A mine truck intelligent scheduling method, system, device and medium

Abstract

The application relates to a mine truck intelligent scheduling method, system, device and medium, which comprises the following steps: collecting real-time state data of all transport trucks in a mine operation area, obtaining a road network node latitude and longitude coordinate sequence, and generating a road topology structure. A production task instruction is received, and a mixed integer programming model containing load constraints, time constraints and task coverage constraints is constructed in combination with the equipment maintenance plan, real-time state data, road topology and task instruction. The model is solved to generate a loading / unloading point distribution scheme for each truck. According to the distribution scheme and real-time road traffic data, a driving path point sequence is planned for each truck. The loading point, unloading point and path sequence are packaged into a scheduling instruction. The actual driving trajectory and operation state of the truck are continuously monitored, and when the trajectory deviation exceeds the threshold or the operation is abnormal, the loading / unloading point is redistributed and the path sequence is updated. The application improves the efficiency and stability of the mine transportation operation.

Description

Technical Field

[0001] This application relates to the field of intelligent scheduling technology, and in particular to an intelligent scheduling method, system, equipment and medium for mining trucks. Background Technology

[0002] With the continuous expansion of open-pit mining and increasingly stringent safety requirements, the transportation link in mining production, as a crucial process connecting mining and processing, directly impacts overall operational efficiency in terms of efficiency and safety. Traditional mine truck dispatching relies heavily on dispatchers' manual experience, issuing instructions to drivers via walkie-talkies and other communication tools. Dispatch decisions are often based on static production plans and limited on-site awareness, lacking systematic integration and analysis of real-time vehicle locations, road congestion, and equipment health status. This manual-based dispatching model may suffice in simple operating environments, but it struggles to achieve precise allocation and efficient coordination of transportation resources in the complex, large-scale, multi-equipment, and dynamically changing mining environment.

[0003] Although some mines have introduced basic vehicle positioning and monitoring systems in recent years, enabling electronic map display of vehicle locations, the level of intelligence in the scheduling process remains insufficient. Existing methods mostly focus on passive monitoring rather than proactive optimization, failing to dynamically adjust based on real-time changes in road conditions, vehicle status, and production progress. For example, when a transport route is blocked due to congestion or temporary maintenance, the system struggles to quickly replan the optimal detour route for affected vehicles; when individual trucks deviate from their original schedule due to mechanical failure or task execution deviations, there is a lack of effective mechanisms to quickly reassign their remaining tasks to other available vehicles, resulting in overall efficiency losses due to idle equipment and interrupted material supply to some work sites. Furthermore, traditional scheduling systems often isolate production task allocation, route planning, and equipment maintenance management, making theoretically feasible scheduling plans significantly less effective in practice due to unforeseen conflicts, thus hindering further improvements in mine transportation efficiency and optimization of operating costs. Summary of the Invention

[0004] To address the aforementioned technical issues, this application provides a method, system, equipment, and medium for intelligent scheduling of mining trucks.

[0005] Firstly, this application provides an intelligent scheduling method for mining trucks, employing the following technical solution:

[0006] A method for intelligent scheduling of mining trucks, the scheduling method comprising:

[0007] Collect real-time status data of all transport trucks in the mining operation area, including location data, vehicle speed data, and load data;

[0008] Obtain the latitude and longitude coordinate sequence of each node in the mine road network and generate the road topology;

[0009] Receive production task instruction data, including the location of the loading point, the location of the unloading point, and the required amount of ore;

[0010] Acquire equipment maintenance plan data, and combine it with the real-time status data, road topology and production task instruction data to construct a mixed integer programming model that includes load constraints, time constraints and task coverage constraints;

[0011] Solve the mixed integer programming model to generate a loading / unloading point allocation scheme for each truck;

[0012] Based on the loading / unloading point allocation scheme and real-time road traffic flow data, a sequence of driving route points is planned for each truck;

[0013] The loading point location, unloading point location, and driving path point sequence are encapsulated into a scheduling instruction and sent to the corresponding truck's on-board terminal.

[0014] The system continuously acquires the actual driving trajectory data and operation status of each truck. When a trajectory deviation exceeds a threshold or an abnormal operation is detected, the system triggers the reallocation of loading / unloading points and updates the driving path point sequence.

[0015] By adopting the above technical solution, the traditional static scheduling method, which relies on manual experience, is upgraded to a data-driven, model-based, and dynamically optimized closed-loop adaptive system. This method achieves optimal matching of resources and tasks from a global perspective through a mixed-integer programming model, improving the efficiency of vehicle-shovel coordination and equipment utilization. Dynamic path planning combined with real-time road conditions effectively reduces empty mileage, waiting time, and congestion, directly lowering operating costs and energy consumption. The introduced closed-loop monitoring and rescheduling mechanism greatly enhances the system's robustness in responding to unexpected situations on-site, ensuring the continuity and stability of production operations. Ultimately, this technical solution maximizes the efficiency of mine transportation operations, streamlines operating costs, and achieves agility and high reliability in the production process.

[0016] Secondly, this application provides an intelligent dispatching system for mining trucks, which adopts the following technical solution:

[0017] A mining truck intelligent dispatching system, the dispatching system comprising:

[0018] The status data acquisition module is used to collect real-time status data of all transport trucks in the mining operation area, including location data, vehicle speed data, and load data.

[0019] The road topology generation module is used to obtain the latitude and longitude coordinate sequence of each node in the mine road network and generate the road topology.

[0020] The production task instruction receiving module is used to receive production task instruction data, including the location of the loading point, the location of the unloading point, and the required amount of ore.

[0021] The scheduling optimization model construction module is used to acquire equipment maintenance plan data, and combine the real-time status data, road topology and production task instruction data to construct a mixed integer programming model that includes load constraints, time constraints and task coverage constraints.

[0022] The task allocation scheme solving module is used to solve the mixed integer programming model and generate a loading / unloading point allocation scheme for each truck.

[0023] The route planning module is used to plan a sequence of driving route points for each truck based on the loading / unloading point allocation scheme and real-time road traffic flow data;

[0024] The scheduling instruction issuing module is used to encapsulate the loading point location, unloading point location, and driving path point sequence into a scheduling instruction and issue it to the corresponding truck's on-board terminal.

[0025] The operation monitoring and rescheduling module is used to continuously acquire the actual driving trajectory data and operation status of each truck. When the trajectory deviation exceeds the threshold or the operation is abnormal, the loading / unloading point is reassigned and the driving path point sequence is updated.

[0026] Thirdly, this application provides a computer device, which adopts the following technical solution:

[0027] A computer device includes a memory, a processor, and a computer program stored in the memory, the processor executing the computer program to perform the steps of the method as described in the first aspect.

[0028] Fourthly, this application provides a computer-readable storage medium, which adopts the following technical solution:

[0029] A computer-readable storage medium storing a computer program that can be loaded by a processor and executed as in any of the methods in the first aspect. Attached Figure Description

[0030] Figure 1 This is a schematic diagram of the first process of a mining truck intelligent scheduling method according to one embodiment of this application.

[0031] Figure 2 This is a second flowchart of a mining truck intelligent scheduling method according to one embodiment of this application.

[0032] Figure 3This is a schematic diagram of the third process of a mining truck intelligent scheduling method according to one embodiment of this application.

[0033] Figure 4 This is a schematic diagram of the fourth process of a mining truck intelligent scheduling method according to one embodiment of this application.

[0034] Figure 5 This is a schematic diagram of the fifth process of a mining truck intelligent scheduling method according to one embodiment of this application.

[0035] Figure 6 This is a schematic diagram of the sixth process of a mining truck intelligent scheduling method according to one embodiment of this application.

[0036] Figure 7 This is a schematic diagram of the seventh process of a mining truck intelligent scheduling method according to one embodiment of this application. Detailed Implementation

[0037] To make the purpose, technical solution, and advantages of this application clearer, the following description is provided in conjunction with the appendix. Figures 1-7 The present application will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the application.

[0038] This application discloses an intelligent scheduling method for mining trucks.

[0039] Reference Figure 1 A method for intelligent scheduling of mining trucks, the scheduling method includes:

[0040] Step S101: Collect real-time status data of all transport trucks in the mining operation area, including location data, vehicle speed data, and load data;

[0041] The system continuously captures key operational parameters through an onboard sensor network deployed on each truck. Location data (typically from GPS / BeiDou systems, combined with inertial navigation to compensate for brief signal loss) provides each truck's precise coordinates in the mine's three-dimensional geographic space, a prerequisite for any spatial association and route planning. Speed ​​data is not only used to monitor driving safety and compliance but is also a crucial input variable for assessing road efficiency, predicting arrival times, and subsequent real-time route optimization. Load data (acquired through onboard weighing sensors) directly reflects the truck's real-time operating load, serving as a core basis for measuring transportation efficiency and calculating actual output, and is a key input for meeting the "load constraint" when building the next optimization model.

[0042] Step S102: Obtain the latitude and longitude coordinate sequence of each node in the mine road network and generate the road topology structure;

[0043] In this system, a simple set of latitude and longitude coordinates consists of discrete points and lacks correlation. By organizing these latitude and longitude coordinate sequences, which represent road intersections and turning points, according to their actual connection relationships, the system generates a network graph consisting of "nodes" and "edges," i.e., the road topology.

[0044] In this structure, nodes represent key locations on the road, and edges represent road segments connecting nodes. Attributes such as base length, default speed, and gradient can be assigned to these edges. This topology forms the basis for all path planning algorithms; it abstracts the complex physical road structure, allowing the system to use graph theory algorithms (such as the A* algorithm) to quickly search for feasible paths between two points.

[0045] Step S103: Receive production task instruction data, including the location of the loading point, the location of the unloading point, and the required amount of ore;

[0046] Among these, production task instruction data serves as the target and driving force for scheduling, specifying "from where" (loading point location), "to where" (unloading point location), and "how much" (ore demand). This instruction data concretizes the abstract production capacity target into a series of transportation tasks with spatial and quantitative attributes. Upon receiving these instructions, the system's core mission is to efficiently and economically complete these tasks while satisfying all constraints. The location information of loading / unloading points is correlated with the road topology to determine their network position. The ore demand serves as the direct basis for the "task coverage constraint" when subsequently constructing the mixed-integer programming model; that is, the total transportation volume of all trucks must meet the demand of each unloading point to ensure the successful completion of the production plan.

[0047] Step S104: Obtain equipment maintenance plan data, and combine it with real-time status data, road topology and production task instruction data to construct a mixed integer programming model that includes load constraints, time constraints and task coverage constraints.

[0048] The underlying logic of this step is to transform the complex systems engineering problem of mining truck scheduling into a mathematical optimization problem that can be precisely solved by a computer. Mixed-Integer Programming (MIP) is a mathematical optimization technique suitable for complex optimization problems where decision variables include both continuous variables (such as transport volume and travel time) and integer variables (such as 0-1 variables, representing whether to assign a task to a truck).

[0049] In this embodiment, the system integrates all prior information: real-time status data (to understand the current capabilities and location of the fleet), road topology (to provide feasible mobile networks), production task instruction data (to clarify operational objectives), and equipment maintenance plan data (to avoid scheduling vehicles that are about to be maintained for long-term tasks and improve the feasibility of the plan) to build this model.

[0050] Specifically, the constraints in the model are concrete manifestations of business rules: load constraints ensure that the tasks assigned to each truck do not exceed its rated load capacity, which is a physical limitation; time constraints may require that specific urgent tasks must be completed within a specified time, which is an efficiency requirement; and task coverage constraints ensure that all production tasks (ore demand) are allocated, which is a rigid requirement of production targets. The objective function is usually to maximize output or minimize total cost (such as total transportation time or total fuel consumption). By constructing this mixed-integer programming model, the system formalizes and mathematically represents the dispatcher's experience and scheduling rules, transforming scheduling decisions from experience-based fuzzy decisions into precise, interpretable optimal or near-optimal decisions based on data and optimization theory.

[0051] Step S105: Solve the mixed integer programming model to generate a loading / unloading point allocation scheme for each truck;

[0052] In this embodiment, the solver (such as CPLEX, Gurobi, etc.) employs advanced algorithms such as branch and bound and cutting plane methods to intelligently search and ultimately find a solution that optimizes the objective function (such as total cost) while satisfying all constraints within a vast space of possible solutions (i.e., all possible allocation methods for all trucks and all tasks). This solution specifically specifies which truck should load at which loading point and then unload at which unloading point within the upcoming scheduling cycle, thus generating a loading / unloading point allocation scheme. This is the top-level planning of the scheduling decision, which optimally matches task requirements with transportation resources from a global perspective. It is the core element for achieving optimal matching and reducing empty runs and waiting times.

[0053] Step S106: Based on the loading / unloading point allocation scheme and real-time road traffic flow data, plan the driving route point sequence for each truck;

[0054] This system introduces dynamic variables, namely real-time traffic flow data, into the road topology. For each truck, the system plans an optimal route from its current location to the designated loading and unloading point. "Optimal" here is typically defined by a cost function that considers both static factors (such as base distance and road gradient) and dynamic factors (such as congestion and queuing caused by real-time traffic flow). The algorithm searches the topology network, calculates the cost of each potential path, and selects the path with the minimum total cost, outputting a series of ordered path points. This ensures that the planned path is not only the shortest in distance but also the fastest or least energy-intensive, and can avoid congestion in real time, achieving dynamic, adaptive, and refined management.

[0055] Step S107: Encapsulate the loading point location, unloading point location, and driving path point sequence into a scheduling instruction and send it to the corresponding truck's on-board terminal.

[0056] The system encapsulates the structured decision results (loading point, unloading point, and path sequence) into a complete, machine-readable dispatch instruction according to a predefined communication protocol. This instruction is then transmitted to the designated truck's onboard terminal via a wireless network (such as 4G / 5G). Upon receiving the instruction, the terminal guides the driver through graphical means (such as displaying the route on an onboard map) and / or voice commands. This ensures that the dispatch intent is accurately, error-free, and timely communicated to the executor, avoiding potential mishearing or misunderstanding issues associated with traditional walkie-talkie dispatching, and significantly improving the accuracy and efficiency of instruction execution.

[0057] Step S108: Continuously acquire the actual driving trajectory data and operation status of each truck. When the trajectory deviation exceeds the threshold or the operation is abnormal, trigger the reassignment of loading / unloading points and update the driving path point sequence.

[0058] The system continuously monitors the performance of each truck: by comparing the actual driving trajectory with the planned route, it calculates the trajectory deviation (such as vertical distance). Once the deviation exceeds a threshold (e.g., the driver detours due to temporary road obstacles), it is considered a planned derailment. Simultaneously, it monitors the operational status (e.g., fault signals reported by onboard sensors). When these abnormal events occur, the system treats them as new triggering conditions, automatically initiating a rescheduling process—that is, re-allocating tasks and replanning routes, and issuing new instructions. This forms a continuous optimization closed loop of "perception-decision-execution-monitoring-re-decision," enabling the entire scheduling system to dynamically respond to various uncertainties at the mine operation site and always maintain the effectiveness and optimality of the scheduling plan.

[0059] In the above implementation, the traditional static scheduling method, which relies on manual experience, is upgraded to a data-driven, model-based, and dynamically optimized closed-loop adaptive system. This method achieves optimal matching of resources and tasks from a global perspective through a mixed-integer programming model, improving the efficiency of vehicle-shovel coordination and equipment utilization. Dynamic path planning combined with real-time road conditions effectively reduces empty mileage, waiting time, and congestion, directly lowering operating costs and energy consumption. The introduced closed-loop monitoring and rescheduling mechanism greatly enhances the system's robustness in responding to unexpected situations on-site, ensuring the continuity and stability of production operations. Ultimately, this technical solution maximizes the efficiency of mine transportation operations, streamlines operating costs, and achieves agility and high reliability in the production process.

[0060] Reference Figure 2 As one implementation of step S104, the step of acquiring equipment maintenance plan data and constructing a mixed-integer programming model that includes load constraints, time constraints, and task coverage constraints, in conjunction with real-time status data, road topology, and production task instruction data, includes:

[0061] Step S201: Read the pre-stored equipment maintenance plan data and extract the next maintenance time and maintenance duration for each truck;

[0062] The underlying logic of this step is to transform long-term equipment management planning into hard constraints that the scheduling system can utilize immediately. Equipment maintenance planning is a core component of mine asset management; it pre-plans the maintenance schedule for each truck based on its operating hours, mileage, or fixed cycles.

[0063] Specifically, the system reads this structured data from the maintenance management database and extracts two key parameters: the next maintenance time (an absolute timestamp that marks the deadline by which the truck must begin maintenance) and the maintenance duration (the estimated maintenance operation time).

[0064] Understandably, by shifting preventative maintenance strategies from a passive, reactive approach to an active scheduling decision factor, the scheduling algorithm ensures that a truck nearing the end of its maintenance period will not be assigned to perform a long-cycle task that may not be completed on time during critical moments (such as during periods of high-load production). This avoids production interruptions or mid-term changes to scheduling instructions due to unplanned downtime, thereby improving the feasibility and robustness of the scheduling plan from the source.

[0065] Step S202: Based on the location data and vehicle speed data, calculate the estimated time for each truck to reach each loading point;

[0066] The system utilizes real-time collected location data (the truck's current location) and speed data (the truck's instantaneous or average speed), combined with existing road topology, to calculate the theoretical shortest path distance from each truck's current location to each potential loading point using path planning algorithms (such as the shortest path algorithm).

[0067] Subsequently, based on the current vehicle speed or the historical average speed of that road segment, the spatial distance is converted into a time estimate, i.e., the estimated travel time. This calculation is not static but dynamically updated. It reflects the real-time accessibility of trucks in the complex terrain of the mine. For example, a truck returning empty and closer to the loading point will naturally have a shorter estimated travel time than a truck that has just unloaded and is located at the other end of the mine. This accurate time estimate is a key input for the next step of determining whether a truck can complete its task before maintenance, and it is also the foundation for achieving efficient task matching.

[0068] Step S203: Generate truck availability tags based on the next maintenance time and estimated time.

[0069] If a truck can return to the maintenance point as planned after completing its current task, it is marked as available; otherwise, it is marked as disabled.

[0070] Specifically, the system logically compares rigid time constraints (maintenance time points) and dynamic time predictions (estimated time consumption). The core criterion is: assuming the current time point is T0, the estimated total time for the truck to complete a new task (including traveling to the loading point, loading, transporting to the unloading point, and unloading) is T_task, and the estimated time to return from the unloading point to the maintenance point is T_return. If T0 + T_task + T_return < the next maintenance time point, it indicates that the truck still has sufficient time to return for planned maintenance after completing this task, and therefore it is marked as "available." Otherwise, it is marked as "disabled."

[0071] This availability label is a highly simplified and abstract signal that transforms complex spatiotemporal judgments into a Boolean value (yes / no). This mechanism cleverly embeds equipment maintenance, a logistical support activity, seamlessly into real-time production scheduling logic, ensuring that production task allocation does not conflict with equipment health management plans and achieving sustainable resource utilization.

[0072] Step S204: Generate a task requirement matrix based on the location of the loading point, the location of the unloading point, and the amount of ore required;

[0073] The production task instructions received by the system (loading / unloading point locations, ore demand) are the concrete manifestation of the scheduling objectives. The task demand matrix is ​​a mathematical abstraction, typically a two-dimensional or multi-dimensional array. Its rows may represent different loading points (bursting piles), and its columns may represent different unloading points (such as crushing stations and spoil heaps). The element values ​​in the matrix represent the ore demand from a specific loading point to a specific unloading point. This matrix quantifies the total amount and flow of all transportation operations that the entire mine needs to complete in the next scheduling cycle.

[0074] Step S205: Construct a load constraint matrix based on load data and truck availability labels;

[0075] The load constraint matrix allows only available state trucks whose load capacity meets the current task's ore requirements to participate in the allocation.

[0076] Specifically, the load constraint matrix is ​​a set of constraints that specifies which truck is eligible to undertake which task. It is based on two key factors: first, the truck's physical capacity, i.e., load data, ensuring that the truck's rated load capacity is greater than or equal to the amount of ore required for a single transport task, avoiding overloading or wasted capacity; second, the truck's operational status, i.e., availability tags, ensuring that only trucks marked as "available" are included in the candidate pool for the current scheduling cycle.

[0077] Understandably, the load constraint matrix essentially defines a feasible decision space for subsequent optimization algorithms. It eliminates all unreasonable allocation schemes (such as using a small truck to haul a heavy-duty task, or assigning a truck that is about to be maintained), greatly narrowing the search range of the algorithm, improving the solution efficiency, and ensuring the physical and planning feasibility of the generated scheme. This is the core step in transforming business rules into mathematical constraints.

[0078] Step S206: Construct a timeliness constraint vector based on the road topology and real-time status data;

[0079] Among them, the time constraint vector specifies the latest completion time of each task, injecting time urgency requirements into production tasks and realizing the upgrade of scheduling optimization from simple spatial path optimization to spatiotemporal joint optimization.

[0080] Specifically, the road topology provides basic information about the path network, while real-time status data (such as average vehicle speed and road congestion) reflects the current traffic efficiency of the road network. By combining these two factors, the system can predict the travel time required to complete a transportation task between any "loading point-unloading point" pair. The timeliness constraint vector specifies a latest completion time threshold (T_max) for each such transportation task pair. This threshold may be set based on factors such as the urgency of the production plan and the material supply demand at the unloading point. This constraint ensures that the optimization algorithm considers not only "whether it can transport" but also "whether it can arrive on time" when allocating tasks. It prevents the algorithm from assigning an urgent task to a truck that, while ultimately saving total time, cannot meet the specific timeliness requirements of the task in pursuit of the shortest overall time, thus ensuring the smooth operation of critical production processes.

[0081] Step S207: Based on the task requirement matrix and the time constraint vector, construct the task coverage constraint;

[0082] Specifically, the task coverage constraint is the top-level constraint in the model. It mandates that the final solution generated by the optimization algorithm must satisfy all ore transportation needs defined in the task requirement matrix and be completed within the time frame specified by the time constraint vector. This means that the algorithm cannot "cut corners" simply to minimize the total transportation time, such as ignoring the demand at a certain unloading point or allowing some tasks to complete overtime. This constraint transforms the rigid requirements of the production plan into mathematical equality or inequality constraints, ensuring that the model's solution (i.e., the scheduling scheme) must be a feasible and complete solution that can effectively guide production, ensuring that each crushing station obtains sufficient ore and each waste dump receives waste rock as planned, thereby guaranteeing the stable operation of the entire mining production chain.

[0083] Step S208: Using minimizing the total transportation time as the objective function, integrate the load constraint matrix, timeliness constraint vector, and task coverage constraint to generate a mixed integer programming model.

[0084] Mixed Integer Programming (MIP) is a powerful mathematical tool for solving complex optimization problems involving discrete decisions (such as whether to assign task A to truck B, which is a 0-1 decision) and continuous variables (such as transportation time).

[0085] In this embodiment, the objective function "minimize total transportation time" indicates the direction of optimization, namely, maximizing overall efficiency. The previously constructed load constraint matrix, timeliness constraint vector, and task coverage constraint serve as constraints on the model, collectively defining what constitutes a "legal" solution. This process is analogous to constructing a building: the objective function is the building's design goal (e.g., maximum stability), the various constraints are building codes (e.g., load-bearing capacity, fire resistance requirements), and the MIP model is the final, complete design blueprint. After generating this model, it can be computed by a professional mathematical optimization solver, ultimately outputting a theoretically optimal or near-optimal truck scheduling scheme that satisfies all real-world business constraints.

[0086] In the above implementation, mutually independent or even conflicting management elements in traditional scheduling (such as production tasks, equipment maintenance, real-time road conditions, and vehicle performance) are deeply integrated and quantitatively modeled. Through the powerful mathematical tool of mixed integer programming, an experience-based scheduling decision-making process is transformed into an automated optimization process based on data and algorithms. This technical solution not only ensures the efficient completion of production tasks (minimizing total transportation time) but also significantly improves the reliability and executability of the scheduling plan by introducing equipment maintenance constraints and timeliness constraints. It effectively avoids unplanned interruptions and production bottlenecks, thereby achieving a synergistic improvement in overall mine transportation efficiency, equipment utilization, and production stability, providing core technical support for the refined operation of smart mines.

[0087] Reference Figure 3 As one implementation of step S105, the step of solving the mixed integer programming model to generate the loading / unloading point allocation scheme for each truck includes:

[0088] Step S301: Invoke the mixed integer programming model and load the load constraint matrix, time constraint vector, and task coverage constraint;

[0089] The mixed-integer programming model can be understood as a mathematical optimization framework with a defined basic structure. This framework includes the objective function (such as minimizing the total transportation time) and the constraint variables to be filled. The loading process involves formally injecting the specific and quantified business rules from the previous steps—namely, the load constraint matrix (which specifies which type of truck can undertake which type of task), the time constraint vector (which specifies when each task must be completed), and the task coverage constraint (which specifies that all tasks must be completed)—into this mathematical framework. Essentially, this step perfectly translates the complex scheduling rules in reality ("large trucks transport large tonnage," "urgent tasks are completed first," "all orders must be cleared") into mathematical language that computers can directly understand and process. It defines clear "rules of the game" for the solver (optimization algorithm): any solution must be found within the scope defined by these rules.

[0090] Step S302: Use the branch and bound algorithm to traverse the solution space and generate an initial feasible solution that satisfies all constraints.

[0091] Specifically, the execution of the branch and bound algorithm includes: defining the combination of trucks and loading / unloading points as decision tree nodes; generating local solutions that satisfy the load constraints through depth-first search; pruning branches and backtracking to the previous node when a local solution violates the task coverage constraint.

[0092] Branch and bound is a classic, exact algorithm for solving combinatorial optimization problems. Faced with a vast solution space (i.e., all possible allocations) of hundreds of tasks assigned to dozens of trucks, this algorithm employs a systematic "divide and conquer" search strategy. It decomposes the problem into a decision tree, where each node represents a partial decision (e.g., "assign task A to truck 1"). The algorithm traverses this tree using methods such as depth-first search. For each decision (branch), it checks whether the current partial solution violates any constraints (e.g., overloading). If a violation occurs, the branch and all its subsequent possibilities are abandoned (bound or pruned), and the algorithm backtracks to the previous node to try other options. This method intelligently avoids most invalid regions, ultimately finding at least one initial feasible solution that satisfies all constraints. This solution may not be optimal, but it is a valuable starting point, proving that a feasible scheduling scheme exists under the current constraints, laying a solid foundation for subsequent local optimization.

[0093] Step S303: Calculate the load utilization rate score of the initial feasible solution based on the matching degree between the truck's load capacity and the task's ore demand.

[0094] The load utilization rate score measures the efficiency of using truck capacity, a valuable resource. Its core idea is to calculate the ratio of the assigned workload to the rated load capacity of each truck, and then average this ratio across all participating trucks. A high score indicates that the overall load capacity of the fleet is fully utilized, avoiding the waste of capacity through overloading; a low score indicates a significant resource mismatch in the allocation scheme, requiring optimization. This scoring mechanism transforms the vague goal of "improving efficiency" into a clear and objective mathematical objective, upgrading the optimization process from "finding a feasible solution" to "finding a high-quality solution."

[0095] Step S304: If the load utilization rate score is lower than the preset score threshold, a neighborhood search is triggered. Under the premise of maintaining the task coverage constraint, the loading / unloading point allocation combination of the two trucks is swapped.

[0096] The triggering conditions for neighborhood search include: when the load utilization rate score S < 0.85; selecting the two trucks with the lowest scores as the exchange targets; and after the exchange, Δt ≤ Tmax, where Δt is the change in total transportation time before and after the exchange, and Tmax is the maximum allowable fluctuation threshold.

[0097] Specifically, when the initial solution receives a poor score, it indicates that the algorithm has failed to fully optimize resource utilization while strictly adhering to the constraints. In this case, the system automatically triggers a neighborhood search (a powerful heuristic optimization technique). The "neighborhood" refers to the set of all new solutions that can be obtained by making a small, structured change to the current solution. Here, the neighborhood operation is defined as "swapping the task assignments of two trucks." First, by limiting the swap to only two trucks, it ensures that the change is local and controllable, and will not completely overturn the original solution. Second, the swap operation must maintain the task coverage constraint; that is, all tasks are still completed, only the trucks performing the tasks have changed.

[0098] Step S305: Re-verify the time constraints of the swapped allocation combinations and generate an optimized feasible solution;

[0099] Task swapping may alter transportation routes and time consumption, potentially triggering or even exceeding time constraints (i.e., the latest task completion time). Therefore, the new allocation combinations resulting from the swaps must be rigorously re-verified. The system recalculates the estimated completion time for each truck under the new task based on road topology and vehicle speed, comparing it with the time thresholds specified in the time constraint vector. Only swap results that fully satisfy all time constraints are accepted as a new, optimized, and feasible solution. This step prevents short-sighted behavior that sacrifices task timeliness for the sake of unilaterally pursuing load utilization, ensuring the robustness and practical operability of the optimization scheme.

[0100] Step S306: Output the feasible solution with the highest load utilization rate as the loading / unloading point allocation scheme.

[0101] After finding an initial solution through branch and bound, and potentially obtaining a series of optimized feasible solutions through multiple neighborhood searches, the system compares the load utilization scores of all these solutions. Ultimately, it determines the solution with the highest score as the formal loading / unloading point allocation scheme, thereby ensuring that the final dispatch instructions issued to the fleet achieve the optimal balance between transportation efficiency and resource utilization under all known information and constraints.

[0102] In the above implementation, a two-layer optimization mechanism combining "global optimization" and "local fine-tuning" was designed to improve the intelligence and refinement of the mine truck scheduling scheme. This method ensures the basic feasibility of the solution through a branch-and-bound algorithm, and then innovatively introduces a load utilization rate scoring system to quantitatively diagnose the quality of the scheme. Based on this, targeted neighborhood search optimization is triggered, ultimately outputting a near-optimal scheduling scheme in terms of capacity utilization while ensuring that all operational constraints are strictly adhered to. By closely integrating mathematical optimization theory with actual business needs (improving truck load factor and reducing unit transportation costs), this method provides direct and crucial technical support for cost reduction and efficiency improvement in mines at the practical operational level.

[0103] Reference Figure 4 As one implementation of step S106, the step of planning a sequence of driving route points for each truck based on the loading / unloading point allocation scheme and real-time road traffic flow data includes:

[0104] Step S401: Based on the loading / unloading point allocation scheme, extract the current location data, loading point location, and unloading point location of each truck;

[0105] The system extracts three key geographic coordinates from the allocation scheme: current location data (real-time GPS coordinates of the truck, representing the dynamic starting point of the route), loading point location data (intermediate target points of this transportation task), and unloading point location data (the final target point of this transportation task). This set of data provides the most fundamental input parameters for subsequent route planning algorithms, transforming the abstract task into a specific sequence of spatial coordinates, ensuring the individualization and accuracy of scheduling instructions.

[0106] Step S402: Obtain real-time road traffic flow data, including current road traffic density, average vehicle speed, and road congestion status indicators.

[0107] The traffic efficiency of mine roads is dynamic; for example, a narrow road may experience slow traffic due to multiple vehicles converging. By collecting real-time road traffic data, including traffic density (number of vehicles per unit length of road), average speed, and road congestion indicators, the algorithm reflects the real-time travel cost of the road network. The path planning algorithm no longer considers only fixed distances but also dynamic travel times. For example, the algorithm will proactively avoid road sections with excessively high traffic density and low average speeds, even if they are geometrically shorter. This step tightly couples real-time perception with decision-making, greatly improving the practicality and efficiency of path planning.

[0108] Step S403: Based on the path planning algorithm, an initial driving path point sequence is generated using the current location data, loading point location, unloading point location, road topology, and real-time road traffic flow data as inputs.

[0109] In some embodiments, the path planning algorithm is a dynamic programming algorithm, and the execution rules include: taking the current position of the truck as the starting point and the unloading point as the ending point; calculating the travel time of each road segment based on the road topology; updating the travel time weights in combination with real-time traffic flow data; and generating an initial travel path point sequence with the minimum total travel time.

[0110] Specifically, by integrating all spatial and dynamic information, the algorithm automatically calculates an optimal or near-optimal feasible path. The path planning algorithm acts as the "navigation brain." It takes all data as input and searches the network graph constructed from the road topology. The algorithm calculates a comprehensive "cost" for each road segment, which is typically determined by the base distance and real-time traffic flow (e.g., cost = distance / average speed, representing travel time). The algorithm's goal is to find a continuous path from the starting point (current location), through intermediate points (loading points), to the ending point (unloading point), minimizing the total cost of all road segments. The final output, the initial travel path point sequence, is an ordered list of latitude and longitude coordinates that precisely indicates which points the truck should pass through in order to efficiently complete the task.

[0111] Step S404: Verify whether the initial driving path point sequence meets the preset safe area constraints. The safe area constraints include the latitude and longitude range of signal blind spots and dangerous areas.

[0112] Specifically, the preset safety zone constraints include: loading signal blind zone data from system storage, which is generated based on historical vehicle reporting frequency analysis; and loading hazardous zone data, which is statically defined based on the slope and obstacle coordinates of the mine's 3D map.

[0113] Among these, the safety zone constraint refers to the restricted areas pre-defined electronically in the system. Signal blind spots refer to areas not covered by the communication network; vehicles entering these areas lose contact with the dispatch center, posing a significant safety risk. Dangerous areas may include steep slopes, geologically unstable areas, and blasting warning zones. This step performs a safety audit on the initial path recommended by the algorithm. The system compares the sequence of path points with the latitude and longitude ranges of these restricted areas to verify whether any path points fall into or cross these regions.

[0114] If not, proceed to step S405; if yes, proceed to step S406.

[0115] Step S405: Adjust the waypoint sequence to avoid signal blind spots or dangerous areas, and output the adjusted driving waypoint sequence as the final driving route for each truck.

[0116] Step S406: Directly output the initial driving path point sequence.

[0117] The underlying logic of this step is to implement a proactive security protection and correction mechanism to repair paths that do not meet security requirements. When verification reveals a security risk in the initial path, the system does not simply reject the path but initiates an automatic adjustment process, which is essentially a local replanning.

[0118] Specifically, the system uses the boundary of the safe zone as a avoidance point and employs a path planning algorithm (such as the A* algorithm) to quickly calculate an alternative path that bypasses the danger zone or blind spot. This adjustment ensures that the final output path maintains the efficiency of the initial path as much as possible while absolutely meeting safety requirements. It demonstrates the system's intelligence and robustness, automatically identifying and correcting potential risks in the solution, thus achieving closed-loop management of safety.

[0119] In the above implementation, route planning is upgraded from a single static shortest path calculation to a multi-objective decision-making process that comprehensively considers real-time traffic efficiency and geospatial safety constraints. This method first uses real-time traffic flow data to ensure the time optimality of the route, and then introduces an independent safety zone verification and adjustment process to provide reliable safety guarantees for efficient routes, ultimately generating a driving route that achieves the best balance between "speed" and "safety." This mechanism enables the truck dispatching system to not only focus on production efficiency but also to proactively integrate safety management into the algorithm's decision-making process, thereby achieving a synergistic improvement in both efficiency and safety in mining transportation operations and effectively reducing operational risks caused by improper route selection.

[0120] Reference Figure 5 As one implementation of step S108, the steps of continuously acquiring the actual driving trajectory data and operation status of each truck, and triggering the reallocation of loading / unloading points and updating the driving path point sequence when a trajectory deviation exceeds a threshold or an abnormal operation status is detected, include:

[0121] Step S501: Receive the truck positioning data sequence uploaded by the vehicle terminal in real time and generate actual driving trajectory data;

[0122] The system continuously receives positioning data sequences (typically latitude and longitude coordinates and timestamps provided by GPS or BeiDou systems) from the onboard terminals of each truck via a wireless communication network. These discrete positioning points, arranged in chronological order, constitute the actual driving trajectory point set of the truck. This process enables precise, real-time digital tracking of the truck's motion state, transforming the truck's physical movement trajectory into time-space sequence data that can be processed and analyzed by a computer. This trajectory point set is the objective basis for the system to perceive whether the vehicle is executing dispatch instructions as planned, and it is the starting point of the entire closed-loop feedback control.

[0123] Step S502: Load the current truck's driving path point sequence and calculate the Euclidean distance deviation between the actual driving trajectory data and the driving path point sequence;

[0124] The system retrieves the sequence of planned driving route points generated for the truck during the preceding path planning phase; this is a pre-defined ideal driving route. By calculating the Euclidean distance deviation (typically the straight-line distance between the actual and planned positions at the same point in time or the nearest path point), the system can quantify the degree of deviation of the actual driving trajectory from the ideal path. This calculation essentially simplifies the complex path-following problem into a measurable spatial deviation problem. Statistical measures such as the maximum deviation (Dmax) can intuitively reflect whether the vehicle has significantly deviated from the predetermined route due to detours, getting lost, traffic congestion, or other unexpected situations. This deviation value is key to transforming the subjective feeling of "deviation" into an objective, quantifiable indicator that can trigger subsequent actions.

[0125] Step S503: Simultaneously receive the operation status data reported by the vehicle-mounted sensors. The operation status data includes fault codes, load change indicators, and abnormal vehicle speed indicators.

[0126] This involves establishing a multi-dimensional vehicle health status monitoring system to identify potential anomalies at the level of internal vehicle operating parameters. Besides external driving trajectories, the vehicle's own operating status is also crucial. The system receives data from the onboard sensor network in parallel, providing in-depth insights into the vehicle's internal operating conditions. Fault codes are standardized abnormal signals issued by the vehicle control system after self-diagnosis, directly indicating specific mechanical or electrical faults. Load change indicators (such as drastic load fluctuations within a very short period) may indicate abnormal loading / unloading, cargo spillage, or weighing sensor malfunction. Abnormal speed indicators (such as prolonged driving far below the speed limit) may suggest problems with the vehicle's powertrain or the presence of unseen obstacles on the road. These operational status data collectively constitute a comprehensive criterion beyond location monitoring for determining whether the vehicle is in normal working condition.

[0127] Step S504: When the Euclidean distance deviation value exceeds the preset path offset threshold, or when the operation status data contains a predefined abnormal flag, mark the truck as an abnormal vehicle.

[0128] This involves implementing a rule-based, multi-condition joint decision-making mechanism to intelligently and automatically identify abnormal vehicles requiring system intervention. The system compares the quantified results of the aforementioned steps with preset safety thresholds and an anomaly flag database. If the vehicle has merely deviated from the path but is unharmed, only minor path adjustments may be needed; however, if the vehicle itself is malfunctioning, the task cannot be completed even if the path is correct.

[0129] Therefore, when the Euclidean distance deviation exceeds a threshold (e.g., 50 meters), it indicates abnormal spatial behavior; or when predefined abnormal flags appear in the operational status data (e.g., engine fault codes, abnormal load changes), it indicates that the truck itself is in an abnormal state. The system will mark the truck as an abnormal vehicle if either condition is met.

[0130] Step S505: Based on the location data of all vehicles in abnormal states and the remaining task load, reconstruct the task coverage constraints and timeliness constraints.

[0131] When one or more trucks are marked as abnormal, their original tasks may become incomplete or severely delayed. The system needs to take these changes into account. It first assesses the remaining workload of the abnormal vehicles (e.g., the amount of ore loaded but not yet delivered by a malfunctioning truck), and then re-adds these unfinished tasks to the task pool.

[0132] Based on this, the system reconstructs the task coverage constraint matrix, specifically including: extracting the remaining ore demand of abnormal vehicles; merging the remaining ore demand into the unassigned task queue; updating the total number of tasks to be assigned at all loading points and the demand gap at unloading points, ensuring that the new scheduling scheme can still meet the total demand of all unloading points. Simultaneously, since time has passed and delays have occurred, the timeliness constraint vector also needs to be updated according to the current time and task urgency. This step involves online and dynamic adjustment of the input parameters of the previously established mixed-integer programming model, enabling the new solution to be based on the latest and most realistic on-site conditions.

[0133] Step S506: Call the mixed integer programming model, and use the reconstructed task coverage constraints and time constraints as inputs to re-solve and generate a new loading / unloading point allocation scheme;

[0134] The mathematical optimization solver is used for calculations, and since the model framework remains unchanged and only the parameters are updated, the solution efficiency is usually high. This means that the system can quickly calculate a new, globally optimal loading / unloading point allocation scheme that takes into account all current abnormal vehicles and remaining tasks within a few minutes.

[0135] Step S507: Update the truck's route point sequence based on the new loading / unloading point allocation scheme and real-time road traffic flow data.

[0136] The system re-plans routes for all trucks involved in task changes (which may include normal vehicles newly assigned tasks and abnormal vehicles that need to complete remaining tasks, if they are recovered), combining the latest real-time road traffic data to generate an updated sequence of driving route points. This new route sequence is immediately sent to the corresponding on-board terminal to guide the driver to perform the new task.

[0137] In the above implementation, when some vehicles in the system are unable to operate as planned due to deviation from their routes or their own malfunctions, the system can automatically and quickly quantify these abnormal situations into manageable mathematical constraint changes, and dynamically adjust the global scheduling plan accordingly, rather than performing localized and short-sighted emergency handling. This method transforms the passive and lagging manual intervention in traditional scheduling systems into proactive and real-time system-level self-healing capabilities, greatly enhancing the robustness of the scheduling system in dealing with various uncertainties on-site, ensuring the continuity and overall efficiency of mine production operations in emergency situations, and achieving a qualitative leap from "static optimization scheduling" to "dynamic fault-tolerant scheduling".

[0138] Reference Figure 6 As a further implementation of the intelligent scheduling method for mining trucks, after the step of planning a sequence of driving route points for each truck based on the loading / unloading point allocation scheme and real-time road traffic flow data, the method further includes:

[0139] Step S601: When the same transportation task can be assigned to multiple trucks in the loading / unloading point allocation scheme, obtain the load-fuel consumption relationship curve data of each pre-configured truck.

[0140] Understandably, in actual scheduling, there are often situations where multiple trucks meet the same task requirements in both time and space (i.e., the task is replaceable). In this case, the system does not randomly assign or only consider the closest distance, but instead activates an energy consumption optimization mechanism.

[0141] In this step, each truck exhibits a specific nonlinear relationship between its load and fuel consumption due to inherent characteristics such as engine model, vehicle condition, and transmission efficiency. The load-fuel consumption curve is a unique characteristic curve obtained through rigorous bench tests and field road tests, collecting a large amount of fuel consumption data under different load gradients (such as no load, half load, and full load), and fitting it using a mathematical model (such as a quadratic function). This curve essentially reveals the objective law of how fuel consumption per unit distance traveled varies with load under standard road conditions.

[0142] Step S602: Calculate the real-time load factor of each truck based on the ore transport volume in the loading / unloading point allocation plan and the truck load data.

[0143] The loading / unloading point allocation scheme determines the amount of ore each truck needs to transport, which is a task requirement. The truck's load capacity (rated load) is its physical capacity limit. The real-time load factor is calculated by dividing the task volume by the rated load to obtain a percentage value, which accurately reflects the degree to which the truck's transport capacity will be utilized under the currently assigned task.

[0144] For example, a truck rated for 50 tons is assigned a 30-ton load, resulting in a load factor of 60%. This calculation transforms the discrete, absolute workload into a continuous, relative load intensity, and its core function is to provide a unified input interface for subsequent fuel consumption prediction. Because different trucks have different rated loads, directly comparing their absolute loads is meaningless. However, through the load factor, a normalized parameter, the system can evaluate and compare the energy efficiency of different vehicles performing the same task on a fair benchmark.

[0145] Step S603: Based on the load-fuel consumption relationship curve data and real-time load rate, predict the fuel consumption per unit mileage for each truck.

[0146] The system uses the calculated real-time load rate as an input variable, substituting it into the truck's specific load-fuel consumption curve for interpolation. The output is the fuel consumption per unit mile, such as "liters / km". This predicted value represents the expected fuel consumption per kilometer for the truck traveling at a constant speed on an ideal, straight road under the current task load. It combines a static curve describing the vehicle's inherent characteristics with a dynamic load state determined by the specific scheduling task, generating a specific, quantified energy consumption expectation for the current task. This provides the most basic granular data unit for comparing energy consumption across different paths.

[0147] Step S604: Load the mine road topology and obtain mine road slope data;

[0148] The roads in the mine are not flat; continuous ups and downs significantly affect the actual fuel consumption of vehicles. The road topology not only includes the horizontal connections of the roads but also embeds the elevation information of each node. By calculating the ratio of the elevation difference to the horizontal distance between adjacent path points, the system can accurately obtain the slope data of each road segment (usually expressed as a percentage, such as an 8% slope). Slope data is a core environmental parameter for correcting the basic fuel consumption prediction because going uphill requires the engine to do extra work to overcome gravity, leading to a significant increase in fuel consumption; while going downhill may reduce power demand or even utilize braking energy.

[0149] Step S605: Calculate the total predicted fuel consumption for each route based on the fuel consumption per unit mileage, road gradient data, and road distance of the driving route point sequence.

[0150] The system integrates and calculates the various data obtained. This process typically involves dividing the entire route into segments based on gradient changes. For each segment, the base fuel consumption is adjusted according to its gradient (e.g., multiplying uphill sections by a correction factor greater than 1), and then multiplied by the segment's length to obtain the predicted fuel consumption for that segment. Finally, the fuel consumption of all segments is summed to obtain the total predicted fuel consumption. This total accurately quantifies the expected fuel consumption for the entire transportation task. Through this step, the system can precisely assess the energy consumption differences between different route options, thus providing indisputable data support for optimization decisions based on the lowest energy consumption.

[0151] Step S606: Select the truck with the lowest total predicted fuel consumption as the final execution vehicle and update the loading / unloading point allocation scheme.

[0152] The system compares the total predicted fuel consumption of all candidate trucks for the task and ultimately selects the truck with the lowest fuel consumption. After selection, the system updates the loading / unloading point allocation scheme accordingly. This step, while meeting all production constraints (such as task completion time and unloading point requirements), elevates the "lowest total energy consumption" economic and environmental indicator to a key basis for the final decision. This makes the scheduling system not only intelligent but also capable of proactively selecting the most energy-efficient equipment and route combinations at the micro-level, thereby achieving the goals of reducing overall operating costs and carbon emissions.

[0153] In the above implementation, by utilizing the unique energy consumption curve of each truck and the precise slope information of the mine roads, it is possible to make forward-looking and quantitative predictions and comparisons of total fuel consumption under different task allocation schemes. This allows for the proactive identification and selection of the most energy-efficient scheme among multiple feasible scheduling options. This breaks through the limitations of traditional scheduling that only focuses on distance or time, achieving an optimization upgrade from "shortest path" to "lowest energy consumption." Thus, while ensuring production efficiency, it significantly reduces fuel consumption and operating costs, and reduces carbon emissions, providing crucial technical support for the green and sustainable development of mines.

[0154] Reference Figure 7 As a further implementation of the intelligent scheduling method for mining trucks, the load-fuel consumption relationship curve data is generated through the following steps:

[0155] Step S701: Obtain the rated load parameter value for each truck;

[0156] Among them, the rated load parameter is a core performance indicator determined by the manufacturer during the design and manufacturing stage of the vehicle based on factors such as its structural strength, material mechanical properties, braking system capabilities, and safety regulations. It defines the maximum effective load that the vehicle can withstand while ensuring safety and durability.

[0157] Step S702: Based on the rated load parameter value, divide the load interval sequence according to the preset gradient and generate the interval boundary value set;

[0158] While the impact of load on fuel consumption is continuous, within a relatively small load variation range (such as from 30% to 40% of the rated load), the relationship can be approximated as linear. This ensures sufficient accuracy for data analysis and fitting within each range, while avoiding the problem of insufficient data and decreased statistical significance caused by overly detailed divisions.

[0159] For example, dividing the load into 10% increments means that the nonlinear characteristics of truck engine fuel efficiency changing with load are approximately linear within each 10% load interval, thus effectively balancing data accuracy and computational complexity. By generating a set of interval boundary values ​​(such as 0%, 10%, 20%, ..., 100% of rated load), this step essentially constructs a systematic test framework covering the entire load range, ensuring that data collection can uniformly cover all possible working load states of the truck, laying a structural foundation for establishing a comprehensive and accurate energy consumption model.

[0160] Step S703: For each load range, extract several sets of load value-fuel consumption value data pairs from the historical database during constant speed driving.

[0161] Among them, constant speed driving is a key data screening condition. The purpose is to control variables and eliminate the extra fuel consumption interference caused by changes in kinetic energy during acceleration or deceleration, so as to purely capture the essential relationship between load and steady-state fuel consumption.

[0162] In some embodiments, at least 30 data sets may be specified for each interval. When the sample size reaches 30 or more, the distribution of the sample mean is closer to a normal distribution, which makes the parameter estimation (such as subsequent curve fitting) based on these data more stable and reliable. These load-fuel consumption data pairs are "chemical samples" extracted from a large amount of historical operating data, reflecting the actual energy consumption level of the vehicle under specific loads, and are the raw materials for model calibration.

[0163] Step S704: Divide the load value by the rated load parameter value to convert it into a set of load rate values;

[0164] This process involves dividing the specific, dimensional load value (in tons) by the vehicle's corresponding rated load parameter (in tons) to obtain a percentage value, i.e., the load factor. This conversion ensures that a truck with a rated load capacity of 60 tons carrying 45 tons has the same load factor (75%) as a truck with a rated load capacity of 100 tons carrying 75 tons. This means that although their absolute loads differ, their relative load levels are the same.

[0165] Step S705: Use the least squares method to perform quadratic function fitting on the load factor value set and the corresponding fuel consumption value set, and output the fitting function expression: F(L)=aL 2 +bL+c; where L is the load factor, and a, b, c are the fitting coefficients.

[0166] Specifically, the core idea of ​​the least squares method is to find a curve that minimizes the sum of the squares of the vertical distances (i.e., residuals) from all actually observed data points to this curve. The choice of this quadratic function formula for fitting is based on the physical understanding of the operating characteristics of truck engines and transmission systems: fuel consumption typically increases non-linearly with increasing load because, in addition to the basic resistance proportional to the load (reflected by the linear term bL), there are factors such as tire rolling resistance and air resistance that increase quadratically with speed (or load) (reflected by the quadratic term aL²), as well as the basic energy consumption when the vehicle is running unloaded (the constant term c).

[0167] Step S706: Store the fitting coefficient values ​​a, b, c in the fitting function expression as calibration parameters for the load-fuel consumption relationship curve data, and associate them with the corresponding truck.

[0168] Storing the fitted coefficients a, b, and c as calibration parameters is a highly efficient data compression and abstraction process. Curves that originally required hundreds of data points to describe can now be accurately represented in the system using only these three coefficients.

[0169] Furthermore, by associating these coefficients with the unique identifiers of the corresponding trucks (such as vehicle IDs), an energy consumption profile is created for each vehicle in the system database, which can be queried at any time. When the scheduling algorithm needs to predict the fuel consumption of a vehicle performing a task, it does not need to access the original massive amounts of data. It only needs to retrieve the three coefficients a, b, and c based on the vehicle ID and substitute them into the formula F(L) to instantly calculate the result. This greatly improves the system's operating efficiency, enabling complex energy consumption optimization calculations to meet the requirements of real-time scheduling.

[0170] In the above implementation, a high-precision personalized energy consumption prediction model is established for each mining truck, which improves the previous fuel consumption estimation that relied on driver experience or rough average values ​​into a quantifiable and predictable scientific tool based on historical big data analysis and mathematical modeling. This enables the scheduling system to quantify the actual fuel consumption cost under different loads and road conditions in route planning.

[0171] This application also discloses an intelligent dispatching system for mining trucks.

[0172] A mining truck intelligent dispatching system, specifically comprising:

[0173] The status data acquisition module is used to collect real-time status data of all transport trucks in the mining operation area, including location data, vehicle speed data, and load data.

[0174] The road topology generation module is used to obtain the latitude and longitude coordinate sequence of each node in the mine road network and generate the road topology.

[0175] The production task instruction receiving module is used to receive production task instruction data, including the location of the loading point, the location of the unloading point, and the required amount of ore.

[0176] The scheduling optimization model construction module is used to acquire equipment maintenance plan data, combine real-time status data, road topology and production task instruction data, and construct a mixed integer programming model that includes load constraints, time constraints and task coverage constraints.

[0177] The task allocation scheme solving module is used to solve the mixed integer programming model and generate the loading / unloading point allocation scheme for each truck.

[0178] The route planning module is used to plan the sequence of driving route points for each truck based on the loading / unloading point allocation scheme and real-time road traffic flow data;

[0179] The dispatch instruction issuing module is used to encapsulate the loading point location, unloading point location, and driving route point sequence into a dispatch instruction and issue it to the corresponding truck's on-board terminal.

[0180] The operation monitoring and rescheduling module is used to continuously acquire the actual driving trajectory data and operation status of each truck. When the trajectory deviation exceeds the threshold or the operation is abnormal, the loading / unloading point is reassigned and the driving path point sequence is updated.

[0181] The intelligent dispatching system for mining trucks according to an embodiment of this application can implement any of the above methods, and the specific working process of each module in the system can be referred to the corresponding process in the above method embodiments.

[0182] In the several embodiments provided in this application, it should be understood that the provided methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for example, the division of a certain module is merely a logical functional division, and in actual implementation there may be other division methods, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0183] This application also discloses a computer device.

[0184] A computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a mining truck intelligent scheduling method as described above.

[0185] This application also discloses a computer-readable storage medium.

[0186] A computer-readable storage medium storing a computer program that can be loaded by a processor and executed as described above in any of the intelligent scheduling methods for mining trucks.

[0187] The computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device; the program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0188] The above are all preferred embodiments of this application and are not intended to limit the scope of protection of this application. Any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is only one example of a series of equivalent or similar features.

Claims

1. A method for intelligent scheduling of mining trucks, characterized in that, The scheduling method includes: Collect real-time status data of all transport trucks in the mining operation area, including location data, vehicle speed data, and load data; Obtain the latitude and longitude coordinate sequence of each node in the mine road network and generate the road topology; Receive production task instruction data, including the location of the loading point, the location of the unloading point, and the required amount of ore; Acquire equipment maintenance plan data, and combine it with the real-time status data, road topology and production task instruction data to construct a mixed integer programming model that includes load constraints, time constraints and task coverage constraints; Solve the mixed integer programming model to generate a loading / unloading point allocation scheme for each truck; Based on the loading / unloading point allocation scheme and real-time road traffic flow data, a sequence of driving route points is planned for each truck; The loading point location, unloading point location, and driving path point sequence are encapsulated into a scheduling instruction and sent to the corresponding truck's on-board terminal. The system continuously acquires the actual driving trajectory data and operation status of each truck. When a trajectory deviation exceeds a threshold or an abnormal operation is detected, the loading / unloading point is reassigned and the driving path point sequence is updated. The steps for acquiring equipment maintenance plan data and, in conjunction with the real-time status data, road topology data, and production task instruction data, constructing a mixed-integer programming model that includes load constraints, time constraints, and task coverage constraints include: Read the pre-stored equipment maintenance plan data and extract the next maintenance time and maintenance duration for each truck; Based on the location data and vehicle speed data, the estimated time for each truck to reach each loading point is calculated; Based on the next maintenance time and the estimated time, generate truck availability tags; Based on the loading point location, unloading point location, and ore demand, a task demand matrix is ​​generated. Based on the load data and truck availability labels, a load constraint matrix is ​​constructed; Based on the road topology and real-time status data, a timeliness constraint vector is constructed; Based on the task requirement matrix and the timeliness constraint vector, a task coverage constraint is constructed. Using minimizing the total transportation time as the objective function, a mixed integer programming model is generated by integrating the load constraint matrix, timeliness constraint vector, and task coverage constraint. The steps for solving the mixed-integer programming model to generate the loading / unloading point allocation scheme for each truck include: Invoke the mixed integer programming model and load the load constraint matrix, time constraint vector, and task coverage constraint; The branch and bound algorithm is used to traverse the solution space and generate an initial feasible solution that satisfies all constraints. Based on the matching degree between the truck's load capacity and the task's ore demand, the load utilization rate score of the initial feasible solution is calculated. If the load utilization rate score is lower than the preset score threshold, a neighborhood search is triggered, and the loading / unloading point allocation combination of the two trucks is swapped while maintaining the task coverage constraint. Re-verify the time constraints of the swapped allocation combinations and generate an optimized feasible solution; The feasible solution with the highest load utilization rate is output as the loading / unloading point allocation scheme.

2. The intelligent scheduling method for mining trucks according to claim 1, characterized in that, Based on the loading / unloading point allocation scheme and real-time road traffic flow data, the steps for planning the sequence of driving route points for each truck include: Based on the loading / unloading point allocation scheme, extract the current location data, loading point location, and unloading point location of each truck; Obtain real-time road traffic flow data, including current road traffic density, average vehicle speed, and road congestion status indicators; Based on the path planning algorithm, the initial driving path point sequence is generated by taking the current location data, loading point location, unloading point location, road topology and real-time road traffic flow data as input. Verify whether the initial driving path point sequence conforms to the preset safe area constraints, which include the latitude and longitude range of signal blind spots and dangerous areas; If not, the path point sequence is adjusted to avoid the signal blind spot or danger zone, and the adjusted driving path point sequence is output as the final driving path for each truck; if yes, the initial driving path point sequence is output directly.

3. The intelligent scheduling method for mining trucks according to claim 2, characterized in that, The steps of continuously acquiring the actual driving trajectory data and operational status of each truck, and triggering the reallocation of loading / unloading points and updating the driving path point sequence when a trajectory deviation exceeds a threshold or an abnormal operational status is detected, include: It receives truck positioning data sequences uploaded by the vehicle terminal in real time and generates actual driving trajectory data; Load the current truck's driving path point sequence and calculate the Euclidean distance deviation between the actual driving trajectory data and the driving path point sequence; The system synchronously receives operational status data reported by onboard sensors, including fault codes, load change indicators, and abnormal vehicle speed indicators. When the Euclidean distance deviation value exceeds the preset path offset threshold, or when the operation status data contains a predefined abnormal flag, the truck is marked as an abnormal vehicle. Based on the location data of all vehicles in abnormal states and the remaining task volume, the task coverage constraints and timeliness constraints are reconstructed. The mixed integer programming model is invoked, and a new loading / unloading point allocation scheme is generated by resolving the reconstructed task coverage constraint and time constraint as input. The truck's route sequence is updated based on the new loading / unloading point allocation scheme and real-time road traffic flow data.

4. A method for intelligent scheduling of mining trucks according to any one of claims 1 to 3, characterized in that, After the step of planning a sequence of driving route points for each truck based on the loading / unloading point allocation scheme and real-time road traffic flow data, the method further includes: When the loading / unloading point allocation scheme has the same transportation task that can be assigned to multiple trucks, obtain the pre-configured load-fuel consumption relationship curve data for each truck. Based on the ore transport volume in the loading / unloading point allocation scheme and combined with the truck load data, calculate the real-time load rate of each truck. Based on the load-fuel consumption relationship curve data and real-time load rate, the fuel consumption per unit mileage of each truck is predicted. Load the mine road topology and obtain mine road slope data; Based on the fuel consumption per unit mileage, road gradient data, and road distance of the driving route point sequence, calculate the total predicted fuel consumption for each route option; The truck with the lowest total predicted fuel consumption is selected as the final execution vehicle, and the loading / unloading point allocation scheme is updated.

5. The intelligent scheduling method for mining trucks according to claim 4, characterized in that, The load-fuel consumption curve data is generated through the following steps: Obtain the rated load capacity parameter value for each truck; Based on the rated load parameter value, the load interval sequence is divided according to a preset gradient to generate an interval boundary value set; For each load range, several sets of load-fuel consumption data pairs during constant speed driving are extracted from the historical database. Divide the load value by the rated load parameter value to convert it into a set of load rate values; The least squares method is used to perform quadratic function fitting on the set of load factor values ​​and the corresponding set of fuel consumption values, and the fitting function expression is output as: F(L)=aL 2 +bL+c; where L is the load factor, and a, b, c are the fitting coefficients. The fitting coefficient values ​​a, b, c in the fitting function expression are stored as calibration parameters for the load-fuel consumption relationship curve data and associated with the corresponding truck.

6. A mining truck intelligent dispatching system, characterized in that, For executing the intelligent dispatching method for mining trucks according to any one of claims 1 to 5, the dispatching system comprises: The status data acquisition module is used to collect real-time status data of all transport trucks in the mining operation area, including location data, vehicle speed data, and load data. The road topology generation module is used to obtain the latitude and longitude coordinate sequence of each node in the mine road network and generate the road topology. The production task instruction receiving module is used to receive production task instruction data, including the location of the loading point, the location of the unloading point, and the required amount of ore. The scheduling optimization model construction module is used to acquire equipment maintenance plan data, and combine the real-time status data, road topology and production task instruction data to construct a mixed integer programming model that includes load constraints, time constraints and task coverage constraints. The task allocation scheme solving module is used to solve the mixed integer programming model and generate a loading / unloading point allocation scheme for each truck. The route planning module is used to plan a sequence of driving route points for each truck based on the loading / unloading point allocation scheme and real-time road traffic flow data; The scheduling instruction issuing module is used to encapsulate the loading point location, unloading point location, and driving path point sequence into a scheduling instruction and issue it to the corresponding truck's on-board terminal. The operation monitoring and rescheduling module is used to continuously acquire the actual driving trajectory data and operation status of each truck. When the trajectory deviation exceeds the threshold or the operation is abnormal, the loading / unloading point is reassigned and the driving path point sequence is updated.

7. A computer device, characterized in that: The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that: The computer program is stored that can be loaded by a processor and executed as described in any one of claims 1 to 5.

Images