A charging and replacing power coordination optimization method and system for multiple unmanned mine cars

Abstract

The present application relates to a kind of multi unmanned mine car-oriented charging and battery swapping collaborative optimization method and system, belong to charging and battery swapping collaborative optimization technical field.It includes: collecting mine car state, infrastructure state and task plan data, based on the improved collaborative filtering algorithm for each mine car generation personalized energy supply recommendation list;Based on personalized energy supply recommendation list, in combination with task plan, determine overall service execution sequence, execute service execution sequence and evaluate the service capability and expected income of each charging and battery swapping facility;In combination with vehicle state data, infrastructure state data, expected service capability and income evaluation report, use deep reinforcement learning algorithm to generate mine car dispatch instruction set, and through multi-aggregator cooperation mechanism optimization as cross-regional collaborative scheduling instruction set;Calculate the comprehensive service performance index when executing cross-regional collaborative scheduling instruction set, if better than preset service performance index, then execute the scheduling scheme;Otherwise, trigger degradation strategy, readjust scheduling scheme.

Description

Technical Field

[0001] This invention belongs to the field of charging and battery swapping collaborative optimization technology, and particularly relates to a charging and battery swapping collaborative optimization method and system for multiple unmanned mining vehicles. Background Technology

[0002] With the continuous advancement of smart mine construction, driverless mining trucks, as an important means to achieve automated and intelligent transportation in mining areas, have been widely used in open-pit coal mines, non-ferrous metal mines, and other scenarios. Compared with traditional manually driven vehicles, driverless mining trucks have advantages such as high operating efficiency, strong safety, and good controllability, especially demonstrating their value in intelligent scheduling in harsh environments and high-risk areas. However, due to the accelerating trend of electrification, more and more driverless mining trucks are adopting pure electric or hybrid electric drive systems. How to scientifically arrange their charging and battery swapping tasks has become a key link in ensuring transportation continuity and improving operational efficiency.

[0003] Currently, most unmanned mining truck charging and battery swapping scheduling still rely on static rules or local optimization strategies. For example, some mining areas adopt a fixed-time-window charging scheme, which automatically schedules charging tasks during shift breaks or when vehicle battery levels are below a threshold. At the equipment level, some systems are equipped with intelligent battery swapping stations or fast-charging piles and complete basic charging queuing through vehicle-to-ground communication, but most scheduling strategies are still primarily based on a first-come, first-served basis, lacking global dynamic scheduling capabilities based on factors such as job priority, remaining battery power, and task deadlines.

[0004] However, existing methods mostly employ local optimization or rule-driven strategies, making it difficult to achieve global coordination under multiple constraints such as task scheduling, power consumption, and power station capacity. This results in low system resource utilization and severe queuing for charging and battery swapping. Furthermore, most systems treat charging and battery swapping as independent processes, lacking a unified collaborative optimization model. This makes it difficult to dynamically select the optimal energy replenishment method based on vehicle status and task urgency, thus limiting overall scheduling efficiency and system flexibility. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the purpose of the invention is to provide a charging and swapping collaborative optimization method and system for multiple unmanned mining vehicles, which can improve the global optimization of the swapping process and the ability to respond to users' personalized needs, thereby ensuring the overall efficiency of the charging and swapping service.

[0006] In a first aspect, the present invention proposes a charging and battery swapping collaborative optimization method for multiple unmanned mining trucks, the method comprising:

[0007] S1 collects vehicle status data, infrastructure status data, and task plans from multiple unmanned mining trucks.

[0008] S2, based on the vehicle status data and the infrastructure status data, an improved collaborative filtering algorithm is used to recommend energy replenishment methods for each of the unmanned mining trucks and generate a personalized service recommendation list;

[0009] S3. Based on the personalized service recommendation list and the task plan, determine the service execution sequence for all mining trucks.

[0010] S4. Based on the service execution sequence, calculate the expected service capacity and revenue assessment report for each charging and swapping facility;

[0011] S5, combining the vehicle status data, the infrastructure status data, the expected service capacity, and the revenue assessment report, a mining truck dispatch instruction set is generated using a deep reinforcement learning algorithm;

[0012] S6, through a multi-aggregator collaboration mechanism, the mining truck dispatch instruction set is coordinated and optimized to obtain a cross-regional collaborative scheduling instruction set;

[0013] S7, calculate the comprehensive service performance index when executing the cross-regional collaborative scheduling instruction set, and determine whether the comprehensive service performance index is greater than the preset service performance index; if yes, proceed to S8; otherwise, trigger the system degradation strategy and readjust the mining truck dispatch instruction set.

[0014] S8, the cross-regional collaborative scheduling instruction is used as the final scheduling scheme, and the final scheduling scheme is executed to complete the charging and swapping collaborative optimization.

[0015] Furthermore, the vehicle status data includes remaining battery power, battery health, current location, and priority.

[0016] The infrastructure status data includes the availability of charging piles, fast charging stations, battery swapping stations, queue lengths, and estimated waiting times.

[0017] The task plan includes the task number, task type, planned departure time, task duration, and latest acceptable arrival time.

[0018] Furthermore, the energy replenishment methods include fast charging, slow charging, and battery swapping; S2 specifically includes:

[0019] S201, Based on the service records and feedback information of multiple unmanned mining vehicles related to different energy replenishment methods during the historical task execution process, a scoring matrix is ​​constructed;

[0020] S202, based on the vehicle status data, calculate the similarity index between mining trucks using the Pearson correlation coefficient, and generate a similarity matrix between each of the unmanned mining trucks based on the similarity index.

[0021] S203, using the similarity matrix and the rating matrix, estimate the service items with missing ratings to obtain the potential rating values ​​of the current unmanned mining truck for various unused energy replenishment methods;

[0022] S204, Based on the infrastructure status data, the potential score value is weighted and adjusted using a preset context correction model to generate a corrected score value that takes into account the actual operating environment;

[0023] S205, select the energy replenishment method with the highest score of the corrected score as the recommendation result, and record the corresponding facility number or service type to generate a personalized service recommendation list of energy replenishment methods for each unmanned mining truck.

[0024] Furthermore, S3 specifically includes:

[0025] S301, parse the personalized service recommendation list to obtain the recommended charging and battery swapping method for each of the unmanned mining trucks;

[0026] S302, confirm the final charging and battery swapping method and expected service time window for each mining truck;

[0027] S303, based on the recommended charging and swapping method, the charging and swapping method and the expected service time window, and in conjunction with the task plan, calculate the comprehensive priority score for each of the unmanned mining vehicles;

[0028] S304, Sort the comprehensive priority scores in descending order to form a preliminary service execution sequence;

[0029] S305, execute the preliminary service execution sequence and perform resource conflict check to determine if there is a conflict; if so, proceed to S306 to resolve the conflict; otherwise, proceed to S307.

[0030] S306, For minor conflicts, the unmanned mining truck with the lowest comprehensive priority score is recommended to be replaced by a nearby charging and swapping facility of the same type with service capacity; for severe conflicts, the demand of unmanned mining trucks is scheduled to available charging and swapping facilities with remaining service windows to form an effective service execution sequence, and proceed to S307.

[0031] S307, Output the service execution sequence containing all driverless mining vehicles.

[0032] Furthermore, S4 specifically includes:

[0033] S401, Based on the service execution sequence, count the number of reserved mining trucks for each of the charging and swapping facilities during the entire scheduling cycle;

[0034] S402, based on the number of reserved mining trucks and the equipment parameters of each charging and swapping facility, calculate the expected service capacity and remaining service capacity of each facility;

[0035] S403, Based on the number of reserved mining trucks, calculate the service load rate of each of the charging and swapping facilities during the scheduling cycle;

[0036] S404, Based on the number of reserved mining trucks, calculate the expected revenue of each of the charging and swapping facilities, and summarize the expected revenue of each facility to obtain the total comprehensive revenue;

[0037] S405, the expected service capacity, the remaining service capacity, the service load rate, the expected revenue, and the total revenue are structured and organized to generate a revenue assessment report for each charging and swapping facility.

[0038] Furthermore, S5 specifically includes:

[0039] S501, based on the vehicle status data, the infrastructure status data, the expected service capacity, and the revenue assessment report, the scheduling decision of the charging and swapping facilities is modeled as a Markov decision process;

[0040] S502, Based on the Markov decision process, the mine truck allocation strategy is trained by reinforcement learning through a deep Q-network to generate the optimal allocation strategy function.

[0041] S503: Sequentially acquire the vehicle status data and current environmental status of each unmanned mining truck to be dispatched, and construct the current state input vector;

[0042] S504, Input the current state input vector into the optimal assignment strategy function to obtain the action values ​​of all possible actions;

[0043] S505, Select the largest valid action among the action values ​​as the current minecart's scheduling response action;

[0044] S506, summarizes all the scheduling response actions of the mining trucks into the mining truck dispatch instruction set.

[0045] Furthermore, S6 specifically includes:

[0046] S601, monitor the operating status of all charging and swapping facilities within the jurisdiction of each aggregator, wherein the operating status includes load index, queue length and expected waiting time;

[0047] S602, determine whether the load index of each aggregator area exceeds the preset load index; if so, proceed to S603 to trigger the cross-aggregator cooperation mechanism; otherwise, execute the original minecart dispatch instruction set.

[0048] S603, the resource-overloaded sending aggregator region sends a cooperation request to the resource-unoverloaded receiving aggregator region. The cooperation request includes a list of mining trucks that cannot be arranged and the corresponding resource demand type.

[0049] S604, Based on the facility status within each of the receiving aggregator jurisdictions, select the receiving aggregator jurisdictions that have sufficient resources to complete the requested task.

[0050] S605, based on resource availability and distance cost, uses a lightweight matching algorithm to perform collaborative matching between the sending aggregator's jurisdiction and the receiving aggregator's jurisdiction to form a cross-aggregator resource allocation scheme;

[0051] S606, Update the mine truck dispatch instruction set according to the cross-aggregator resource allocation scheme to generate the cross-regional collaborative scheduling instruction set.

[0052] Furthermore, after S6 and before S7, the following is also included:

[0053] When re-acquiring and executing the cross-regional collaborative scheduling instruction set, the expected revenue and expected waiting time of each of the unmanned mining vehicles, as well as the service load rate of each charging and swapping facility, are determined.

[0054] Furthermore, the calculation method for the comprehensive service performance indicators is as follows:

[0055]

[0056] in, Indicates comprehensive service performance indicators. Indicates the return weight. Indicates total revenue. Indicates the theoretical maximum benefit, Indicates the weight of resource utilization. This represents the average service load rate. Indicates the service efficiency weight. This indicates the average expected waiting time. Indicates the maximum allowed waiting time. Indicates the weight of satisfaction. This represents the average satisfaction rating. This represents the highest satisfaction rating.

[0057] In a second aspect, the present invention proposes a charging and swapping collaborative optimization system for multiple unmanned mining trucks, comprising: a memory and a processor;

[0058] The memory stores an application program adapted to be executed by the processor to implement the charging and swapping collaborative optimization method for multiple unmanned mining vehicles as described in the first aspect.

[0059] The beneficial effects of this invention are as follows:

[0060] In this embodiment of the invention, an improved collaborative filtering algorithm is introduced to deeply integrate multi-source data such as the operating status of mining trucks and the status of infrastructure. This intelligently recommends the optimal energy replenishment method for each unmanned mining truck, enabling flexible selection and unified modeling between charging and battery swapping, overcoming the limitations of fragmented processing in traditional strategies. Based on this, the system gradually constructs a global collaborative scheduling mechanism with adaptive and closed-loop iterative capabilities, from multi-vehicle service ranking, infrastructure service capacity assessment, and deep reinforcement learning-based scheduling decisions to cross-regional collaborative optimization involving multiple aggregators. This mechanism can dynamically perceive key elements such as task plans, power consumption, power station load, and economic benefits, achieving efficient, flexible, and globally optimized charging and battery swapping scheduling strategies, significantly improving the system's resource utilization efficiency and overall operational performance. Attached Figure Description

[0061] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. It is obvious that the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings.

[0062] Figure 1 This is a flowchart illustrating a collaborative optimization method for charging and swapping of multiple unmanned mining trucks provided in an embodiment of the present invention.

[0063] Figure 2 This is a schematic diagram of a charging and swapping collaborative optimization system for multiple unmanned mining trucks provided in an embodiment of the present invention. Detailed Implementation

[0064] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0065] Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts disclosed in this invention.

[0066] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance. The terms "installed," "connected," and "linked" should be interpreted broadly; for example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0067] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of methods and systems consistent with some aspects of the invention as detailed in the appended claims.

[0068] This invention proposes a collaborative optimization method and system for charging and battery swapping of multiple unmanned mining vehicles. It addresses the shortcomings of existing methods, which often employ local optimization or rule-driven strategies, making it difficult to achieve global coordination under multiple constraints such as task scheduling, power consumption, and power station capacity. This results in low system resource utilization and severe queuing for charging and battery swapping. Furthermore, most systems treat charging and battery swapping as independent processes, lacking a unified collaborative optimization model. This makes it difficult to dynamically select the optimal energy replenishment method based on vehicle status and task urgency, thus limiting overall scheduling efficiency and system flexibility.

[0069] Method Implementation Examples

[0070] Reference Figure 1 The diagram shows a flowchart of a charging and swapping collaborative optimization method for multiple unmanned mining vehicles provided by an embodiment of the present invention.

[0071] This invention provides a charging and battery swapping collaborative optimization method for multiple unmanned mining trucks, the method comprising:

[0072] Specifically, the method includes steps S1 to S8.

[0073] S1 collects vehicle status data, infrastructure status data, and task plans from multiple unmanned mining trucks.

[0074] In one possible implementation, vehicle status data includes remaining battery power, battery health, current location, and priority.

[0075] Infrastructure status data includes the availability of charging stations, fast charging stations, battery swapping stations, queue lengths, and estimated waiting times.

[0076] The task plan includes the task number, task type, planned departure time, task duration, and latest acceptable arrival time.

[0077] S2, based on vehicle status data and infrastructure status data, uses an improved collaborative filtering algorithm to recommend energy replenishment methods for each unmanned mining truck and generate a personalized service recommendation list.

[0078] It should be noted that the improved collaborative filtering algorithm in this application introduces a multi-dimensional utility score and a context-aware correction mechanism on the basis of traditional collaborative filtering, so as to realize personalized decision-making for unmanned mining trucks under different energy replenishment methods.

[0079] In one possible implementation, energy replenishment methods include fast charging, slow charging, and battery swapping. S2 specifically includes:

[0080] S201, based on service records and feedback information related to different energy replenishment methods during the historical mission execution of multiple unmanned mining trucks, constructs a scoring matrix:

[0081]

[0082]

[0083] Where R represents the rating matrix, This represents the overall score of the i-th mining truck for the j-th energy replenishment method. This represents the charging and battery swapping efficiency score for the i-th mining truck when choosing the j-th energy replenishment method. The score represents the impact of the i-th mining truck choosing the j-th energy replenishment method on the task scheduling time. This represents the cost-benefit score of the i-th mining truck choosing the j-th energy replenishment method. The value represents the potential impact score of the i-th mining truck choosing the j-th energy replenishment method on battery health, m represents the number of mining trucks, w1 represents the efficiency score weight, w2 represents the time impact score weight, w3 represents the cost score weight, and w4 represents the health score weight.

[0084] It should be noted that, Based on experience.

[0085] Optionally, charging / swapping efficiency rating ,in, Indicates the ideal resupply time. Indicates the actual resupply time. Impact rating of task scheduling time. Where max represents maximization, and Δt represents the delay time caused by resupply. Indicates the maximum allowable delay. Cost-benefit score. ,in, express, Indicates the maximum acceptable charging cost. Potential impact rating for battery health. ,in, This indicates the decrease in battery health caused by the charging operation. This represents the minimum acceptable baseline value for health decline, and | represents the absolute value.

[0086] S202, based on vehicle status data, calculates the similarity index between mining trucks using the Pearson correlation coefficient, and generates a similarity matrix between each unmanned mining truck based on the similarity index:

[0087]

[0088]

[0089] Where S represents the similarity matrix, Let represent the similarity between the i-th mining car and the k-th mining car. This represents the value of the i-th mining truck on the d-th state feature. This represents the value of the k-th mining truck on the d-th state feature. Let represent the average value of the state characteristics of the i-th mining car. Let represent the average state characteristics of the k-th mining truck, and D represent the total number of state dimensions.

[0090] The Pearson correlation coefficient is a statistical indicator that measures the degree of linear correlation between two variables, reflecting the consistency of their changing trends. Its value ranges from -1 to 1. A coefficient close to 1 indicates a high positive correlation, meaning that an increase in one variable tends to increase the other. A coefficient close to -1 indicates a strong negative correlation. A coefficient close to 0 indicates almost no linear relationship between the two variables.

[0091] S203, using the similarity matrix and the rating matrix, estimates the service items with missing ratings, and obtains the potential rating values ​​of the current unmanned mining truck for various unused energy replenishment methods:

[0092]

[0093] in, This represents the predicted score of the i-th mining truck for the j-th energy replenishment method. This represents the first k mine cars that are similar to the i-th mine car. Let represent the similarity between the i-th mining car and the k-th mining car. This represents the historical score of the k-th mining truck on the j-th energy replenishment method. This indicates taking the absolute value.

[0094] In this embodiment of the invention, by predicting missing items in the rating matrix, the problem of sparse or incomplete mining truck rating data can be effectively alleviated. This allows for the inference of potential preferences based on the ratings of similar mining trucks, even if some vehicles have never used a certain energy replenishment method. This process not only improves the coverage and practicality of the collaborative filtering algorithm but also enhances the system's adaptability to new vehicles or cold start scenarios, ensuring that the completeness and predictive ability of the recommendation list remain high-quality even with incomplete information.

[0095] S204: Based on infrastructure status data, the potential score values ​​are weighted and adjusted using a pre-defined context correction model to generate corrected score values ​​that take into account the actual operating environment.

[0096] The context-correction model is a type of model that dynamically adjusts the basic recommendation results by incorporating real-time operating environment information. It calculates a context-correction coefficient for each energy replenishment method by analyzing contextual factors such as the availability of charging and swapping facilities, queue length, estimated waiting time, distance, electricity price, utilization rate, and renewable energy supply. When the facility is in good condition, the distance is short, and the waiting time is short, the correction coefficient increases, raising the score of the chosen option. Conversely, it lowers the score. This model enables recommendations to not only rely on historical preferences but also reflect real-time environmental changes, thus achieving more flexible, intelligent, and efficient energy replenishment decisions.

[0097] Specifically, for each energy replenishment facility (or method), we first examine its current status. For example, is there any available space (availability)? Is there a queue? How long is the queue? How long is the expected wait? How far is it from the current location of the mining truck? Is the current utilization rate high? Is it too busy? Is the electricity price high? Is there any PV green electricity available? This constitutes the "contextual information." Next, based on these factors, we assign a context adjustment coefficient λ to each energy replenishment method under the current conditions. j If the queue is short, the location is close, and the utilization rate is low, then λ j For higher speed (bonus points), if the queue is long, too far, or the fast charging stations are all under maintenance, then λ jThe lower score (deduction) is multiplied into the previous potential score. Finally, the corrected score is saved. Each vehicle will have an updated list that tells it: "After considering the current queue, distance, and electricity price, which option is the most worthwhile for you now?" This is the recommendation list we will use in the following steps.

[0098] In this embodiment of the invention, the context correction model dynamically adjusts the recommendation results by incorporating the real-time operating status of facilities (such as availability, queue length, waiting time, distance, electricity price, PV green electricity supply, etc.), overcoming the limitation of traditional collaborative filtering algorithms that cannot perceive environmental changes. This mechanism can weight the scores according to real-time operating conditions, making the recommendations more timely and adaptable to reality.

[0099] S205. Select the energy replenishment method with the highest correction score as the recommendation result, and record the corresponding facility number or service type to generate a personalized service recommendation list for the energy replenishment method for each unmanned mining truck.

[0100] Personalized service recommendation lists refer to a collection of service solutions with customized recommendations generated for a specific individual based on individual differences (such as behavioral preferences and status characteristics) and the real-time environment.

[0101] Optionally, the personalized service recommendation list for each mining truck includes the following: Recommended energy replenishment method: for example, fast charging, slow charging, or battery swapping. Recommended specific facility number: for example, fast charging station F05, battery swapping station B03, etc. Overall score: The final score derived based on historical preferences, similar vehicle behavior, and context correction.

[0102] In this embodiment of the invention, an improved collaborative filtering algorithm is introduced to intelligently recommend the most suitable energy replenishment method for each unmanned mining truck by integrating historical service feedback and current facility status, significantly improving the scientific rigor and personalization of energy replenishment decisions. This method not only integrates multi-dimensional utility scores (efficiency, time, cost, and health) to construct a more expressive preference model, but also solves the problems of score sparsity and cold start through vehicle similarity calculation and score prediction. Furthermore, it combines a context-correction model to perceive the actual operating status of the energy replenishment facility in real time, achieving dynamic optimization and adjustment of historical preference results. The resulting personalized service recommendation list better matches the operating needs and environmental constraints of each mining truck, providing a key basis for subsequent scheduling priority ranking and execution optimization, and improving the intelligence and adaptability of the entire charging and swapping system.

[0103] S3 determines the service execution sequence for all mining trucks based on a personalized service recommendation list and a task plan.

[0104] The service execution sequence refers to the arrangement of energy replenishment operations that are executed sequentially according to priority, after determining the energy replenishment method for each unmanned mining truck and taking into account factors such as its task plan, power urgency, time window and facility location.

[0105] In one possible implementation, S3 specifically includes:

[0106] S301 analyzes the personalized service recommendation list to obtain the recommended charging and battery swapping methods for each driverless mining truck.

[0107] S302 confirms the final charging / battery swapping method and expected service window for each mining truck.

[0108] S303, based on the recommended charging / battery swapping method, the expected service time window, and the task plan, calculate the comprehensive priority score for each unmanned mining truck:

[0109]

[0110] in, This indicates the priority of the i-th mining truck. Indicates the urgency weight of the battery charge. This represents the current remaining power of the i-th mining truck. This indicates the maximum design capacity of the vehicle's rechargeable battery. Indicates the weight of time urgency. This represents the remaining schedulable time before the i-th mining truck performs its next task. This represents the maximum allowable scheduling time. Indicates the urgency weight of distance. This represents the distance between the i-th mining truck and the target refueling facility. This indicates the maximum effective scheduling distance.

[0111] S304: Sort the overall priority scores in descending order to form a preliminary service execution sequence.

[0112] S305: Execute the initial service execution sequence and perform a resource conflict check to determine if a conflict exists. If so, proceed to S306 to resolve the conflict. Otherwise, proceed to S307.

[0113] For example, within the same time period, the number of mining trucks trying to be assigned to the same facility exceeds the facility's physical limit (e.g., there are only 3 charging spots, but 5 trucks want to go there).

[0114] S306: For minor conflicts, the unmanned mining truck with the lowest overall priority score will be recommended to be replaced by a nearby charging / swapping facility with available service capacity. For severe conflicts, the demand of unmanned mining trucks will be scheduled to available charging / swapping facilities with remaining service windows, forming an effective service execution sequence, and proceeding to S307.

[0115] S307 outputs the service execution sequence of all driverless mining trucks.

[0116] Specifically, the system first analyzes the recommended list to extract the recommended charging method for each mining truck, such as fast charging, slow charging, or battery swapping. Next, considering the vehicle's current operating status and task sequence, it confirms the final charging method for each truck and estimates its ideal service time window. Then, the system calculates a comprehensive priority score for each truck, based on three dimensions: battery urgency (lower battery levels mean higher priority), task time urgency (closer to the task execution time means higher priority), and distance from charging facilities (closer trucks are prioritized). These factors are weighted to form a priority score for ranking. The system then arranges all vehicles in descending order of priority, forming a preliminary service execution sequence. However, in real-world scenarios, resource conflicts may occur, such as multiple trucks being assigned to the same charging station at the same time, where the number of charging stations is limited. In such cases, the system automatically resolves the conflict: for minor conflicts, lower-priority vehicles are directed to adjacent, available similar facilities. For severe conflicts, the service time or charging location of these vehicles is rescheduled until resource matching is conflict-free.

[0117] In this embodiment of the invention, a unified comprehensive priority scoring mechanism is constructed by comprehensively considering multiple key factors such as the task time constraints, power urgency, and distance to refueling facilities of mining trucks, based on personalized service recommendation results. This mechanism generates a service execution sequence across the entire mining area in an orderly manner. This sequence not only ensures that mining trucks with critical power levels, urgent tasks, or better geographical locations receive service resources first, but also effectively solves the service congestion or allocation conflicts that may occur under limited facility capacity conditions through resource conflict checking and hierarchical resolution strategies, thereby improving resource utilization efficiency and task scheduling coordination.

[0118] S4, based on the service execution sequence, calculates the expected service capacity and revenue assessment report for each charging and swapping facility.

[0119] In one possible implementation, S4 specifically includes:

[0120] S401, based on the service execution sequence, counts the number of reserved mining trucks for each charging and swapping facility during the entire scheduling cycle.

[0121] S402 calculates the expected service capacity and remaining service capacity of each facility based on the number of reserved mining trucks and the equipment parameters of each charging and swapping facility.

[0122] Optionally, remaining service capacity represents the difference between expected service capacity and the actual number of tasks assigned.

[0123] S403 calculates the service load rate of each charging and swapping facility during the scheduling cycle based on the number of reserved mining trucks.

[0124] Alternatively, the actual number of reservations can be divided by the maximum service capacity.

[0125] S404 calculates the expected revenue of each charging and swapping facility based on the number of reserved mining trucks, and sums up the expected revenue of each facility to obtain the total comprehensive revenue.

[0126] Alternatively, the revenue model may consider unit service fee × number of services.

[0127] S405 structures and organizes the expected service capacity, remaining service capacity, service load rate, expected revenue, and total revenue to generate a revenue assessment report for each charging and swapping facility.

[0128] Specifically, the revenue assessment report integrates the facility's reservation service data, physical capacity limitations, service load, and potential revenue performance during the current scheduling cycle, providing visualized and data-driven analysis results to facilitate the management system's decisions on subsequent scheduling, coordination, expansion, or price adjustments.

[0129] In this embodiment of the invention, calculating the expected service capacity and revenue assessment report for each facility not only helps to identify potential resource bottlenecks and redundancies, but also provides a quantitative basis for cross-regional scheduling, revenue-oriented optimization, and dynamic pricing strategies, thereby improving the resource allocation efficiency, service balance, and operational revenue level of the entire energy replenishment system.

[0130] S5 combines vehicle status data, infrastructure status data, expected service capacity, and revenue assessment reports to generate a mining truck dispatch instruction set through a deep reinforcement learning algorithm.

[0131] Among them, deep reinforcement learning algorithm is an intelligent optimization method that combines the decision-making mechanism of reinforcement learning with the perception capability of deep neural networks. It can learn the optimal strategy by continuously interacting with the environment in a complex, high-dimensional state space.

[0132] In one possible implementation, S5 specifically includes:

[0133] S501 models the scheduling decisions for charging and battery swapping facilities as a Markov decision process based on vehicle status data, infrastructure status data, expected service capacity, and revenue assessment reports.

[0134] Specifically, define the state space, action space, and reward function.

[0135] Optionally, the state space is defined as follows: the operational state of a facility is described as a discretized set of states. The state space includes the following state variables: the current queue length, the current utilization rate, the expected waiting time, and the renewable energy power supply capacity (PV Power). These variables, after quantization and discretization, form a finite set of states, i.e., the state space.

[0136]

[0137] in, Representing the state space, This represents the remaining power of the i-th mining truck. This indicates the priority of the i-th mining truck. This represents the current queue length for the l-th facility. PVPower represents the current utilization rate of the l-th facility, W represents the renewable energy power supply capacity, and W represents the number of mining trucks to be dispatched.

[0138] Definition of the action space: When a facility receives a scheduling request, it can choose from the following three actions: Accept the request; Reject the request; or Redirect to another facility. These actions constitute the action space.

[0139]

[0140] in, Represents the action space. This indicates that the service request is assigned to the j-th facility, M represents the total number of facilities, and Reject indicates that the service request is rejected.

[0141] Definition of reward function: Construct the following immediate reward function to evaluate the overall benefit brought about by performing a certain action:

[0142]

[0143] in, Represents the reward function, , , Both represent adjustable weighting coefficients. This indicates the service revenue of mining car i in facility l. This represents the waiting cost factor per unit of time. Let represent the expected waiting time of mine car i at facility l, and max represent maximizing . This indicates the real-time utilization rate of facility l. This indicates the maximum utilization rate.

[0144] S502, based on Markov decision process, uses a deep Q-network to train the mine truck allocation strategy through reinforcement learning to generate the optimal allocation strategy function.

[0145] Deep Q-Network (DQN) is a classic deep reinforcement learning algorithm that combines traditional Q-learning methods with deep neural networks to solve optimal decision-making problems in high-dimensional state spaces. In DQN, the agent approximates the Q-value function of the state-action pair through a neural network, that is, learns a function Q(s,a;θ) to estimate the expected reward obtained after performing action a in state s.

[0146] Specifically, a neural network is used to fit the optimal action-value function, denoted as: ,in, This indicates the action to be performed under the current state s estimated by the neural network. value, This represents the maximum expected reward obtainable by following the optimal policy after taking action a in state s. The following loss function is defined to guide parameter updates:

[0147]

[0148] in, Represents the loss function. Here, E represents the network parameters, r represents the expected value, r represents the immediate reward obtained in the current step, and γ represents the discount factor. Indicates the state at the next moment. Indicates the state One of all possible actions, Indicates the target network parameters. This represents the maximum Q-value for all actions in the next state within the target network. During the learning process, the state-action-reward-next state tuple (...) is used... , , , ),in, The state space for the next time step is stored in the experience replay pool. Then, random sampling is used for model training to improve sample utilization efficiency and reduce correlation between samples. A soft update strategy is used to update the target network parameters, with the following update formula: ,in, Let θ represent the target network parameters, τ represent the soft update coefficients, and θ represent the current main network parameters. After training, an approximately optimal dispatch policy function can be obtained to guide the response behavior of each facility to scheduling requests. For each mining vehicle to be scheduled, the effective action with the highest Q value is selected.

[0149] S503: Sequentially acquire the vehicle status data and current environmental status of each unmanned mining truck to be dispatched, and construct the current state input vector.

[0150] S504: Input the current state input vector into the optimal assignment policy function to obtain the action values ​​of all possible actions.

[0151] S505: Select the action with the largest effective action value as the current minecart's scheduling response action.

[0152] Specifically, if the selected action is a specific facility F l If the result is positive, the minecart will be assigned to that facility. If the result is negative, adjustments will be made based on the suboptimal action or readjustment strategy.

[0153] S506 summarizes all the scheduling response actions of the mining trucks into a mining truck dispatch instruction set.

[0154] Optionally, the mining truck dispatch instruction set refers to the set of charging and battery swapping task instructions generated by the system for each unmanned mining truck to be dispatched after comprehensively considering the mining truck status, facility load, task plan, and optimization strategy. It includes the following fields: mining truck number, dispatch method, target facility number, priority score, and instruction status.

[0155] For example, vehicle number: V102, energy replenishment method: battery swapping, target facility: B03, dispatch priority: 0.86, operation instruction: proceed immediately, queue up and perform battery swapping operation.

[0156] In this embodiment of the invention, by modeling the mine truck allocation problem as a Markov decision process and introducing the deep Q-network algorithm from deep reinforcement learning for training and inference, this implementation can autonomously learn the optimal mine truck allocation strategy in a complex environment with multiple variables, high dimensionality, and dynamic changes. Compared with traditional rule-based scheduling or static optimization methods, this method can not only make full use of the real-time status of mine trucks and the operating status of facilities, but also dynamically adapt to fluctuations in service capacity and changes in resource supply and demand, thereby achieving precise allocation of each mine truck to be scheduled. In addition, by introducing a reward function and experience replay mechanism during the training process, the system can continuously accumulate experience and optimize strategies, so that the scheduling result achieves a globally optimal balance between service revenue, waiting time, and resource load.

[0157] S6, through a multi-aggregator collaboration mechanism, coordinates and optimizes the mine truck dispatch instruction set to obtain a cross-regional collaborative scheduling instruction set.

[0158] The multi-aggregator collaboration mechanism is an intelligent coordination strategy for regional distributed resource management, designed to address issues such as local facility overload and scheduling bottlenecks. Under this mechanism, the system monitors the operational status of charging and swapping facilities within each aggregator's jurisdiction in real time, including load rate, queue length, and service capacity, and determines whether resource overload exists. If a region experiences excessive scheduling pressure, the system automatically initiates a cross-regional collaboration request, transferring some scheduling tasks to neighboring aggregator jurisdictions with available resources. Through task and facility matching optimization, the collaboration mechanism achieves cross-regional resource complementarity and task balancing, effectively alleviating service congestion, improving overall resource utilization and scheduling response efficiency, and ensuring the entire system maintains high efficiency and stability even in complex and dynamic environments.

[0159] In one possible implementation, S6 specifically includes:

[0160] S601 monitors the operating status of all charging and swapping facilities within the jurisdiction of each aggregator, including load index, queue length, and estimated waiting time.

[0161] S602, determine whether the load index of each aggregator's jurisdiction exceeds the preset load index. If so, proceed to S603 to trigger the cross-aggregator cooperation mechanism. Otherwise, execute the original minecart dispatch instruction set.

[0162] It should be noted that those skilled in the art can set the size of the preset load index according to actual needs, and this invention does not limit this.

[0163] S603, the resource-overloaded sending aggregator region sends a cooperation request to the resource-unoverloaded receiving aggregator region. The cooperation request includes a list of mining trucks that cannot be scheduled and the corresponding resource requirement type.

[0164] Optionally, the load index is calculated as the ratio between the number of actual scheduled tasks and the maximum service capacity of the resource during the scheduling period.

[0165] S604, based on the facility status within each receiver aggregator's jurisdiction, select receiver aggregator jurisdictions with sufficient resource reserves to complete the requested task.

[0166] S605, based on resource availability and distance cost, uses a lightweight matching algorithm to perform collaborative matching between the sending aggregator's jurisdiction and the receiving aggregator's jurisdiction, forming a cross-aggregator resource allocation scheme.

[0167] Optionally, a task-resource bipartite graph is constructed between the task set I submitted by the requesting aggregator (denoted as Aggregator_A) and the facility set L with remaining resources within the jurisdiction of the responding aggregators (such as Aggregator_B and Aggregator_C). The following objective function is used for task allocation modeling:

[0168]

[0169]

[0170] Where min represents minimizing, I represents the set of mining truck tasks that need to be reallocated within the resource overload aggregator's jurisdiction, and L represents the set of charging and swapping facilities with remaining service capacity within the responder aggregator's jurisdiction. Let $\frac{i}{i}$ be the matching cost incurred in assigning mine car $i$ to facility $l$. ∈{0,1} indicates whether to assign mine car i to facility l. This represents the distance between mine car i and facility l. This represents the cost coefficient per unit distance. This indicates the current estimated waiting time for facility l. This represents the cost coefficient per unit waiting time. This indicates the current resource utilization rate of facility l. This represents the utilization rate adjustment factor.

[0171] The constraints are:

[0172]

[0173]

[0174] in, This indicates the remaining service capacity of facility l within the current scheduling period.

[0175] Under the constraints of each task being allocated at most once and each facility's capacity not being overloaded, the allocation scheme that minimizes the objective function value is solved to form a cross-aggregator resource allocation scheme.

[0176] S606 updates the mine truck dispatch instruction set according to the cross-aggregator resource allocation scheme, and generates a cross-regional collaborative scheduling instruction set.

[0177] It should be noted that by introducing a multi-aggregator collaboration mechanism, this implementation method can effectively address issues such as service congestion and queuing exceeding limits caused by localized facility resource shortages during the scheduling process. Under this mechanism, the system monitors operational status indicators such as facility load index, queue length, and waiting time in real time within each aggregator's jurisdiction. Once it detects that the resource pressure in a certain area exceeds a set threshold, a collaboration process is immediately triggered, transferring some of the energy replenishment tasks of the scheduled mining trucks across regions to neighboring aggregator jurisdictions with surplus resources. This mechanism optimizes the model based on the matching cost between tasks and facilities, comprehensively considering distance, waiting time, and facility utilization, and uses a lightweight algorithm to obtain a globally optimal or suboptimal task redistribution scheme. Finally, the original assignment instruction set is adjusted according to the matching results to generate a cross-regional collaborative scheduling instruction set, achieving dynamic resource complementarity and orderly balanced task distribution.

[0178] In this embodiment of the invention, the mining truck dispatch instruction set is coordinated and optimized through a multi-aggregator cooperation mechanism. This not only alleviates local scheduling peaks and reduces the waiting time and service failure rate of mining trucks, but also improves the resource utilization efficiency and service response capability of the entire system. This makes the scheduling strategy more flexible, coordinated and robust, ensuring that it can maintain an efficient and stable operating level in a complex and ever-changing mining environment.

[0179] In one possible implementation, the process after S6 and before S7 includes:

[0180] When reacquiring and executing cross-regional collaborative scheduling instruction sets, the expected revenue and expected waiting time for each unmanned mining vehicle, as well as the service load rate of each charging and swapping facility, are considered.

[0181] S7: Calculate the comprehensive service performance index when executing the cross-regional collaborative scheduling instruction set, and determine whether the comprehensive service performance index is greater than the preset service performance index. If yes, proceed to S8. Otherwise, trigger the system degradation strategy and readjust the mining truck dispatch instruction set.

[0182] Among them, the comprehensive service performance index is a multi-dimensional comprehensive evaluation index used to assess the overall scheduling effect and system operation quality. It aims to fully reflect the performance of the current scheduling scheme in terms of profitability, resource utilization, service efficiency and user satisfaction.

[0183] It should be noted that those skilled in the art can set the size of preset service performance indicators according to actual needs, and this invention does not limit this.

[0184] In one possible implementation, the calculation method for the comprehensive service performance index is as follows:

[0185]

[0186] in, Indicates comprehensive service performance indicators. Indicates the return weight. Indicates total revenue. Indicates the theoretical maximum benefit, Indicates the weight of resource utilization. This represents the average service load rate. Indicates the service efficiency weight. This indicates the average expected waiting time. Indicates the maximum allowed waiting time. Indicates the weight of satisfaction. This represents the average satisfaction rating. This represents the highest satisfaction rating.

[0187] It should be noted that when the mining truck assignment instruction set is continuously triggered for adjustment, in order to avoid scheduling oscillations, frequent changes and execution instability, the system introduces a dynamic adjustment suppression mechanism. Specifically, this includes setting a minimum change threshold. When the change ratio of the new round of instructions compared with the previous round is lower than a preset threshold (such as 5%), the update is not triggered, ensuring that the scheduling instructions have execution stability and convergence while responding to dynamic changes.

[0188] In this embodiment of the invention, the system determines whether the current scheduling scheme meets the service quality requirements based on a set performance threshold. If it does not meet the threshold, it automatically returns to the assignment stage for optimization and readjustment, thus forming a closed-loop mechanism of "evaluation to feedback to optimization". The advantage of this approach is that it not only provides a scientific and unified performance criterion, effectively avoiding resource bias or service imbalance under a single objective (such as maximizing revenue), but also enables the scheduling system to have self-correction and dynamic adjustment capabilities, significantly improving the stability of the scheduling strategy.

[0189] S8 takes the cross-regional collaborative scheduling instruction as the final scheduling scheme and executes the final scheduling scheme to complete the collaborative optimization of charging and swapping.

[0190] The beneficial effects of this invention are as follows:

[0191] In this embodiment of the invention, an improved collaborative filtering algorithm is introduced to deeply integrate multi-source data such as the operating status of mining trucks and the status of infrastructure. This intelligently recommends the optimal energy replenishment method for each unmanned mining truck, enabling flexible selection and unified modeling between charging and battery swapping, overcoming the limitations of fragmented processing in traditional strategies. Based on this, the system gradually constructs a global collaborative scheduling mechanism with adaptive and closed-loop iterative capabilities, from multi-vehicle service ranking, infrastructure service capacity assessment, and deep reinforcement learning-based scheduling decisions to cross-regional collaborative optimization involving multiple aggregators. This mechanism can dynamically perceive key elements such as task plans, power consumption, power station load, and economic benefits, achieving efficient, flexible, and globally optimized charging and battery swapping scheduling strategies, significantly improving the system's resource utilization efficiency and overall operational performance.

[0192] System Implementation Examples

[0193] Reference manual attached Figure 2 The diagram shows a structural schematic of a charging and swapping collaborative optimization system for multiple unmanned mining trucks provided by an embodiment of the present invention.

[0194] This invention proposes a charging and swapping collaborative optimization system 30 for multiple unmanned mining trucks, including: a memory 303 and a processor 301.

[0195] The memory 303 stores an application program adapted to be executed by the processor 301 to implement the charging and swapping collaborative optimization method for multiple unmanned mining vehicles in the method embodiment.

[0196] The charging and swapping collaborative optimization system 30 for multiple unmanned mining trucks includes a processor 301 and a memory 303. The processor 301 and the memory 303 are connected, for example, via a bus 302.

[0197] The structure of the charging and swapping collaborative optimization system 30 for multiple unmanned mining trucks does not constitute a limitation on the embodiments of the present invention.

[0198] Processor 301 may be a CPU, a general-purpose processor, a DSP, an ASIC, an FPGA, or other programmable logic device, transistor logic device, hardware component, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in connection with this disclosure. Processor 301 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0199] Bus 302 may include a pathway for transmitting information between the aforementioned components. Bus 302 may be a PCI bus or an EISA bus, etc. Bus 302 may be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the figure, but this does not mean that there is only one bus or one type of bus.

[0200] The memory 303 may be a ROM or other type of static storage device capable of storing static information and instructions, RAM or other type of dynamic storage device capable of storing information and instructions, or it may be an EEPROM, CD-ROM or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto.

[0201] Computer-readable storage medium embodiments

[0202] This invention proposes a computer-readable storage medium storing a computer program that can be loaded and executed by a processor for a first aspect of a charging and swapping collaborative optimization method for multiple unmanned mining vehicles.

[0203] The applicant of this invention has provided a detailed description of the embodiments of the invention in conjunction with the accompanying drawings. However, those skilled in the art should understand that the above embodiments are merely preferred embodiments of the invention. The detailed description is only intended to help readers better understand the spirit of the invention and is not intended to limit the scope of protection of the invention. On the contrary, any improvements or modifications made based on the inventive spirit of the invention should fall within the scope of protection of the invention.

[0204] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the protection scope of the present invention.

Claims

1. A method for collaborative optimization of charging and battery swapping for multiple unmanned mine vehicles, characterized in that, The charging and battery swapping collaborative optimization method for multiple unmanned mining vehicles includes: S1 collects vehicle status data, infrastructure status data, and task plans from multiple unmanned mining trucks. S2, based on the vehicle status data and the infrastructure status data, an improved collaborative filtering algorithm is used to recommend energy replenishment methods for each of the unmanned mining trucks and generate a personalized service recommendation list; S3. Based on the personalized service recommendation list and the task plan, determine the service execution sequence for all mining trucks. S4. Based on the service execution sequence, calculate the expected service capacity and revenue assessment report for each charging and swapping facility; S5, combining the vehicle status data, the infrastructure status data, the expected service capacity, and the revenue assessment report, a mining truck dispatch instruction set is generated using a deep reinforcement learning algorithm; S6, through a multi-aggregator collaboration mechanism, the mining truck dispatch instruction set is coordinated and optimized to obtain a cross-regional collaborative scheduling instruction set; S7, calculate the comprehensive service performance index when executing the cross-regional collaborative scheduling instruction set, and determine whether the comprehensive service performance index is greater than the preset service performance index; if yes, proceed to S8; otherwise, trigger the system degradation strategy and readjust the mining truck dispatch instruction set. S8, the cross-regional collaborative scheduling instruction is used as the final scheduling scheme, and the final scheduling scheme is executed to complete the charging and swapping collaborative optimization; S3 specifically includes: S301, parse the personalized service recommendation list to obtain the recommended charging and battery swapping method for each of the unmanned mining trucks; S302, confirm the final charging and battery swapping method and expected service time window for each mining truck; S303, based on the recommended charging and swapping method, the final determined charging and swapping method, and the expected service time window, and in conjunction with the task plan, calculate the comprehensive priority score for each of the unmanned mining vehicles; S304, Sort the comprehensive priority scores in descending order to form a preliminary service execution sequence; S305, execute the preliminary service execution sequence and perform resource conflict check to determine if there is a conflict; if so, proceed to S306 to resolve the conflict; otherwise, proceed to S307. S306, For minor conflicts, the unmanned mining truck with the lowest comprehensive priority score is recommended to be replaced by a nearby charging and swapping facility of the same type with service capacity; for severe conflicts, the demand of unmanned mining trucks is scheduled to available charging and swapping facilities with remaining service windows to form an effective service execution sequence, and proceed to S307. S307, Output the service execution sequence containing all driverless mining vehicles.

2. The method of claim 1, wherein, The vehicle status data includes remaining battery power, battery health, current location, and priority. The infrastructure status data includes the availability of charging piles, the availability of fast charging stations, the availability of battery swapping stations, queue length, and estimated waiting time. The task plan includes the task number, task type, planned departure time, task duration, and latest acceptable arrival time.

3. The method of claim 1, wherein, The energy replenishment methods include fast charging, slow charging, and battery swapping; S2 specifically includes: S201, Based on the service records and feedback information of multiple unmanned mining vehicles related to different energy replenishment methods during the historical task execution process, a scoring matrix is ​​constructed; S202, based on the vehicle status data, calculate the similarity index between mining trucks using the Pearson correlation coefficient, and generate a similarity matrix between each of the unmanned mining trucks based on the similarity index. S203, using the similarity matrix and the rating matrix, estimate the service items with missing ratings to obtain the potential rating values ​​of the current unmanned mining truck for various unused energy replenishment methods; S204, Based on the infrastructure status data, the potential score value is weighted and adjusted using a preset context correction model to generate a corrected score value that takes into account the actual operating environment; S205, select the energy replenishment method with the highest correction score as the recommendation result, and record the corresponding facility number or service type to generate a personalized service recommendation list of energy replenishment methods for each unmanned mining truck.

4. The method of claim 1, wherein, S4 specifically includes: S401, Based on the service execution sequence, count the number of reserved mining trucks for each of the charging and swapping facilities during the entire scheduling cycle; S402, based on the number of reserved mining trucks and the equipment parameters of each charging and swapping facility, calculate the expected service capacity and remaining service capacity of each facility; S403, Based on the number of reserved mining trucks, calculate the service load rate of each of the charging and swapping facilities during the scheduling cycle; S404, Based on the number of reserved mining trucks, calculate the expected revenue of each of the charging and swapping facilities, and summarize the expected revenue of each facility to obtain the total comprehensive revenue; S405, the expected service capacity, the remaining service capacity, the service load rate, the expected revenue, and the total revenue are structured and organized to generate a revenue assessment report for each charging and swapping facility.

5. The method of claim 1, wherein, S5 specifically includes: S501, based on the vehicle status data, the infrastructure status data, the expected service capacity, and the revenue assessment report, the scheduling decision of the charging and swapping facilities is modeled as a Markov decision process; S502, Based on the Markov decision process, the mine truck allocation strategy is trained by reinforcement learning through a deep Q-network to generate the optimal allocation strategy function. S503: Sequentially acquire the vehicle status data and current environmental status of each unmanned mining truck to be dispatched, and construct the current state input vector; S504, Input the current state input vector into the optimal assignment strategy function to obtain the action values ​​of all possible actions; S505, Select the largest valid action among the action values ​​as the current minecart's scheduling response action; S506, summarizes all the scheduling response actions of the mining trucks into the mining truck dispatch instruction set.

6. The method of claim 1, wherein, S6 specifically includes: S601, monitor the operating status of all charging and swapping facilities within the jurisdiction of each aggregator, wherein the operating status includes load index, queue length and expected waiting time; S602, determine whether the load index of each aggregator area exceeds the preset load index; if so, proceed to S603 to trigger the cross-aggregator cooperation mechanism; otherwise, execute the original minecart dispatch instruction set. S603, the resource-overloaded sending aggregator region sends a cooperation request to the resource-unoverloaded receiving aggregator region. The cooperation request includes a list of mining trucks that cannot be arranged and the corresponding resource demand type. S604, Based on the facility status within each of the receiving aggregator jurisdictions, select the receiving aggregator jurisdictions that have sufficient resources to complete the requested task. S605, based on resource availability and distance cost, uses a lightweight matching algorithm to perform collaborative matching between the sending aggregator's jurisdiction and the receiving aggregator's jurisdiction to form a cross-aggregator resource allocation scheme; S606, Update the mine truck dispatch instruction set according to the cross-aggregator resource allocation scheme to generate the cross-regional collaborative scheduling instruction set.

7. The method of claim 5, wherein, The process includes the following steps after S6 and before S7: When re-acquiring and executing the cross-regional collaborative scheduling instruction set, the expected revenue and expected waiting time of each of the unmanned mining vehicles, as well as the service load rate of each charging and swapping facility, are determined.

8. The method of claim 1, wherein, The calculation method for the comprehensive service performance indicators is as follows: in, Indicates comprehensive service performance indicators. Indicates the return weight. Indicates total revenue. Indicates the theoretical maximum benefit, Indicates the weight of resource utilization. This represents the average service load rate. Indicates the service efficiency weight. This indicates the average expected waiting time. Indicates the maximum allowed waiting time. Indicates the weight of satisfaction. This represents the average satisfaction rating. This represents the highest satisfaction rating.

9. A charging and swapping coordination optimization system for multiple unmanned mine cars, characterized in that, include: Memory and processor; The memory stores an application program adapted to be executed by the processor to implement the charging and swapping collaborative optimization method for multiple unmanned mining vehicles as described in any one of claims 1 to 8.

Images