A garbage transfer vehicle intelligent scheduling method and system

Abstract

This invention belongs to the field of data processing technology, specifically relating to an intelligent scheduling method and system for garbage transfer vehicles. The method includes: constructing feature vectors by performing spatiotemporal alignment and quantization processing on multi-source data, and dynamically generating a priority set of transfer tasks using a pollution prediction model; constructing a dynamic comprehensive travel time model that integrates real-time traffic congestion, predicted traffic trends, and vehicle operation modes; using an improved genetic algorithm to evaluate path fitness with dynamic travel time as the core indicator, and optimizing the path; smoothing the optimized path and issuing control commands. This invention, by introducing dynamic time-varying costs into the core of path planning decision-making, can plan realistically optimal paths that intelligently avoid traffic congestion, significantly improving the operational efficiency of garbage transfer vehicles and their adaptability to dynamic urban environments.

Description

Technical Field

[0001] This invention relates to the field of data processing technology. More specifically, this invention relates to an intelligent scheduling method and system for garbage transfer vehicles. Background Technology

[0002] Garbage transfer vehicles are an important component of smart city infrastructure, and route planning is a core technology that determines their operational efficiency, resource utilization, and operating costs.

[0003] Currently, one existing technical approach is to use a two-stage planning method: First, in the task generation stage, by integrating multi-source data such as historical transfer data, urban pedestrian activity data, and weather information, time series prediction models such as the Autoregressive Integrated Moving Average (ARIMAX) model with exogenous variables are used to predict the future waste generation of various urban road sections, thereby dynamically identifying the task points that need to be transferred. Second, in the route optimization stage, the generated set of task points is treated as a vehicle routing problem, and heuristic optimization algorithms such as genetic algorithms are used to solve the optimal route for the waste transfer fleet to complete all tasks.

[0004] However, when using genetic algorithms to evaluate path selection, the fitness function typically relies on static geographic information system (GIS) data, using fixed road mileage as the primary measure of path cost. This approach inherently assumes that road tolls are constant. However, in real urban environments, actual road tolls are primarily reflected in travel time, which is significantly influenced by traffic flow, exhibiting high dynamism and time-varying characteristics. Therefore, while genetic algorithms relying entirely on static distances can find a theoretically shortest path, this path may, in practice, result in excessively long travel times due to traffic congestion, severely reducing the operational efficiency of garbage trucks. Summary of the Invention

[0005] To address the technical problem of low operational efficiency caused by static evaluation of path costs, this invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides an intelligent scheduling method for garbage transfer vehicles, comprising: acquiring multi-source data and using a preset pollution prediction model to predict the future garbage index of each road segment in the urban road network, and generating a priority set of transfer tasks by combining the regional importance coefficient of the road segments; constructing a dynamic comprehensive travel time model to calculate the dynamic comprehensive travel time required for a garbage transfer vehicle to pass through any road segment at the expected arrival time, wherein the dynamic comprehensive travel time is related to the static geographical distance of the road segment, real-time traffic congestion, historical traffic trends, and vehicle operation mode; embedding the dynamic comprehensive travel time model into the fitness function of a genetic algorithm, using the total travel time of the route plan and the priority of untransferred tasks as evaluation indicators, iteratively solving the optimal transfer path corresponding to the set of transfer tasks; smoothing the optimal transfer path to generate control instructions that can be executed by the garbage transfer vehicle, thereby realizing dynamic intelligent planning of the transfer path of the garbage transfer vehicle.

[0007] This invention solves the problem of the disconnect between planning and reality by introducing a travel time model that reflects real-world traffic dynamics as the core basis for route optimization, enabling garbage transfer vehicles to intelligently avoid congestion and thus significantly improving overall operational efficiency.

[0008] Preferably, the step of acquiring multi-source data and using a preset pollution prediction model to predict the future waste index of each road segment in the urban road network includes: performing spatiotemporal alignment processing on multi-source data, including historical transfer volume, pedestrian flow heat map, weather, and public activities, to ensure consistency in road segment division and time granularity; generating a unified format numerical feature vector based on the spatiotemporally aligned multi-source data, wherein pedestrian flow heat map data is converted into pedestrian flow density scores, weather data is converted into corresponding weather condition level values, and public activity data is converted into event weight values ​​reflecting the degree of impact of the activities; using the historical waste index of each road segment as an endogenous variable and the numerical feature vector as an exogenous variable, both are input into an autoregressive integral moving average model with exogenous variables for training, and the trained model is used to predict the future waste index of each road segment.

[0009] This invention transforms heterogeneous data, which was originally impossible to directly compare and calculate, into unified numerical features by standardizing spatiotemporal alignment and quantization. This provides high-quality input for the accurate prediction of subsequent models, thereby improving the accuracy and reliability of pollution prediction.

[0010] Preferably, the dynamic integrated passage time satisfies the following relationship: Where i is the starting node of the road segment, and j is the ending node of the road segment. Let i be the estimated time when the vehicle arrives at the starting node. In order to be in The dynamic overall travel time from node i to node j at any given moment. Let be the static geographic distance of the road segment from node i to node j. Let be the free-flow velocity of the road segment from node i to node j, used to characterize the average vehicle speed under ideal free-flow conditions. Real-time traffic congestion factors, To predict traffic trend factors, This represents the influencing factor of work mode.

[0011] Preferably, the real-time traffic congestion factor satisfies the following relationship: ;in, For the road segment from node i to node j Real-time average vehicle speed at any given time and Adjust parameters for the preset Sigmoid function used to control the severity of congestion penalties and trigger points.

[0012] By introducing real-time traffic congestion factors, route planning can respond instantly to sudden traffic congestion on the road, ensuring the immediate effectiveness of the planning.

[0013] Preferably, the predicted traffic trend factors satisfy the following relationship: Where N is the number of historical weeks used for prediction. Let be the weight corresponding to the k-th historical week, and the sum of the weights of all historical weeks is 1. For the road segment from node i to node j in the historical week k, and... Historical traffic congestion factors for the corresponding weekday date and hour.

[0014] By introducing traffic trend prediction factors, the route planning system is given the ability to anticipate periodic congestion, enabling it to plan detour routes in advance. This achieves a shift from passive response to proactive avoidance, enhancing the forward-looking nature of the planning.

[0015] Preferably, the operation mode influence factor is determined as follows: if the road segment from node i to node j belongs to the transfer task set, the operation mode influence factor is equal to the ratio of the free flow speed of the road segment to the preset transfer operation speed; if the road segment from node i to node j does not belong to the transfer task set, the operation mode influence factor is 1.

[0016] By distinguishing between the traffic status and the operation status of vehicles, the model can more accurately calculate the total route time, thus improving the accuracy of the planning results.

[0017] Preferably, the process of iteratively solving for the optimal transit path corresponding to the transit task set is as follows: for the sequence of task points... Any vehicle path formed, where For station nodes, to For task point nodes, The total number of task points; the total travel time is calculated recursively, and the next segment... The dynamic integrated travel time is based on the estimated arrival time of the destination of the previous road segment. The calculations are performed, among which This is the index of the task point sequence.

[0018] By recursively accumulating the total time, the genetic algorithm ensures that it fully considers the time evolution and road condition changes during the path execution process when evaluating each potential path, thus enabling it to search for a truly time-optimal solution.

[0019] Preferably, the specific process of smoothing the optimal transfer path is to fit the optimal transfer path composed of discrete task points using a B-spline curve to generate a smooth trajectory with continuous curvature.

[0020] Secondly, the present invention provides an intelligent scheduling system for garbage transfer vehicles, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned intelligent scheduling method for garbage transfer vehicles is implemented.

[0021] By adopting the above technical solution, a computer program for the intelligent scheduling method of garbage transfer vehicles is generated and stored in a memory so that it can be loaded and executed by a processor. Terminal devices can then be made based on the memory and processor for convenient use.

[0022] This invention constructs a multi-factor dynamic comprehensive travel time model and deeply embeds it into the fitness evaluation core of the genetic algorithm. It abandons the dependence of traditional planning methods on static distance, and can not only respond to the current traffic congestion in real time, but also predict the upcoming periodic congestion through weighted learning of historical data. At the same time, it distinguishes the travel and operation status of vehicles in different road sections, so as to calculate a more accurate travel time cost in the real world.

[0023] Furthermore, the improved genetic algorithm can search for the most efficient path in a dynamic traffic environment within a vast path solution space, rather than the theoretical shortest path. This makes the planning results more consistent with actual execution, effectively improving the operational efficiency of the garbage transfer fleet and its adaptability to complex urban environments. Attached Figure Description

[0024] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent upon reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of the invention are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein:

[0025] Figure 1 This is a flowchart illustrating an intelligent scheduling method for garbage transfer vehicles according to the present invention;

[0026] Figure 2 This is a schematic diagram comparing the effects of paths planned by the method according to an embodiment of the present invention with those planned by traditional methods. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments 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 are within the scope of protection of the present invention.

[0028] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0029] This invention discloses an intelligent scheduling method for garbage transfer vehicles, referring to... Figure 1 This includes steps S1-S4:

[0030] S1. Acquire multi-source data and use a preset pollution prediction model to predict the future waste index of each road segment in the urban road network, and generate a set of priority transfer tasks by combining the regional importance coefficient of the road segment.

[0031] In an optional embodiment, the data collected by the system includes historical garbage transfer volume recorded hourly for each road segment, pedestrian flow heat map, weather forecast and public event calendar, etc. For unified processing, all data are mapped onto a grid with urban road segments as the basic spatial unit and hours as the basic time unit. Through spatiotemporal alignment processing, the data is kept consistent in terms of road segment division and time granularity.

[0032] In this optional embodiment, the aligned data is used as a numerical feature vector in a uniform format: the pedestrian heat map is converted into a pedestrian density score in the range of [0,1]; the weather conditions are converted into level values ​​(e.g., sunny = 0, light rain = 2); and public activities are converted into event weight values ​​(e.g., the affected road segment has a weight of 1 during the event, otherwise it is 0).

[0033] Furthermore, an autoregressive integral moving average (ARIMAX) model with exogenous variables can be used to predict the future garbage index. In this model, the historical garbage index time series of each road segment is used as the endogenous variable, while the resulting feature vector matrix is ​​used as the exogenous variable. The model learns the laws of the endogenous variables and the influence of the exogenous variables on them through training, and then predicts the garbage index of each road segment in each hour in the future after training.

[0034] In an optional embodiment, to prioritize the allocation of transportation resources to the areas most in need, a prioritized set of transportation tasks can be generated by combining the regional importance coefficient of road segments. Each road segment can be assigned a regional importance coefficient based on an urban functional zoning map; for example, hospitals and schools are assigned a coefficient of 1.0, commercial areas 0.8, and ordinary residential areas 0.6. The final priority of the tasks satisfies the following relationship:

[0035]

[0036] in, Let i be the final priority of the task at time t. It is the predicted garbage index for road segment i in time slice t. It is the regional importance coefficient of road segment i. and These are adjustable weighting coefficients corresponding to the waste index and regional importance, for example, The value is 0.7. The value is 0.3.

[0037] In this optional embodiment, when the calculated final priority of the task is greater than a preset threshold such as 0.8, a corresponding transfer task can be generated to form a set of transfer tasks to be planned.

[0038] In this way, by predicting pollution levels and prioritizing them based on regional importance, it is possible to ensure that limited sanitation resources are used for the most critical and urgent transportation tasks, thereby improving the intelligence level and response speed of urban management.

[0039] S2. Construct a dynamic comprehensive travel time model to calculate the dynamic comprehensive travel time required for garbage transfer vehicles to pass through any road segment at the expected arrival time. The dynamic comprehensive travel time is related to the static geographical distance of the road segment, real-time traffic congestion, historical traffic trends, and vehicle operation mode.

[0040] In an optional embodiment, for any road segment in the road network that connects to the starting node i and the ending node j, its dynamic comprehensive travel time satisfies the following relationship:

[0041]

[0042] Where i is the starting node of the road segment, and j is the ending node of the road segment. Let i be the estimated time when the vehicle arrives at the starting node. In order to be in The dynamic overall travel time from node i to node j at any given moment. Let be the static geographic distance of the road segment from node i to node j. Let be the free-flow velocity of the road segment from node i to node j, used to characterize the average vehicle speed under ideal free-flow conditions. Real-time traffic congestion factors, To predict traffic trend factors, This represents the influencing factor of work mode.

[0043] Specifically, the calculation formula for dynamic integrated travel time is derived from the basic travel time. It consists of three dynamic correction factors, where the real-time traffic congestion factor satisfies the following relationship:

[0044]

[0045] in, For the road segment from node i to node j Real-time average vehicle speed at any given time and For example, parameters are adjusted for a preset Sigmoid function used to control the severity and trigger point of congestion penalties. The value is 5. The value is 1.5.

[0046] Specifically, This information can be obtained in real time by the system through the city's intelligent transportation system or an API provided by a third-party map service provider. near At that time, the speed ratio is close to 1, and the function output is close to 0. When the value is close to 1, the path cost is not significantly affected; when It begins to decline, the speed ratio increases, and once it exceeds the threshold of 1.5, The value will increase rapidly and approach 2, imposing a significant penalty on travel time. This non-linear adjustment is more in line with the abrupt change in traffic flow from smooth to congested.

[0047] To address predictable periodic congestion, the system can maintain a historical traffic database for each key road segment, storing traffic congestion factors for a specific hour (e.g., 8:00 AM) on a specific day (e.g., Wednesday) of the past week. The predicted traffic trend factors satisfy the following relationship:

[0048]

[0049] Where N is the number of historical weeks used for prediction. Let be the weight corresponding to the k-th historical week, and the sum of the weights of all historical weeks is 1. For the road segment from node i to node j in the historical week k, and... Historical traffic congestion factors for the corresponding weekday date and hour.

[0050] Specifically, to make recent data play a greater role, you can set... Decays over time, for example, ,in This is the attenuation coefficient, which can be set to 0.9. By weighting and averaging the congestion data from the same historical period, the model can predict the most likely level of congestion on that road segment at a future point in time. By assigning higher weights to recent data, the model can adapt to gradual changes in traffic patterns more quickly.

[0051] Specifically, the historical traffic congestion factor is not an independent raw data, but the result of recalculating historical data. For example, the historical traffic congestion factor at a certain historical moment, such as 8 a.m. last Wednesday, is calculated using the historical vehicle speed data at that moment, through the same formula used to calculate the real-time traffic congestion factor.

[0052] The method for determining the operation mode influence factor is as follows: if the road segment from node i to node j belongs to the transfer task set, the operation mode influence factor is equal to the ratio of the free flow speed of the road segment to the preset transfer operation speed; if the road segment from node i to node j does not belong to the transfer task set, the operation mode influence factor is 1.

[0053] Thus, by constructing a multi-dimensional cost model that integrates real-time traffic conditions, historical patterns, and operational characteristics, a cost basis that is closer to real-world travel time can be provided for route planning.

[0054] S3. Embed the dynamic integrated travel time model into the fitness function of the genetic algorithm, and use the total travel time of the route scheme and the priority of the untransfer tasks as evaluation indicators to iteratively solve the optimal transfer path corresponding to the transfer task set.

[0055] In an optional embodiment, the dynamic integrated travel time model can be embedded into the fitness function of the genetic algorithm. The fitness function is used to evaluate the merits of different path plans. A good path plan should complete as many high-priority tasks as possible in the shortest possible time. In this invention, the fitness function satisfies the following relationship:

[0056]

[0057] in, Let P be the fitness function value of the path scheme P. Let P be the path of the o-th vehicle. Complete the path for vehicle o. Total time required, where O is the total number of vehicles. Let P be the set of tasks not covered by the path scheme P. The priority of task c that has not been transferred. and These are the weighting coefficients for time and priority.

[0058] In this optional embodiment, the total time cost plays a decisive role in the fitness function, determining the quality of a path scheme. For the path of the o-th vehicle in path scheme P... Its path starts from a station node. and task nodes are to Composed of, i.e., paths The task point sequence is , The total number of task points; the total travel time is calculated recursively, and the next segment... The dynamic integrated travel time is based on the estimated arrival time of the destination of the previous road segment. The calculations are performed, among which This is the index of the task point sequence.

[0059] in, Satisfying the relation:

[0060]

[0061] in, For the mission point The service time required to perform transfer operations is calculated in a recursive manner, ensuring that the correct estimated arrival time can be used to query and calculate the dynamic transit time for each segment of the route when evaluating it.

[0062] Furthermore, the algorithm iterates using standard genetic operators such as selection, crossover, and mutation, calculating the real time cost of each path recursively until the path with the highest fitness is found. Tournament selection is used to reduce the risk of premature convergence; ordered crossover (OX) is used to ensure that the generated offspring paths are valid task permutations; and swap mutation is used to randomly swap two task points in the path to increase population diversity. The algorithm terminates when it reaches the maximum number of iterations or when the optimal solution shows no significant improvement over multiple generations, outputting the path with the highest fitness.

[0063] Thus, by deeply binding the recursive calculation logic of dynamic travel time with total path time, the fitness evaluation of the genetic algorithm can reflect the real time cost of the path in a dynamic traffic environment, thereby ensuring that the optimization result is the optimal solution in the real world.

[0064] S4. Smooth the optimal transfer path and generate control commands that can be executed by the garbage transfer vehicles, thereby realizing dynamic intelligent planning of the transfer path of the garbage transfer vehicles.

[0065] In an optional embodiment, the output of the genetic algorithm is a discrete sequence of task point visits, which is processed in this step to enable the vehicle to execute smoothly and safely. First, a B-spline curve is used to fit the path composed of discrete task points, thereby generating a smooth trajectory with continuous curvature that conforms to the vehicle's kinematic characteristics.

[0066] Furthermore, this smooth trajectory is discretized into a dense series of path points. Each path point contains not only precise location coordinates but also the desired travel speed, heading angle, and operational instructions to be executed on that section of the road, such as the transfer brush rotation speed and water spray volume. This complete sequence of instructions is ultimately sent to the vehicle's underlying control unit to directly guide the vehicle in completing the transfer operation.

[0067] like Figure 2The diagram illustrates a comparison of path planning results using the method of this invention and the traditional method in a city transportation task. The red dashed line represents the path planned by the traditional method, which, in pursuit of the shortest geometric distance (498.15), directly traverses the gray congested area, resulting in an actual travel time as high as 649.89 seconds. In contrast, the green solid line planned by this invention, by dynamically anticipating congestion, actively chooses a detour. Even when some task points are located in congested areas, although their theoretical distance is longer, the congestion is successfully avoided, and the actual travel time (545.35 seconds) is significantly lower than that of the red path.

[0068] In this way, by smoothing the macro path and generating micro control commands, the entire link from high-level task planning to low-level vehicle control is connected, ensuring that the planning results can be executed accurately and safely by the vehicle.

[0069] This invention also discloses an intelligent scheduling system for garbage transfer vehicles, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent scheduling method for garbage transfer vehicles according to this invention is implemented.

[0070] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0071] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise expressly and specifically defined.

[0072] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A method for intelligent scheduling of garbage transfer vehicles, characterized in that, include: Acquire multi-source data and use a pre-set pollution prediction model to predict the future waste index of each road segment in the urban road network, and generate a set of priority transfer tasks by combining the regional importance coefficient of the road segment. A dynamic integrated travel time model is constructed to calculate the dynamic integrated travel time required for garbage transfer vehicles to pass through any road segment at their expected arrival time. The dynamic integrated travel time satisfies the following relationship: Where i is the starting node of the road segment, and j is the ending node of the road segment. Let i be the estimated time when the vehicle arrives at the starting node. In order to be in The dynamic overall travel time from node i to node j at any given moment. Let be the static geographic distance of the road segment from node i to node j. Let be the free-flow velocity of the road segment from node i to node j, used to characterize the average vehicle speed under ideal free-flow conditions. Real-time traffic congestion factors, To predict traffic trend factors, Factors influencing work patterns; Real-time traffic congestion factors satisfy the following relationship: in, For the road segment from node i to node j Real-time average vehicle speed at any given time and The parameters are adjusted for the preset Sigmoid function used to control the severity of congestion penalties and trigger points; The dynamic integrated travel time model is embedded into the fitness function of the genetic algorithm. The total travel time of the route scheme and the priority of the untransfer tasks are used as evaluation indicators to iteratively solve the optimal transfer path corresponding to the transfer task set. The optimal transfer path is smoothed to generate control commands that can be executed by waste transfer vehicles, thereby realizing dynamic intelligent planning of the transfer path of waste transfer vehicles.

2. The intelligent scheduling method for garbage transfer vehicles according to claim 1, characterized in that, The process of acquiring multi-source data and using a pre-set pollution prediction model to predict the future waste index of each road segment in the urban road network includes: Spatiotemporal alignment processing is performed on multi-source data, including historical transit volume, pedestrian flow heat maps, weather, and public activities, to ensure consistency in road segment division and time granularity. Numerical feature vectors in a unified format are generated based on spatiotemporally aligned multi-source data. Among them, pedestrian flow heat map data is converted into pedestrian flow density scores, weather data is converted into corresponding weather condition level values, and public activity data is converted into event weight values ​​that reflect the degree of impact of the activities. The historical garbage index of each road segment is used as an endogenous variable, and the numerical feature vector is used as an exogenous variable. They are jointly input into an autoregressive integral moving average model with exogenous variables for training, and the future garbage index of each road segment is predicted using the trained model.

3. The intelligent scheduling method for garbage transfer vehicles according to claim 1, characterized in that, The predicted traffic trend factors satisfy the following relationship: Where N is the number of historical weeks used for prediction. Let be the weight corresponding to the k-th historical week, and the sum of the weights of all historical weeks is 1. For the road segment from node i to node j in the historical week k, and... Historical traffic congestion factors for the corresponding weekday date and hour.

4. The intelligent scheduling method for garbage transfer vehicles according to claim 1, characterized in that, The method for determining the influencing factors of the aforementioned work mode is as follows: If the road segment from node i to node j belongs to the set of transfer tasks, then the operation mode influence factor is equal to the ratio of the free flow speed of the road segment to the preset transfer operation speed. If the route from node i to node j does not belong to the set of transfer tasks, then the operation mode influence factor is 1.

5. The intelligent scheduling method for garbage transfer vehicles according to claim 1, characterized in that, The iterative process of finding the optimal transfer path corresponding to the set of transfer tasks is as follows: For a sequence of task points... Any vehicle path formed, where For station nodes, to For task point nodes, The total number of task points; the total travel time is calculated recursively, and the next segment... The dynamic integrated travel time is based on the estimated arrival time of the destination of the previous road segment. The calculations are performed, among which This is the index of the task point sequence.

6. The intelligent scheduling method for garbage transfer vehicles according to claim 1, characterized in that, The specific process of smoothing the optimal transfer path involves fitting the optimal transfer path, which consists of discrete task points, with a B-spline curve to generate a smooth trajectory with continuous curvature.

7. An intelligent dispatching system for garbage transfer vehicles, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement a smart scheduling method for garbage transfer vehicles according to any one of claims 1-6.

Images