Intelligent port operation vehicle scheduling system and scheduling robot

Abstract

The application discloses a kind of wisdom port operation vehicle scheduling system and scheduling robot, it is related to wisdom port operation technical field.To solve the problem of slow response, high error rate, uneven resource allocation, high labor cost and difficult to meet the efficient and intelligent needs of modern port in traditional port vehicle scheduling;The application reduces the empty distance of vehicle by using two-part graph matching strategy, improves the heavy load rate, while the prediction scheduling module plans in advance, reduces the backlog of tasks and vehicle idle, effectively improves the operation efficiency, the constraint processing module optimizes task allocation according to the task and vehicle condition, the multi-objective optimization module considers the heavy load rate and single fairness, so that the resource allocation is more reasonable, the scheduling robot integrated system function, through data acquisition and online learning technology continuous optimization, can adapt to different working environment, realize intelligent scheduling, not only improve scheduling accuracy and efficiency, but also reduce labor cost, improve the overall competitiveness of port.

Description

Technical Field

[0001] This invention relates to the field of smart port operation technology, and in particular to a smart port operation vehicle dispatching system and dispatching robot. Background Technology

[0002] As a key hub for international trade, the operational efficiency of ports is crucial for the smooth operation of the logistics chain and cost control. In port operations, vehicle scheduling directly impacts cargo loading and unloading efficiency, vehicle turnover rate, and overall operating costs. However, traditional manual scheduling methods have significant drawbacks, including slow response times, high error rates, uneven resource allocation, and high labor costs. In the context of modern ports pursuing efficient and intelligent operations, traditional scheduling methods are no longer sufficient to meet the demands of modern ports. Summary of the Invention

[0003] The purpose of this invention is to provide a smart port operation vehicle scheduling system and scheduling robot, which uses a bipartite graph strategy to perform intelligent operation vehicle scheduling, improve port operation efficiency, optimize the vehicle scheduling process, and reduce operating costs, thereby solving the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A smart port operation vehicle dispatching system includes:

[0006] The task matching module is configured to match the acquired job tasks and vehicle information based on a bipartite graph matching strategy to obtain the task allocation results;

[0007] The constraint processing module is configured to prioritize job tasks, adjust the corresponding vehicle information weights based on weight factors, and optimize the task allocation results based on the bipartite graph matching strategy based on task priority and vehicle information weights.

[0008] The multi-objective optimization module is configured to obtain the overload rate and order dispatch fairness evaluation index based on the optimized task allocation results, and adjust the weight factors based on the overload rate and order dispatch fairness evaluation index.

[0009] The predictive scheduling module is configured to acquire historical data on port vehicle scheduling, predict vehicle locations and future task arrival based on the historical data, integrate the prediction results with the current task allocation results to generate scheduling strategies, and monitor and provide early warnings for the scheduling and travel paths of vehicles during port transportation.

[0010] Furthermore, the bipartite graph matching strategy specifically includes:

[0011] The port vehicle scheduling task is transformed into a bipartite graph matching problem, and a bipartite graph framework is built.

[0012] Each job to be assigned is treated as a set of vertices in a bipartite graph framework, and the information on schedulable vehicles is treated as another set of vertices in the bipartite graph framework.

[0013] Between the vertex set of the task and the vertex set of vehicle information, edges are constructed based on the potential relationship between the task and vehicle information, and the weight of each edge is determined. No two edges have a common vertex.

[0014] The weight of each edge is set according to the distance between the vehicle's current position and the corresponding task starting point;

[0015] Each job task to be assigned is matched with the information of schedulable vehicles based on the weight of each edge, and the corresponding job task and vehicle information are marked with status based on the matching results.

[0016] Furthermore, before matching the acquired task and vehicle information, the process also includes:

[0017] The acquired task and vehicle information are read and divided into a multi-level topic tree. Each level is divided into multiple clusters according to task type and vehicle attributes.

[0018] Clustering is performed on the data information corresponding to each layer to analyze the task characteristics of the operation and the attribute information of the vehicle, resulting in several cluster sets, and each cluster set is assigned to the corresponding cluster.

[0019] The task description statements and vehicle information description statements contained in the cluster set in each cluster are segmented into words, and then cleaned and deduplicated to obtain a keyword set and perform word vector conversion.

[0020] Based on the word vectors representing the core features of the cluster as defined by port business standards, the keyword corresponding to the word vector with the smallest distance is selected from the keyword set and used as the target keyword for that cluster.

[0021] Based on the target keywords, key information corresponding to the task and vehicle information is labeled with keywords to determine the key features of the task and vehicle information.

[0022] Furthermore, the constraint processing module optimizes the task allocation results, specifically as follows:

[0023] Based on the key characteristics of the task, determine the importance, urgency and type of goods involved in the task, determine the priority score of the task, and rank the tasks from high to low based on the priority score.

[0024] Based on the key features of vehicle information, the vehicle's performance is evaluated to determine the vehicle's comprehensive performance coefficient. At the same time, the driver's working hours and work experience corresponding to the vehicle information are extracted to evaluate the driver's work efficiency coefficient.

[0025] The vehicle's overall performance coefficient is compared with the preset vehicle performance reference value to obtain the vehicle performance ratio. The driver's work efficiency coefficient is compared with the preset driver efficiency reference value to obtain the corresponding driver efficiency ratio.

[0026] Vehicle weights are adjusted based on vehicle performance ratios and corresponding driver efficiency ratios, and in conjunction with the priority of the task, vehicles with higher weights that meet the requirements of the task's vehicle type and cargo type are given priority allocation.

[0027] Furthermore, the multi-objective optimization module obtains evaluation indicators for overload rate and dispatch fairness, specifically including:

[0028] During the task allocation process, the task load of the assigned vehicles is continuously monitored to obtain the ratio of the actual weight of the cargo loaded on the vehicle to the maximum load capacity of the vehicle, and the heavy load rate of each vehicle is calculated.

[0029] Statistical analysis of the load factor of all vehicles is performed to obtain overall vehicle load factor data.

[0030] The number of tasks received by each vehicle within a certain time period is counted, and the variance of the number of tasks for different vehicles is calculated to assess the fairness of task assignment. The larger the variance, the greater the difference in the number of tasks between vehicles, and the worse the fairness of task assignment. The smaller the variance, the better the fairness of task assignment.

[0031] Furthermore, the multi-objective optimization module also includes:

[0032] Based on the results, extract the driver's historical driving data corresponding to the vehicle information, and analyze the driver's driving experience value under the priority of the current task based on the weight value of the driver's historical driving data in the total historical driving data.

[0033] Determine the driver's maximum continuous driving time based on historical driving data;

[0034] Input the vehicle's maximum speed and all historical speeds from the historical speed list into the preset data axis to obtain the speed difference between the maximum speed and each historical speed.

[0035] The maximum driving speed is adjusted based on the average value of the speed difference to generate the adjustment speed;

[0036] The system generates a suggested driving speed range based on the adjustment speed and driving experience value. At the same time, it adjusts the maximum continuous driving time based on driving experience value and generates a suggested driving time.

[0037] Based on the recommended driving speed range and recommended driving time, a driving plan for the task is generated, and the driver's driving behavior is analyzed in real time during the task to determine whether it conforms to the driving plan.

[0038] If the conditions are not met, extract the driver's abnormal items and trigger the corresponding alarm.

[0039] Furthermore, the predictive scheduling module also includes:

[0040] During the task allocation process, if it is predicted that a large number of tasks will occur in the target area and there are not enough vehicles in the area, vehicles from non-target areas will be guided to the target area to stand by.

[0041] At the same time, the task type and vehicle status of the corresponding operation task are determined based on the prediction results, and the corresponding vehicle is matched to perform the corresponding operation task.

[0042] Real-time feedback data is acquired, and the actual scheduling situation is compared with the prediction results. The prediction results are then dynamically adjusted based on the comparison results.

[0043] Furthermore, the predictive scheduling module monitors and provides early warnings regarding the vehicle's travel path during port transportation, specifically including:

[0044] Based on the vehicle's dispatching origin and transportation priority, and combined with the road condition information of the road network system, the initial dispatching route of the vehicle is determined;

[0045] Real-time acquisition of traffic flow, weather conditions, accident and road maintenance information under the initial scheduling path;

[0046] The congestion value of the initial dispatch route is determined based on the traffic flow, the driving difficulty value of the initial dispatch route is determined based on the weather conditions, and the probability value of emergencies of the initial dispatch route is determined based on accident and road maintenance information.

[0047] Based on the congestion value, driving difficulty value, and probability value of emergencies, a comprehensive evaluation of the initial scheduling path is determined to obtain a path evaluation value. It is then determined whether the path evaluation value is greater than a first preset evaluation value. If so, the initial scheduling path is determined to be the optimal scheduling path; otherwise, it is determined whether the path evaluation value is greater than a second preset evaluation value.

[0048] If so, determine the initial scheduling path as a candidate scheduling path;

[0049] Otherwise, the initial scheduling path is determined to be an unqualified scheduling path;

[0050] When the initial scheduling path is the optimal scheduling path, the road condition characteristics and surrounding environment characteristics of the optimal scheduling path are obtained, and abnormal alarm conditions are formulated based on the road condition characteristics and surrounding environment characteristics.

[0051] When the initial scheduling path is a candidate scheduling path, other candidate scheduling paths are obtained based on the comprehensive evaluation value of the candidate scheduling path determined by the road network system within a preset difference. The candidate scheduling paths and other candidate scheduling paths are integrated to obtain a comprehensive scheduling path. When the initial scheduling path is an unqualified scheduling path, other scheduling paths with a value greater than a second preset evaluation value are determined based on the road network system. The other scheduling paths are integrated to obtain a comprehensive scheduling path.

[0052] Based on the road condition characteristics and surrounding environment characteristics of the comprehensive scheduling path, abnormal alarm conditions are formulated.

[0053] Furthermore, based on the road condition characteristics and surrounding environment characteristics of the comprehensive scheduling path, abnormal alarm conditions are formulated, including:

[0054] Based on the road network system, important nodes are marked in the comprehensive scheduling path to obtain important road condition nodes and important environmental nodes;

[0055] Based on the road conditions at key road nodes, the maximum dwell time at those key road nodes is calculated using the following formula; in, This indicates the maximum dwell time at the key road condition nodes. Indicates the preset minimum stay time. This represents the second preset evaluation value. This represents the path evaluation value at the critical node of the road condition. This indicates the number of other important road condition nodes adjacent to the aforementioned important road condition node. This represents the path evaluation value for the i-th other important road condition node;

[0056] Based on the environment of important environmental nodes, the maximum stay time at the important environmental nodes is calculated according to the following formula; in, This indicates the maximum dwell time at the key nodes in the environment. represents the natural constant, with a value of 2.72. This indicates the probability of an operational anomaly occurring under standard conditions. This indicates the probability of an operational anomaly occurring at a critical node in the environment.

[0057] Based on the route of the comprehensive scheduling path, location anomaly conditions are established. Based on the maximum dwell time of important road condition nodes and the maximum dwell time of important environmental nodes, dwell time anomaly conditions are established. Based on the location anomaly conditions and dwell time anomaly conditions, anomaly alarm conditions are formulated.

[0058] This invention provides another technical solution: a smart port operation vehicle dispatching robot, comprising: configuring a port operation vehicle dispatching system in the operation vehicle dispatching robot; the operation vehicle dispatching robot acquires heavy load rate and dispatch fairness evaluation indicators, performs functions such as task and vehicle matching, handles complex task constraints, achieves multi-objective optimization, and makes dispatching decisions based on prediction results; and establishes a data interaction channel with vehicles in the port operation scenario to intelligently dispatch vehicles and collect data in real time during the dispatching process through the data interaction channel.

[0059] Compared with the prior art, the beneficial effects of the present invention are:

[0060] By utilizing a bipartite graph matching strategy to reduce vehicle idle distance and increase load factor, while the predictive scheduling module plans ahead to reduce task backlog and vehicle idleness, effectively improving operational efficiency, the constraint processing module optimizes task allocation based on task and vehicle conditions, and the multi-objective optimization module balances load factor and dispatch fairness, making resource allocation more reasonable. The scheduling robot integrates system functions and continuously optimizes through data collection and online learning technology, enabling it to adapt to different operating environments and achieve intelligent scheduling. This not only improves scheduling accuracy and efficiency but also reduces labor costs, enhances the overall competitiveness of the port, promotes the intelligent development of the port, and brings efficient and intelligent solutions to port operations. Attached Figure Description

[0061] Figure 1 This is a block diagram of the intelligent port operation vehicle dispatching system of the present invention. Detailed Implementation

[0062] 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 embodiments of the present invention, and not all embodiments. 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.

[0063] Please see Figure 1 This embodiment provides the following technical solution:

[0064] A smart port operation vehicle dispatching system includes:

[0065] The task matching module is configured to match the acquired work tasks and vehicle information based on a bipartite graph matching strategy, with the primary goal of improving the heavy load rate, and to obtain the task allocation results.

[0066] The constraint processing module is configured to prioritize job tasks, adjust the corresponding vehicle information weights based on weight factors, and optimize the task allocation results based on the bipartite graph matching strategy based on task priority and vehicle information weights.

[0067] The multi-objective optimization module is configured to obtain the overload rate and dispatch fairness evaluation index based on the optimized task allocation results, and further dynamically optimize the task allocation results based on the adjustment weight factors of the overload rate and dispatch fairness evaluation index to achieve the optimization of the overall scheduling effect.

[0068] The predictive scheduling module is configured to acquire historical data on port vehicle scheduling, including vehicle trajectories, task completion times, and changes in task volume over different time periods. Based on this historical data, it predicts vehicle locations and future task arrivals. For example, by analyzing past data, it learns vehicle driving patterns, the time and location distribution patterns of tasks, and predicts the possible locations of vehicles in specific time periods and areas, as well as the probability and approximate time of different types of tasks. The prediction results are then integrated with the current task allocation results to generate a scheduling strategy, assisting in the development of more scientific and reasonable scheduling plans. Other features include:

[0069] During the task allocation process, if it is predicted that a large number of tasks will appear in the target area and there are not enough vehicles in the area, vehicles from non-target areas will be guided to the target area to stand by, so as to avoid task backlog and vehicle idleness.

[0070] At the same time, based on the prediction results, the task type and vehicle status of the corresponding operation task are determined, and the corresponding vehicle is matched to execute the corresponding operation task, thereby improving the task execution efficiency and vehicle utilization.

[0071] By acquiring feedback data in real time, comparing the actual scheduling situation with the predicted results, and dynamically adjusting the prediction results based on the comparison results, the accuracy of predictions and the scientific nature of scheduling decisions are continuously improved, ensuring that the port vehicle scheduling system can adapt to the ever-changing operating environment and achieve efficient and intelligent scheduling.

[0072] In this embodiment, vehicle resources, including vehicle type, capacity, and operational capabilities, are initialized using real-time location, status, and task requirements of port operation vehicles. Task requests, including cargo type, weight, and destination, are received from the port management system. Tasks are categorized and prioritized, distinguishing between VIP and regular tasks. This clear categorization and prioritization of vehicle resources and tasks provides a basis for the rational arrangement of vehicle task execution, ensuring that important and urgent tasks are handled first, thereby improving the overall operational efficiency of the port.

[0073] In this embodiment, the bipartite graph matching strategy specifically includes:

[0074] The port vehicle scheduling task is transformed into a bipartite graph matching problem, and a bipartite graph framework is built.

[0075] Each job to be assigned is treated as a set of vertices in a bipartite graph framework, and the information on schedulable vehicles is treated as another set of vertices in the bipartite graph framework.

[0076] Between the vertex set of the task and the vertex set of vehicle information, edges are constructed based on the potential relationship between the task and vehicle information, and the weight of each edge is determined. No two edges have a common vertex.

[0077] The weight of each edge is set according to the distance between the vehicle's current position and the corresponding task starting point;

[0078] Based on the weight of each edge, each job task to be assigned and the information of schedulable vehicles are matched, and the corresponding job task and vehicle information are marked with a status, such as "assigned status", based on the matching results.

[0079] In this embodiment, if a vehicle has the capability to perform a certain task, an edge is drawn between the vertex representing the task and the vehicle, and each edge is assigned a corresponding weight. Through this setting, the goal of reducing empty distance in port vehicle scheduling is transformed into a minimum weight maximum matching problem in a bipartite graph. By using a bipartite graph matching strategy, the complex port vehicle scheduling problem is transformed into a solvable mathematical model. Through the setting of edge weights and the matching algorithm, the optimal or relatively optimal task allocation scheme can be found quickly, effectively reducing the empty distance of vehicles, increasing the load rate, thereby improving the transportation efficiency of port vehicles and reducing resource waste.

[0080] In this embodiment, before matching the acquired task and vehicle information, the following steps are also included:

[0081] The acquired task and vehicle information are read and segmented into a multi-layered topic tree. Each layer is divided into multiple clusters according to task type and vehicle attributes. For example, according to cargo type, task information is divided into clusters such as general cargo and dangerous goods; according to vehicle purpose, vehicle information is divided into clusters such as container transport vehicles and bulk cargo transport vehicles.

[0082] Clustering is performed on the data information corresponding to each layer. Task characteristics such as weight and destination of the task are analyzed, as well as attribute information such as cargo capacity and driving area restrictions of the vehicle. Several cluster sets are obtained, and each cluster set is assigned to the corresponding cluster.

[0083] The task descriptions and vehicle information descriptions contained in the cluster set of each cluster are segmented into words, and then cleaned and deduplicated to remove some words that have no practical meaning or do not conform to the port business scenario, so as to obtain a keyword set and perform word vector conversion.

[0084] Based on the word vectors representing the core features of the cluster as defined by port business standards, the keyword corresponding to the word vector with the smallest distance is selected from the keyword set and used as the target keyword for that cluster.

[0085] Based on the target keywords, the key information corresponding to the task and vehicle information is marked with keywords to determine the key features of the task and vehicle information.

[0086] In this embodiment, key features include the type of cargo to be transported (whether it is high-value cargo, dangerous goods, etc.), the urgency of the task (delivery time requirements), the special operational requirements corresponding to the type of vehicle required for the task (container truck, flatbed truck, etc.), as well as the vehicle type (cargo capacity, volume), vehicle speed, vehicle maintenance status (whether it has been recently maintained, frequency of failures), driver's working hours and work experience (types of cargo to be transported, ability to handle complex road conditions), etc.

[0087] In this embodiment, the constraint processing module optimizes the task allocation results, specifically as follows:

[0088] Based on the key characteristics of the task, the importance, urgency, and type of goods involved in the task are determined, and a priority score is assigned to the task. The tasks are then ranked from high to low based on the priority score. For example, VIP tasks involving important clients, urgent tasks with strict delivery time limits, and tasks transporting dangerous goods are given higher priority; while ordinary cargo transportation tasks with ample delivery time are given relatively lower priority.

[0089] Based on the key features of vehicle information, the vehicle's performance is evaluated to determine the vehicle's comprehensive performance coefficient. At the same time, the driver's working hours and work experience corresponding to the vehicle information are extracted to evaluate the driver's work efficiency coefficient.

[0090] The vehicle's overall performance coefficient is compared with the preset vehicle performance reference value to obtain the vehicle performance ratio. The driver's work efficiency coefficient is compared with the preset driver efficiency reference value to obtain the corresponding driver efficiency ratio.

[0091] Vehicle weights are adjusted based on vehicle performance ratios and corresponding driver efficiency ratios. If both vehicle performance ratios and driver efficiency ratios are high, a higher overall fit weight is given to the vehicle. If one ratio is low and the other is high, the overall fit weight is determined by weighted summation based on the port's actual operational needs and priorities. Combined with the priority of the task, vehicles with higher weights that meet the vehicle and cargo types required for the task are given priority. For example, high-priority dangerous goods transportation tasks are given priority to vehicles with dangerous goods transportation qualifications, good overall performance, and experienced drivers. For ordinary priority tasks, vehicles are allocated sequentially according to the overall fit weight and task requirements, while ensuring that high-priority tasks are completed.

[0092] In this embodiment, by analyzing the key characteristics of the work tasks, the importance, urgency, and cargo type of the tasks are determined, thereby assigning priority scores to the work tasks and ranking them according to the scores. At the same time, the comprehensive performance coefficient of the vehicles and the work efficiency coefficient of the drivers are evaluated, and the vehicle performance ratio and driver efficiency ratio are compared to adjust the vehicle weights. Vehicles with high weights that meet the task requirements are given priority in allocation, ensuring that important and urgent tasks are handled first, optimizing the allocation of vehicle resources, improving the execution efficiency and accuracy of work tasks, reducing operating costs, and improving the overall efficiency and safety of port operations.

[0093] In this embodiment, the multi-objective optimization module obtains the overload rate and dispatch fairness evaluation indicators, specifically including:

[0094] During the task allocation process, the task load of the assigned vehicles is continuously monitored to obtain the ratio of the actual weight of the cargo loaded on the vehicle to the maximum load capacity of the vehicle, and the heavy load rate of each vehicle is calculated.

[0095] Statistical analysis of the load factor of all vehicles is performed to obtain overall vehicle load factor data. For example, the average, minimum and maximum load factors of all vehicles operating in the port are statistically analyzed within a certain period of time, so as to comprehensively evaluate the load situation under the current task allocation.

[0096] The number of tasks received by each vehicle within a certain time period is counted, and the variance of the number of tasks for different vehicles is calculated to assess the fairness of dispatching. The larger the variance, the greater the difference in the number of tasks between vehicles, and the worse the fairness of dispatching. The smaller the variance, the better the fairness of dispatching. For example, the variance of the number of tasks received by each vehicle in the past hour is calculated as an evaluation indicator of the current fairness of dispatching.

[0097] In this embodiment, the multi-objective optimization module acquires data in real time during task execution, updates the overload rate and dispatch fairness evaluation indicators, and adjusts the weight factors according to the evaluation indicators to provide a more accurate basis for the next task allocation. As tasks continue to proceed and data continues to accumulate, the multi-objective optimization module continuously iterates and optimizes the task allocation scheme, and further adjusts the weight factors and task allocation strategies.

[0098] In this embodiment, the obtained task and vehicle information are constructed using linear programming, and decision variables and constraints are set to obtain the optimal task and vehicle matching scheme.

[0099] The decision variable is each edge between the vertices of the task and vehicle information. ,(in The vertex representing the task. Each vertex representing vehicle information is assigned a variable. When the task Assigned to a vehicle hour, =1 (e.g., set to "assigned status"); otherwise, =0 (e.g., set to "unassigned state"), and construct an objective function based on this. Based on the objective function, extract the matching scheme with the minimum total unloaded driving distance of the vehicle from all task assignment results;

[0100] The constraints include any task in the vertex set of the job tasks. , must meet This ensures that each task must be assigned to one and only one vehicle; for any vehicle in the vertex set of vehicle information... , must meet This ensures that each vehicle can perform at most one task; when And when the task and vehicles When there is a connection, the corresponding variables will be... The expression input into the task allocation calculation is as follows: In the formula, Represented as decision variables, This is represented as the vertex of the task. The vertex representing vehicle information. Represented as a vehicle To the mission A represents the set of vertices for the task, B represents the set of vertices for vehicle information, and E represents the set of edges.

[0101] In this embodiment, a linear programming method is used to construct a task-vehicle matching scheme. By precisely setting decision variables, objective functions, and constraints, the optimal task allocation scheme is solved, which effectively reduces the total empty driving distance of vehicles and increases the vehicle load rate. This provides a scientific and precise method for port vehicle scheduling, further improving port operation efficiency and resource utilization.

[0102] Furthermore, the predictive scheduling module also includes monitoring and early warning of vehicle scheduling routes during port transportation, specifically including:

[0103] Based on the vehicle's dispatching origin and transportation priority, and combined with the road condition information of the road network system, the initial dispatching route of the vehicle is determined;

[0104] Real-time acquisition of traffic flow, weather conditions, accident and road maintenance information under the initial scheduling path;

[0105] The congestion value of the initial dispatch route is determined based on the traffic flow, the driving difficulty value of the initial dispatch route is determined based on the weather conditions, and the probability value of emergencies of the initial dispatch route is determined based on accident and road maintenance information.

[0106] Based on the congestion value, driving difficulty value, and probability value of emergencies, a comprehensive evaluation of the initial scheduling path is determined to obtain a path evaluation value. It is then determined whether the path evaluation value is greater than a first preset evaluation value. If so, the initial scheduling path is determined to be the optimal scheduling path; otherwise, it is determined whether the path evaluation value is greater than a second preset evaluation value.

[0107] If so, determine the initial scheduling path as a candidate scheduling path;

[0108] Otherwise, the initial scheduling path is determined to be an unqualified scheduling path;

[0109] When the initial scheduling path is the optimal scheduling path, the road condition characteristics and surrounding environment characteristics of the optimal scheduling path are obtained, and abnormal alarm conditions are formulated based on the road condition characteristics and surrounding environment characteristics.

[0110] When the initial scheduling path is a candidate scheduling path, other candidate scheduling paths are obtained based on the comprehensive evaluation value of the candidate scheduling path determined by the road network system within a preset difference. The candidate scheduling paths and other candidate scheduling paths are integrated to obtain a comprehensive scheduling path. When the initial scheduling path is an unqualified scheduling path, other scheduling paths with a value greater than a second preset evaluation value are determined based on the road network system. The other scheduling paths are integrated to obtain a comprehensive scheduling path.

[0111] Based on the road condition characteristics and surrounding environment characteristics of the comprehensive scheduling path, abnormal alarm conditions are formulated.

[0112] In this embodiment, the initial scheduling path of the vehicle during port operations is the optimal path determined only based on basic information such as path information, such as distance and road width.

[0113] In this embodiment, the higher the congestion value, the greater the driving difficulty value, the greater the probability of unexpected situations, and the lower the corresponding path evaluation value.

[0114] In this embodiment, the first preset evaluation value is greater than the second preset evaluation value. Paths greater than the first preset evaluation value can be considered excellent, while paths greater than the second preset evaluation value and less than the first preset evaluation value can be considered qualified.

[0115] In this embodiment, the comprehensive scheduling path includes the path information traversed by all scheduling paths.

[0116] In this embodiment, the abnormal alarm conditions include that the vehicle's position cannot deviate from the set path, and that the vehicle's dwell time cannot exceed the set time.

[0117] The beneficial effects of the above design scheme are: by evaluating the route based on traffic flow, weather conditions, accidents and road maintenance information of the port route, the final executable scheduling route is determined, and abnormal alarm conditions are formulated based on two indicators: vehicle location and vehicle dwell time, ensuring the accuracy of the formulated abnormal alarm conditions and providing an accurate pre-judgment mechanism for the reliability of vehicle operations in the port.

[0118] Furthermore, based on the road condition characteristics and surrounding environment characteristics of the comprehensive scheduling path, abnormal alarm conditions are formulated, including:

[0119] Based on the road network system, important nodes are marked in the comprehensive scheduling path to obtain important road condition nodes and important environmental nodes;

[0120] Based on the road conditions at key road nodes, the maximum dwell time at those key road nodes is calculated using the following formula; in, This indicates the maximum dwell time at the key road condition nodes. Indicates the preset minimum stay time. This represents the second preset evaluation value. This represents the path evaluation value at the critical node of the road condition. This indicates the number of other important road condition nodes adjacent to the aforementioned important road condition node. This represents the path evaluation value for the i-th other important road condition node;

[0121] Based on the environment of important environmental nodes, the maximum stay time at the important environmental nodes is calculated according to the following formula; in, This indicates the maximum dwell time at the key nodes in the environment. represents the natural constant, with a value of 2.72. This indicates the probability of an operational anomaly occurring under standard conditions. This indicates the probability of an operational anomaly occurring at a critical node in the environment.

[0122] Based on the route of the comprehensive scheduling path, location anomaly conditions are established. Based on the maximum dwell time of important road condition nodes and the maximum dwell time of important environmental nodes, dwell time anomaly conditions are established. Based on the location anomaly conditions and dwell time anomaly conditions, anomaly alarm conditions are formulated.

[0123] In this embodiment, the evaluation value ranges from (0, 1).

[0124] In this embodiment, the path evaluation value of the important road condition nodes in the port is the evaluation value of the path of a preset length passing through the important road condition nodes. The evaluation method is the same as the evaluation method of the overall scheduling path, and the path evaluation value of the important road condition nodes is greater than the second preset evaluation value.

[0125] In this embodiment, the probability of operational anomalies occurring at critical environmental nodes is related to environmental conditions. For example, the probability of operational anomalies increases if there are many people nearby. Operational anomalies include traffic jams, collisions, or car accidents.

[0126] The beneficial effects of the above design scheme are as follows: by marking important nodes in the comprehensive scheduling path based on the road network system, important road condition nodes and important environmental nodes are obtained, the maximum dwell time at important road condition nodes and the maximum dwell time at important environmental nodes are determined, location anomaly conditions are established based on the route of the comprehensive scheduling path, dwell time anomaly conditions are established based on the maximum dwell time at important road condition nodes and the maximum dwell time at important environmental nodes, and anomaly alarm conditions are formulated based on location anomaly conditions and dwell time anomaly conditions, ensuring the accuracy and comprehensiveness of the formulated anomaly alarm conditions, and providing an accurate judgment mechanism for identifying vehicle operation anomalies.

[0127] In this embodiment, the multi-objective optimization module further includes:

[0128] Based on the results, the historical driving data of the driver corresponding to the vehicle information is extracted. Based on the weight value of the driver's historical driving data in the total historical driving data, the weight value is determined according to factors such as the driver's participation in port operations and the amount of tasks completed. The driver's driving experience value under the priority of the current task is analyzed. If the driver has rich experience in transporting heavy cargo, his driving experience value under heavy load tasks is higher.

[0129] Determine the driver's maximum continuous driving time based on historical driving data;

[0130] Input the vehicle’s maximum speed (which can be obtained from the vehicle information) and all historical speeds from the historical speed list of historical driving data into the preset data axis to obtain the speed difference between the maximum speed and each historical speed.

[0131] The maximum driving speed is adjusted based on the average value of the speed difference to generate the adjustment speed; for example, if the historical driving speed is generally low, it indicates that the driver is accustomed to conservative driving, and the maximum driving speed can be appropriately reduced to obtain the adjustment speed.

[0132] The recommended driving speed range is generated based on the adjustment speed and driving experience value. Experienced drivers can appropriately widen the speed range, while less experienced drivers can narrow it. At the same time, the maximum continuous driving time is adjusted based on driving experience value to generate a recommended driving time. Experienced drivers can appropriately extend the continuous driving time, but must still comply with safety regulations.

[0133] Based on the recommended driving speed range and recommended driving time, a driving plan for the task is generated, and the driver's driving behavior is analyzed in real time during the task to determine whether it conforms to the driving plan.

[0134] If the conditions are not met, extract any abnormal items from the driver, such as speeding or excessive driving time, and issue corresponding alarms. For example, issue a speed alarm for speeding and a fatigue driving alarm for excessive driving time to ensure the safety of port operation vehicles.

[0135] In this embodiment, by comprehensively considering factors such as driving experience and historical driving data to generate a reasonable driving plan and monitoring driving behavior in real time, it not only helps to improve the safety of vehicle driving and reduce the risk of accidents, but also allows for the reasonable arrangement of tasks according to the actual situation of the driver, improving the driver's work efficiency and comfort, further improving the function of the port vehicle dispatching system, and ensuring the safe and stable operation of port operations.

[0136] This invention provides another technical solution: a smart port operation vehicle dispatching robot, comprising: configuring a port operation vehicle dispatching system in the operation vehicle dispatching robot; the operation vehicle dispatching robot acquires heavy load rate and dispatch fairness evaluation indicators, performs functions such as task and vehicle matching, handles complex task constraints, achieves multi-objective optimization, and makes dispatching decisions based on prediction results; and establishes a data interaction channel with vehicles in the port operation scenario to intelligently dispatch vehicles, collect data in real time during the dispatching process through the data interaction channel, and continuously optimizes its own model using online learning technology to improve dispatching performance and adapt to different port operation conditions.

[0137] In this embodiment, the smart port operation vehicle dispatching robot integrates the core functions of the dispatching system. It collects data in real time through the data interaction channel with the vehicles and continuously optimizes its own model by combining online learning technology. This enables the robot to adapt to changes in the port operation environment and achieve intelligent and efficient dispatching. It not only improves the accuracy and efficiency of dispatching but also reduces labor costs, promotes the development of port operations towards intelligence and automation, and enhances the overall competitiveness of the port.

[0138] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A smart port operations vehicle dispatching system, characterized in that, Comprise: The task matching module is configured to match the obtained job tasks and vehicle information based on a bipartite graph matching strategy, and obtain a task allocation result; The bipartite graph matching strategy specifically comprises: Convert the port vehicle scheduling task into a bipartite graph matching problem and build a bipartite graph framework; Each to-be-allocated job task is taken as a vertex set of the bipartite graph framework, and the schedulable vehicle information is taken as another vertex set of the bipartite graph framework; Between the vertex set of the job tasks and the vertex set of the vehicle information, edges are constructed according to the potential association relationship between the job tasks and the vehicle information, and the weight of each edge is determined, wherein any two edges do not have a common vertex; Wherein, the weight of each edge is set according to the distance between the current position of the vehicle and the starting point of the corresponding task; Based on the weight of each edge, each to-be-allocated job task and schedulable vehicle information are matched, and the corresponding job task and vehicle information are marked based on the matching result; The constraint processing module is configured to allocate priorities to the job tasks, adjust the weight of the corresponding vehicle information based on a weight factor, and optimize the task allocation result based on the bipartite graph matching strategy based on the task priority and the vehicle information weight; The multi-objective optimization module is configured to obtain a heavy load rate and a dispatch fairness evaluation index based on the optimized task allocation result, and adjust the weight factor based on the heavy load rate and the dispatch fairness evaluation index; The predictive scheduling module is configured to obtain historical data of port vehicle scheduling, predict the vehicle position and future task arrival based on the historical data, integrate the prediction result with the current task allocation result to generate a scheduling strategy, and monitor and warn the scheduling driving path of the vehicle during port transportation.

2. The smart port vehicle dispatching system of claim 1, wherein, Before matching the obtained job tasks and vehicle information, it further comprises: Reading the obtained job tasks and vehicle information, and dividing the job tasks and vehicle information into a multi-layer topic tree, each layer being divided into multiple clusters according to the task type and vehicle attribute; Cluster processing of the data information corresponding to each layer, analyzing the task characteristics of the job tasks and the attribute information of the vehicles, obtaining a plurality of cluster sets, and assigning each cluster set to a corresponding cluster; Carrying out word segmentation operation on the job task description sentences and vehicle information description sentences contained in each cluster set, and carrying out cleaning and deduplication processing to obtain a keyword set and perform word vector conversion; According to the word vector representing the core features of the cluster set by the port business standard, the keyword corresponding to the minimum distance word vector is selected from the keyword set as the target keyword in the cluster; Based on the target keyword, the key information corresponding to the job task and the vehicle information is marked with the keyword, and the key features of the job task and the vehicle information are determined.

3. A smart port vehicle dispatching system as claimed in claim 2, wherein, The constraint processing module optimizes the task allocation result, specifically: Determine the importance, urgency and cargo type involved in the job task based on the key features of the job task, determine the priority score of the job task, and arrange the job tasks from high to low based on the priority score; The vehicle is evaluated according to the key features of the vehicle information, and the comprehensive performance coefficient of the vehicle is determined, and the working time and working experience of the driver corresponding to the vehicle information are extracted, and the working efficiency coefficient of the driver is evaluated; The comprehensive performance coefficient of the vehicle is compared with the preset vehicle performance reference value to obtain the vehicle performance ratio, and the working efficiency coefficient of the driver is compared with the preset driver efficiency reference value to obtain the corresponding driver efficiency ratio; Based on the vehicle performance ratio and the corresponding driver efficiency ratio, the vehicle weight is adjusted, the priority of the operation task is combined, and the vehicle with high vehicle weight and meeting the required vehicle type and cargo type of the task is preferentially allocated.

4. A smart port vehicle dispatching system as claimed in claim 3, wherein, The multi-objective optimization module obtains the overload rate and the order distribution fairness evaluation index, specifically including: During the task allocation process, the task load of the allocated vehicle is continuously monitored, the ratio of the actual loaded cargo weight of the vehicle to the maximum carrying weight of the vehicle is obtained, and the overload rate of each vehicle is calculated; The overload rates of all vehicles are statistically analyzed to obtain the overall vehicle overload rate data; The number of tasks received by each vehicle within a certain time period is counted, and the variance of the number of tasks of different vehicles is calculated to evaluate the order distribution fairness. The larger the variance, the greater the difference in the number of tasks between vehicles, and the worse the order distribution fairness. The smaller the variance, the better the order distribution fairness.

5. A smart port vehicle dispatching system as claimed in claim 4, wherein, The multi-objective optimization module further includes: Based on the result, the historical driving data of the driver corresponding to the vehicle information is extracted, and the driving experience value of the driver under the priority of the current operation task is analyzed based on the weight value of the historical driving data of the driver in the total historical driving data; The maximum continuous driving time of the driver is determined based on the historical driving data; The maximum driving speed of the vehicle and all historical speeds in the historical speed list of the historical driving data are input into a preset data axis to obtain the speed difference corresponding to each historical speed; The maximum driving speed is adjusted based on the average value of the speed difference to generate an adjusted speed; The recommended driving speed range is generated according to the adjusted speed combined with the driving experience value, and the recommended driving time is generated by adjusting the maximum continuous driving time based on the driving experience value; Based on the recommended driving speed range and the recommended driving time, a driving plan for the operation task is generated, and whether the driver's driving behavior conforms to the driving plan is analyzed in real time during the execution of the operation task by the driver; If not, the abnormal items of the driver are extracted and corresponding alarm work is performed.

6. A smart port vehicle dispatching system as claimed in claim 5, wherein, The prediction scheduling module further includes: During the task allocation process, if it is predicted that a large number of operation tasks will appear in the target area and the number of vehicles in the area is insufficient, then the vehicles in non-target areas are guided to the target area for standby; At the same time, according to the prediction result, the task type and vehicle state of the corresponding operation task are determined, and the corresponding vehicle is matched to execute the corresponding operation task; Real-time feedback data is obtained, the actual scheduling situation monitored is compared with the prediction result, and the prediction result is dynamically adjusted based on the comparison result.

7. A smart port vehicle dispatching system as claimed in claim 1, wherein, The prediction scheduling module monitors and warns the scheduling driving path of the vehicle in the port transportation process, specifically including: The vehicle-based scheduling system determines an initial scheduling path of the vehicle based on a transportation starting point and a transportation ending point of the vehicle and in combination with road condition information of a road network system; Real-time traffic flow, weather conditions, accident and road maintenance information on the initial scheduling path are obtained; A congestion value of the initial scheduling path is determined based on the traffic flow, a driving difficulty value of the initial scheduling path is determined based on the weather conditions, and a sudden situation probability value of the initial scheduling path is determined based on the accident and road maintenance information; A comprehensive evaluation of the initial scheduling path is determined based on the congestion value, the driving difficulty value and the sudden situation probability value to obtain a path evaluation value, and it is determined whether the path evaluation value is greater than a first preset evaluation value; If yes, the initial scheduling path is determined as an optimal scheduling path, otherwise, it is determined whether the path evaluation value is greater than a second preset evaluation value; If yes, the initial scheduling path is determined as a candidate scheduling path; Otherwise, the initial scheduling path is determined as an unqualified scheduling path; When the initial scheduling path is the optimal scheduling path, road condition features and surrounding environment features of the optimal scheduling path are obtained, and an abnormal alarm condition is formulated based on the road condition features and the surrounding environment features; When the initial scheduling path is the candidate scheduling path, other candidate scheduling paths are obtained based on the road network system and the comprehensive evaluation value of the candidate scheduling path is within a preset difference, the candidate scheduling path and the other candidate scheduling paths are integrated to obtain a comprehensive scheduling path, and when the initial scheduling path is the unqualified scheduling path, other scheduling paths greater than the second preset evaluation value are determined based on the road network system, and the other scheduling paths are integrated to obtain the comprehensive scheduling path; 8. A smart port vehicle dispatching system as claimed in claim 7, wherein, An abnormal alarm condition is formulated based on road condition features and surrounding environment features of the comprehensive scheduling path. An abnormal alarm condition is formulated based on road condition features and surrounding environment features of the comprehensive scheduling path, including: Important nodes in the comprehensive scheduling path are marked based on the road network system to obtain road condition important nodes and environment important nodes; The maximum stay duration at the road condition important nodes is calculated based on the road condition of the road condition important nodes; The maximum stay duration at the environment important nodes is calculated based on the environment of the environment important nodes; 9. A smart port operation vehicle scheduling robot, applying the smart port operation vehicle scheduling system according to claim 1, characterized in that, Positioning abnormal conditions are established based on the route of the comprehensive scheduling path, stay time abnormal conditions are established based on the maximum stay duration of the road condition important nodes and the maximum stay duration of the environment important nodes, and the abnormal alarm condition is formulated based on the positioning abnormal conditions and the stay time abnormal conditions. Including: A port work vehicle scheduling system is configured in a work vehicle scheduling robot, the work vehicle scheduling robot runs to obtain a heavy load rate and a single dispatch fairness evaluation index, performs functions of task and vehicle matching, complex task constraint processing, multi-objective optimization and scheduling decision based on prediction results, and constructs a data interaction channel with the vehicle in a port work scene, intelligently schedules the vehicle, and collects data in the scheduling process in real time through the data interaction channel.

Images