) as the global historical optimal fitness value Gbest at the kth iteration k (At this time, the number of iterations k = 0). (5-7) Obtain the velocity vector and position vector of each particle in the particle swarm at the kth iteration according to the initialized particle swarm parameters obtained in step (5-1), and calculate the velocity vector and position vector of the particle at the k+1th iteration according to the velocity vector and position vector of the particle at the kth iteration. (5-8) According to the speed limit and position limit of each particle in the particle swarm preset in step (5-1), the speed vector and position vector of the particle obtained in step (5-7) at the k+1th iteration are screened to obtain a plurality of screened particles. (5-9) The fitness evaluation function obtained in step (5-4) is used to recalculate the fitness value of each particle screened in step (5-8) to obtain the fitness value of the particle at the k+1th iteration. (5-10) The individual historical optimal fitness value of the particle at the kth iteration calculated according to step (5-5) The fitness value of each particle at the k+1th iteration calculated by step (5-9) is used to obtain the individual historical optimal fitness value of each particle in the particle swarm at the k+1th iteration after adjustment. (5-11) The global historical optimal fitness value Gbest at the kth iteration calculated according to step (5-6) k And the individual historical optimal fitness value of each particle at the k+1th iteration obtained in step (5-10) Get the global historical optimal fitness value Gbest of each particle in the particle swarm at the k+1th iteration after adjustment k+1 : (5-12) Set k=k+1 and return to step (5-3). (5-13) Output the global historical optimal fitness value as the minimum total transportation cost of all mine carts.
7. The method for optimizing the dispatching of mine vehicles based on improved PSO-ACO according to claim 6 is characterized in that: In step (5-1), the particle swarm parameters of the improved PSO algorithm are initialized as follows: the initial velocity vector and initial position vector of the particles in the particle swarm are set to random values; the number of particles in the particle swarm is set to NP = 50; the speed limit of the particles in the particle swarm in the search space is set to [v b ,v s ]=[-1,1]; Set the position limit of particles in the particle swarm in the search space [x b ,x s ]=[-1,1], where x b represents the maximum limit position of particles in the particle swarm, x s represents the minimum restricted position of particles in the particle swarm, v b represents the maximum speed limit of particles in the particle swarm, v s Indicates the minimum speed limit of particles in the particle swarm; sets the inertia weight, individual learning factor and group learning factor of particles in the particle swarm as random perturbation adaptive adjustment mode. The velocity update formula and position update formula of each particle i in the particle swarm at the k+1th iteration in step (5-7) are as follows: Where ω represents the inertia weight, and ω=-0.4k / K+(1 / n)·randnd(), randnd() represents a random value that obeys the normal distribution; c1, c2 represent the individual learning factor and the group learning factor, c1=-2k / K+(1 / n)·randnd(), c2=2k / K+n·randnd(); r1 and r2 represent random numbers between the interval [0,1]; represents the velocity vector of particle i at the k+1th iteration; represents the velocity vector of particle i at the kth iteration; Gbest represents the individual historical optimal fitness value of particle i at the kth iteration; k Represents the global historical optimal fitness value at the kth iteration; represents the position vector of particle i at the k+1th iteration; represents the position vector of particle i at the kth iteration. Step (5-8) is specifically as follows: first, the velocity vector of each particle obtained in step (5-7) at the k+1th iteration is screened according to the velocity limit of the particles in the particle group preset in step (5-1), so as to obtain a first group of screened particles; Then, the position vector of each particle obtained in step (5-7) at the k+1th iteration is screened according to the position restriction of the particles in the particle group preset in step (5-1) (the specific expression is shown in the following formula), thereby obtaining a second group of screened particles. Finally, the first group of particles and the second group of particles obtained above are combined to obtain the screened particles.
8. The method for optimizing the dispatching of mine vehicles based on improved PSO-ACO according to claim 7 is characterized in that: In step (6), the specific calculation formula for converting the minimum total transportation cost of all mine carts obtained in step (5) into the initial pheromone enhancement amount of the improved ACO algorithm is as follows: τ(0)=τ Β (0)+Δt P Among them, τ(0) represents the initial pheromone distribution value; τ Β (0) represents the initial value of pheromone, τ Β (0) = 0; Δτ P Indicates the conversion of the minimum total transportation cost of all mine carts obtained in the improved PSO algorithm into the increment of the pheromone intensity of the ACO algorithm, Δτ p =Gbest k+1 , Gbest k+1 is the global historical optimal fitness value obtained at the k+1th iteration obtained in step (5), and k+1=Num.
9. The method for optimizing the dispatching of mine vehicles based on improved PSO-ACO according to claim 8 is characterized in that: Step (7) specifically includes the following sub-steps: (7-1) Initialize the ant colony related parameters in the improved ACO algorithm, and use the improved ACO algorithm initial pheromone increment obtained in step (6) as the initial value of the pheromone in the improved ACO algorithm to obtain the initialized ant colony parameters. Specifically, the step is as follows: first, set the number of ants in the ant colony R = 50; set the time window heuristic factor λ = 1; set the pheromone heuristic factor α∈[0,5], preferably 1; set the expected heuristic factor β∈[0,10], preferably 7; set the pheromone volatility coefficient The constant C0∈[0,1] is preferably 0.
5. Then, the initial pheromone increment of the improved ACO algorithm obtained in step (6) is used as the initial value of the pheromone in the improved ACO algorithm. (7-2) Set counter NC=1. (7-3) Determine whether NC is greater than the preset maximum number of iterations NC max (The maximum number of iterations NC max The value range of is 100 to 500, preferably 200), if yes, go to step (7-15), otherwise, go to step (7-4). (7-4) Set the current number of ants r = 1. (7-5) Determine whether r is greater than the number of ants R in the ant colony preset in the ant colony parameters after initialization in step (7-1). If so, proceed to step (7-11); otherwise, proceed to step (7-6). (7-6) Place the rth ant in the ant colony at the departure point of the mine cart, and create a taboo table to record each unloading point reached by the rth ant in this cycle, add the departure point of the mine cart to the taboo table, and add multiple unloading points of the mine cart that the rth ant has not yet reached to the preset unloading point set m r middle. (7-7) Determine whether the rth ant has reached the preset unloading point set m r If yes, go to step (7-10), otherwise go to step (7-8). (7-8) In combination with the actual transportation situation of the mine, the time window constraint data of the unloading point of the mine car in the preprocessed basic operation scheduling data obtained in step (1) is combined with the state transition probability formula of the traditional ACO algorithm to obtain an improved state transition probability formula. According to the obtained improved state transition probability formula, the next unloading point that the rth ant will reach is obtained, and the next unloading point is added to the taboo table, and then step (7-9) is entered. In this step, the time window constraint data of the unloading point of the mine car obtained in step (1) is combined with the state transition probability formula of the traditional ACO algorithm to obtain an improved state transition probability formula, the specific expression of which is shown in the following formula. in represents the probability selection function of the rth ant at time t choosing the unloading point v of the mine car at the unloading point u of the mine car as the next node; τ uv (t) represents the pheromone concentration on the transport path (u, v) at time t; η uv (t) The expected degree of the ant from the unloading point u of the mine car to the unloading point v of the target mine car at time t, η uv =1 / c uv ;c uv represents the transportation cost required for the ant to travel from the unloading point u of the mine cart to the unloading point v of the mine cart; φ uv (t) represents the relative coincidence value of the time window of the unloading point of the mine car at time t; p represents a random number between the interval [0,1]; (7-9) Obtain the current position information of the rth ant according to the next unloading point that the rth ant will reach in step (7-8), and update the unloading point set according to the current position information of the rth ant to obtain an updated unloading point set, and return to step (7-7). (7-10) Set r=r+1 and return to step (7-5). (7-11) Obtain the path formed by all the unloading points reached by each ant, and calculate the transportation cost of the corresponding path based on the mathematical model of minimizing the total transportation cost of all mine carts obtained in step (3), and then select the minimum value from the transportation costs corresponding to the paths generated by all ants as the transportation cost pbest(t) corresponding to the path generated by the optimal ant at the NC-th iteration at time t, where if the NC-th iteration is the first iteration, then set the transportation cost corresponding to the path generated by the optimal ant at the NC-th iteration at time t to the transportation cost gbest(t) corresponding to the path generated by the global optimal ant at the NC-th iteration at time t, and then enter step (7-12). (7-12) Based on the transportation cost pbest(t) corresponding to the path generated by the optimal ant at the NC-th iteration at time t obtained in step (7-11), the transportation cost gbest(t-1) corresponding to the path generated by the global optimal ant at the NC-th iteration at time t-1 is updated to obtain the transportation cost gbest(t) corresponding to the path generated by the global optimal ant at the NC-th iteration at time t, and then enter step (7-13). The transportation cost update formula corresponding to the path generated by the global optimal ant at the NC-th iteration at time t is as follows: (7-13) Determine whether NC is less than or equal to 20. If so, update the pheromone concentration on the path at the NC-th iteration at time t according to the transportation cost pbest(t) corresponding to the path generated by the optimal ant at the NC-th iteration at time t obtained in step (7-11) to obtain the pheromone concentration on the path at the NC-th iteration at time t+1. Otherwise, update the pheromone concentration on the path at the NC-th iteration at time t according to the transportation cost gbest(t) corresponding to the path generated by the global optimal ant at the NC-th iteration at time t obtained in step (7-12) to obtain the pheromone concentration on the path at the NC-th iteration at time t+1, and then enter step (7-14). The formula for updating the pheromone concentration on the path at the NC-th iteration at time t+1 is as follows: where Δτ uv (t) represents the pheromone increment on the path from unloading point u to unloading point v at time t; Δτ uv (t) = 1 / b t , where b t The values are as follows: (7-14) Set NC=NC+1, clear the taboo table, and return to step (7-3). (7-15) Output the NC at time t+1 max The transportation cost gbest(t+1) corresponding to the path generated by the global optimal ant at the iteration is taken as the minimum total transportation cost of all mine carts.
10. A mine car optimization scheduling system based on improved PSO-ACO, characterized in that: include: The first module is used to obtain basic operation scheduling data and basic work data of the unloading point of the mine car, and preprocess the basic operation scheduling data and basic work data respectively to obtain preprocessed basic operation scheduling data and preprocessed basic work data respectively. The second module is used to obtain the comprehensive parameters of mine car scheduling and operation, and obtain the time parameters of the mine car arriving at the unloading point based on the basic operation scheduling data and the basic work data preprocessed by the first module. The third module is used to construct a mathematical model for minimizing the total transportation cost of all mine cars based on the comprehensive parameters of mine car scheduling and operation obtained in the second module and the parameters of the time when the mine cars arrive at the unloading point; The fourth module is used to add multiple constraints to the mathematical model for minimizing the total transportation cost of all mine carts obtained in the third module to obtain an updated mathematical model for minimizing the total transportation cost of all mine carts. The fifth module is used to use the improved PSO algorithm and the updated mathematical model of minimizing the total transportation cost of all mine carts obtained in the fourth module to convert the mathematical model of minimizing the total transportation cost of all mine carts obtained in the third module into a fitness evaluation function of the improved PSO algorithm, and obtain the fitness value of each particle in the particle swarm according to the fitness evaluation function to obtain the minimum total transportation cost of all mine carts. The sixth module is used to convert the minimum total transportation cost of all mine carts obtained in the fifth module into the initial pheromone enhancement amount of the improved ACO algorithm. The seventh module is used to obtain the minimum total transportation cost of all mine carts using the improved ACO algorithm and the initial pheromone increment of the improved ACO algorithm obtained in the sixth module.
Citation Information
Cited By
Traffic scheduling method and system based on real-time data
CN120338398A
A traffic scheduling method and system based on real-time data
CN120338398B
Wharf short-distance transfer vehicle scheduling method and system based on hybrid cost optimization
CN121684762A
Multi-robot path planning method and device based on hybrid optimization and space-time prediction
CN122170866A
