A transit logistics dynamic path optimization scheduling method

Abstract

The application provides a kind of transfer logistics dynamic path optimization scheduling method, it is related to logistics scheduling technical field, by dynamic collection and analysis to handling equipment position, cargo distribution time window etc. Data, combined with multi-level ant colony path algorithm and equipment conflict sensing mechanism, realize the dynamic update of path information element intensity and the real-time adjustment of path selection probability, to effectively optimize path scheduling efficiency. In this way, the risk of channel congestion and the probability of equipment concurrent conflict can be significantly reduced, the utilization rate of path resources and the rationality of task allocation can be improved, which helps to improve the dynamic response capability and overall operation efficiency of transfer station logistics operation. At the same time, the application fully considers the matching relationship between distribution time window and equipment idle state in task scheduling, which can improve the reliability of timely delivery of goods and reduce the delay of sorting and transfer.

Description

Technical Field

[0001] This invention relates to the field of logistics scheduling technology, and more specifically, to a dynamic path optimization scheduling method for transit logistics. Background Technology

[0002] With the rapid development of global economic integration and the e-commerce industry, the importance of modern logistics systems has become increasingly prominent. As a key node in the logistics system, transit logistics hubs undertake the core functions of cargo collection, sorting, and transshipment. The efficiency of transit logistics stations not only affects the operating speed of the entire logistics chain but also has a profound impact on logistics costs and customer experience. Currently, modern transit logistics is gradually developing towards intelligence, including the widespread application of unmanned transport equipment (AGVs, Automated Guided Vehicles) and the introduction of various optimization algorithms in path planning, which has significantly improved the operational efficiency of transit stations. However, many logistics stations still face problems such as equipment concurrency conflicts, unreasonable path selection, and insufficient dynamic scheduling response capabilities. Dynamic path optimization scheduling technology, which adapts to complex and ever-changing order arrival patterns and cargo delivery needs, has become a key research direction for improving the efficiency of transit logistics hubs.

[0003] Existing technologies typically employ fixed-path scheduling methods or basic optimization strategies based on classical pathfinding algorithms (such as shortest path algorithms and genetic algorithms) for equipment task allocation. However, fixed-path scheduling cannot adapt to the dynamic allocation needs of goods, resulting in low path resource utilization and increased channel congestion risks. While classical pathfinding algorithms can achieve a certain degree of dynamic scheduling, they do not fully consider the real-time status of path resources and the issue of concurrent equipment conflicts, leading to decreased efficiency in resource scheduling and task allocation. Furthermore, when facing goods with stringent delivery time windows, existing methods often neglect the coordination between the remaining processing time of goods and the idle time of equipment, resulting in severe delays for some goods. In addition, the lack of initial pheromone settings, path conflict monitoring, and congestion response mechanisms renders existing transit logistics hubs inadequate in terms of dynamic scheduling performance in complex order scenarios, making it difficult to meet the demands for efficient and accurate sorting and transfer. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention is proposed. This invention provides a dynamic path optimization scheduling method for transit logistics, which to some extent solves the problem that existing scheduling methods cannot handle dynamic conflicts between multiple handling paths within a transit station and the real-time allocation of equipment resources. This leads to the inability to promptly replan cargo flow and adjust equipment scheduling sequences when congestion occurs in key channels, resulting in reduced overall operational efficiency and cargo backlog.

[0005] According to one aspect of the present invention, a method for dynamic path optimization scheduling of transit logistics is provided, comprising: Receive the location coordinates of each handling equipment in the transfer station and the cargo delivery time window data, set the initial value of pheromone according to the idle time of the equipment and the remaining processing time of the cargo, and generate a path pheromone intensity matrix; Based on the path pheromone intensity matrix, multi-level ant colony pathfinding calculation is performed to construct a device conflict perception mechanism to monitor the concurrent usage status of each path segment in real time and output a path selection probability distribution containing conflict weights. The path selection probability distribution is dynamically adjusted in accordance with the real-time channel congestion status, and the attractiveness weight of each path is updated using a pheromone decay and enhancement mechanism to form an adaptive path preference sequence. Based on the adaptive path preference sequence, the device task allocation is performed, and the handling tasks are assigned to the corresponding devices according to the path preference degree. The pheromone intensity distribution is updated synchronously to complete the dynamic path optimization scheduling within the transfer station.

[0006] Furthermore, the initial pheromone value is set by matching the idle time of the equipment and the remaining processing time of the goods to obtain a time matching coefficient. The pheromone intensity adjustment factor is determined based on the deviation between the time matching coefficient and the average matching coefficient of all equipment and goods pairings. Finally, the pheromone intensity adjustment factor is used to correct the basic pheromone intensity to obtain the initial pheromone value of the path segment.

[0007] Furthermore, when the device conflict perception mechanism detects that multiple virtual ants are simultaneously selecting the same path segment, it immediately triggers a conflict marker update operation, calculates the conflict weight value based on the number of virtual ants in the conflict path segment, calculates the path attractiveness score based on the conflict weight value, and then normalizes it to obtain the path selection probability distribution. The device conflict perception mechanism establishes a status monitoring table for each path segment to record the virtual ant identifiers in use. When multiple ants enter the same path segment at the same time, a concurrent conflict is detected and the path segment status is updated. Real-time monitoring of concurrent usage status is achieved by continuously scanning the monitoring table to count the number of occupants in each path segment.

[0008] Furthermore, the conflict weight value is obtained by comprehensively evaluating the conflict intensity coefficient based on the comparison between the number of virtual ants in the conflict path segment and the carrying capacity of the path segment, as well as the comparison between the conflict duration and the standard processing time. Then, the conflict intensity coefficient is converted into a conflict weight value through an inverse proportional conversion.

[0009] Furthermore, the path selection probability value in the path selection probability distribution is calculated as follows: in, Let be the combined selection probability value from device i to region j. The initial pheromone value for the path from device i to region j is used as the pheromone concentration factor. Let be the actual length of the path segment from device i to region j. This represents the real-time conflict weight values ​​for the path segment from device i to region j. This represents the positive promoting weighting coefficient of the pheromone concentration factor. Optimize the weighting coefficients for the path length factor based on distance. The congestion penalty weight coefficient is the conflict weight factor. For conflict sensitivity adjustment parameters, This represents the total number of all possible path segments. This is the numerical stability protection constant. Let be a natural constant, and k be all possible regions. This iterates through all possible target regions starting from the current device i. This represents the real-time conflict weight value of the path segment from device i to region k.

[0010] Furthermore, the path selection probability distribution is dynamically adjusted in relation to the real-time channel congestion. The deviation between the path selection probability and the actual congestion level of the channel is calculated to determine the probability adjustment range. The probability adjustment range is then used to directly attenuate or enhance the pheromone concentration, thereby achieving dynamic matching between the path attraction weight and the current channel condition.

[0011] Furthermore, the probability adjustment amplitude is calculated as follows: in, Let $\frac{i}{j}$ be the path probability adjustment value from device $i$ to region $j$. The sign function determines the adjustment direction: when the probability deviation is positive, it outputs a negative sign to decrease the selection probability; when the probability deviation is negative, it outputs a positive sign to increase the selection probability. The probability deviation value is equal to the selection probability value minus the current channel usage value. The maximum number of devices that can be carried by this path channel is determined by reading the channel configuration database. This refers to the remaining throughput capacity of the channel for this path, obtained from channel monitoring data. This is the probability bias response intensity coefficient. For probability bias standardization parameters, Adjust the basic weighting coefficient for capacity. For capacity response sensitivity parameters, It is a natural constant.

[0012] According to another aspect of the present invention, a dynamic route optimization and scheduling system for transit logistics is provided, comprising: The equipment conflict perception module is used to monitor the equipment occupancy status of each path segment in the transfer station in real time, detect path competition conflicts between virtual ants and calculate the conflict weight value. The comprehensive probability calculation module is used to integrate information from three dimensions: pheromone concentration, path length factor, and real-time conflict weight, to calculate the comprehensive selection probability distribution of each path segment. The probability adjustment range calculation module is used to compare the matching degree between the path selection probability and the actual congestion status of the channel, and calculate the probability adjustment range value required for each path. The pheromone dynamic update module is used to attenuate or enhance the pheromone concentration of each path according to the probability adjustment amplitude, forming an adaptive path preference sequence. The intelligent task allocation module is used to perform device task allocation based on the adaptive path preference sequence, synchronously update the pheromone intensity distribution, and complete dynamic path optimization scheduling.

[0013] Compared with existing technologies, the dynamic path optimization scheduling method for transit logistics provided by this invention dynamically collects and analyzes data such as the location of handling equipment and the delivery time window of goods. Combined with a multi-level ant colony pathfinding algorithm and an equipment conflict perception mechanism, it achieves dynamic updates of path pheromone intensity and real-time adjustments to path selection probabilities, thereby effectively optimizing path scheduling efficiency. This significantly reduces the risk of channel congestion and the probability of concurrent equipment conflicts, improves path resource utilization and task allocation rationality, and helps enhance the dynamic response capability and overall operational efficiency of transit station logistics operations. Simultaneously, this invention fully considers the matching relationship between delivery time windows and equipment idle status in task scheduling, which can improve the reliability of timely delivery of goods and reduce sorting and transfer delays. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a flowchart of a dynamic path optimization scheduling method for transit logistics according to an embodiment of the present invention.

[0015] Figure 2 This is a flowchart of the dynamic adjustment and adaptive path preference sequence generation process in the transit logistics dynamic path optimization scheduling method according to an embodiment of the present invention. Detailed Implementation

[0016] Hereinafter, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments of the present invention. It should be understood that the present invention is not limited to the exemplary embodiments described herein.

[0017] Figure 1 This is a flowchart of a dynamic path optimization scheduling method for transit logistics according to an embodiment of the present invention. Figure 1 As shown, the dynamic path optimization scheduling method for transit logistics includes: S1: Receives the location coordinates of each handling equipment in the transfer station and the cargo delivery time window data, sets the initial value of pheromone based on the idle time of the equipment and the remaining processing time of the cargo, and generates a path pheromone intensity matrix. First, the real-time location coordinates of all handling equipment are obtained from the equipment management system within the transfer station, including the precise positions of various handling equipment such as AGVs, conveyor belt nodes, and sorting robotic arms in the station's two-dimensional coordinate system. Simultaneously, delivery time window data for pending goods is obtained from the order management system, i.e., the time limit required for each batch of goods to complete its departure from the transfer station.

[0018] After obtaining the basic data, the idle time for each handling device from the completion of the current task to the next schedulable time point is calculated. This time reflects the availability of the equipment. For goods, the remaining processing time for each batch of goods is calculated based on the difference between the delivery deadline and the current time. This time reflects the urgency of the goods processing.

[0019] Furthermore, the idle time of each device is used as the numerator, and the remaining processing time of the corresponding goods is used as the denominator. A time matching coefficient is obtained by dividing the two. All time matching coefficients for device-goods pairings are collected, and their average value is calculated as a reference. When the time matching coefficient of a pairing is higher than the average, a pheromone intensity value higher than the baseline level is assigned to the path segment connecting that device. If the time matching coefficient equals the average, a pheromone intensity value at the baseline level is assigned to the corresponding path segment. When the time matching coefficient is lower than the average, a pheromone intensity value lower than the baseline level is assigned to the relevant path segment. By quantifying the deviation of each time matching coefficient from the average, a corresponding pheromone intensity adjustment factor is obtained. The baseline pheromone intensity value is multiplied by the adjustment factor to obtain the final initial path pheromone value. All path segments are traversed, and the calculated initial pheromone values ​​are filled into the corresponding positions of the path pheromone intensity matrix according to the starting device number as the row index and the ending region number as the column index. More specifically, the calculation of the initial path pheromone value is shown in the following formula: in, in, Let be the initial path pheromone value from device i to region j, and it is a dimensionless value. The basic pheromone intensity coefficient is set to 1.0. The idle time of device i is in minutes. The remaining processing time for cargo j is in minutes. The time smoothing parameter is set to 0.1 minutes to prevent the denominator from being zero. The average time matching coefficient for all equipment and cargo pairings is dimensionless. The standard deviation of the time matching coefficient is dimensionless. The numerical stability constant is set to 0.01. The parameter for the hyperbolic tangent function is set to 2.0, which is dimensionless. The parameter for the sigmoid function is adjusted to a value of 3.0, which is dimensionless. This represents the total number of equipment and goods paired together. It is the hyperbolic tangent function. It is an exponential function.

[0020] Finally, all possible transport paths are divided into path segments according to the starting equipment, intermediate nodes, and ending areas. The initial pheromone values ​​of each path segment are stored and organized in matrix form to form a complete path pheromone intensity matrix.

[0021] S2: Perform multi-level ant colony pathfinding calculation based on the path pheromone intensity matrix. By constructing a device conflict perception mechanism, monitor the concurrent usage status of each path segment in real time. When multiple virtual ants are detected to select the same path segment at the same time, trigger the conflict mark update and output the path selection probability distribution containing conflict weights. When initiating multi-level ant colony pathfinding computation, the virtual ant colony is divided into three levels according to the urgency of the goods: high-priority, medium-priority, and normal-priority. The high-priority layer handles urgent goods tasks, the medium-priority layer handles regular goods tasks, and the normal-priority layer handles non-urgent goods tasks. Virtual ants at each level perform path exploration tasks in parallel. Each level is allocated an independent computation time slice. High-priority ants have priority in obtaining path selection rights. After a high-priority layer completes one round of path selection, the medium-priority layer begins execution, and so on. Each virtual ant performs a preliminary path evaluation based on the pheromone concentration values ​​in the path pheromone intensity matrix, calculating the attractiveness index of each available path as the basis for path selection.

[0022] The attraction index is calculated by extracting the pheromone concentration value for each available path and simultaneously measuring the actual distance from the starting point to the destination area. The pheromone concentration value is directly used as the positive contribution component of the attraction, while the path distance is used as the negative contribution component. When processing the negative contribution component, it is converted to a positive value by taking the reciprocal of the distance; the longer the distance, the smaller the reciprocal, and vice versa. The pheromone concentration value is multiplied by the reciprocal of the distance to obtain the overall attraction value for that path. If a path has high pheromone concentration and short distance, the product results in a high attraction value; conversely, if the path has low pheromone concentration or long distance, the product results in a relatively low attraction value. The overall attraction values ​​of all available paths are collected, and the largest value is used as the normalization divisor. The overall attraction value of each path is divided by the largest value divisor to obtain a standardized attraction index ranging from zero to one, thus completing the quantitative calculation process for the attraction index of each available path.

[0023] Furthermore, an independent status monitoring table is established for each path segment within the transfer station. This table records the identifiers and entry timestamps of the virtual ants currently using that path segment. When a virtual ant selects a path segment, its unique identifier is immediately written into the corresponding path segment's status monitoring table, and the path segment is marked from an idle state to an occupied state. If other virtual ants subsequently attempt to enter an already occupied path segment, the status monitoring table is checked, and if a prior occupant is found, the identifier of the newly entering ant is appended to the waiting queue of the monitoring table, and the path segment's status is updated to a concurrent conflict state. By continuously scanning the status monitoring tables of all path segments, the number of occupying ants and the length of the waiting queue for each path segment are counted in real time. When a virtual ant finishes using a path segment and leaves, its identifier is deleted from the corresponding status monitoring table. If other ant identifiers remain in the monitoring table, the occupied state is maintained. When the monitoring table is completely cleared, the path segment's status is restored to an idle state. This dynamic update mechanism enables real-time perception and monitoring of the concurrent usage status of each path segment.

[0024] It should be noted that the equipment conflict awareness mechanism was built to address the blind spot problem in existing scheduling methods when handling path conflicts within transfer stations. Traditional scheduling methods can only make decisions based on static path information and cannot perceive the real-time conflict situation arising when multiple handling tasks compete for the same path segment simultaneously. This results in a lack of predictive ability for path congestion when allocating tasks. When congestion occurs in critical channels, existing methods cannot detect the problem in time due to the lack of real-time conflict monitoring, and can only respond passively when congestion has already formed and affected overall efficiency. The equipment conflict awareness mechanism, by monitoring the concurrent usage status of each path segment in real time, can immediately identify and record congestion information the moment a conflict occurs, providing dynamic conflict awareness. When multiple virtual ants are detected competing for the same path segment simultaneously, the mechanism immediately triggers a conflict marker update, converting the conflict information into a negative weight for path selection, thereby guiding subsequent handling tasks to automatically avoid congested areas. This proactive conflict awareness and response mechanism enables the replanning of cargo flow before congestion spreads, avoiding the time loss of passive waiting.

[0025] Furthermore, when the device's conflict detection mechanism detects multiple virtual ants simultaneously selecting the same path segment, it immediately triggers a conflict marker update operation. A conflict intensity coefficient is calculated based on the number of virtual ants on the conflicting path segment; the more conflicting virtual ants, the higher the conflict intensity coefficient. This conflict intensity coefficient is converted into a conflict weight value, which is then used as a penalty factor in calculating the selection probability of that path segment. A higher conflict weight value results in a lower probability of that path segment being selected in subsequent path selections, thus guiding other virtual ants to avoid congested path segments.

[0026] The process involves calculating the conflict intensity coefficient and converting it into a conflict weight value. This is achieved by statistically analyzing the total number of virtual ants currently using the conflict path segment and retrieving the maximum allowed number of ants for that path segment at a single moment from the path segment attribute database. The current load ratio of the path segment is calculated by dividing the actual number of ants used by the maximum allowed limit. Simultaneously, the start time of the conflict is recorded, and the duration of the conflict is obtained by subtracting the conflict start time from the current time. The standard completion time of a single transport task is obtained from the configuration as a time reference, and the time overrun multiple is obtained by dividing the conflict duration by the standard completion time. The conflict intensity coefficient of the path segment is obtained by multiplying the current load ratio by the time overrun multiple. When converting the conflict weight, a unit value is set as the dividend, and the conflict intensity coefficient is used as the divisor. A larger conflict intensity coefficient results in a smaller division result, and vice versa. This reciprocal relationship converts the conflict intensity coefficient into a conflict weight value, achieving an inverse correspondence between conflict severity and path selection weight.

[0027] It should be noted that the construction of conflict weight values ​​is to convert the real-time conflict information collected by the device conflict sensing mechanism into quantitative parameters that can directly influence path selection decisions. The device conflict sensing mechanism is responsible for monitoring and identifying path congestion, but its output conflict information is still descriptive data and cannot be directly integrated into the path selection probability calculation process of the ant colony algorithm. The conflict weight values ​​act as a bridge between the two, converting the conflict characteristic parameters such as congestion level and duration captured by the conflict sensing mechanism into standardized numerical weights, enabling conflict information to be comprehensively calculated within the same mathematical framework as traditional factors such as pheromone concentration and path distance.

[0028] Furthermore, considering the original pheromone concentration, path length, and real-time conflict weights, the overall selection probability of each path segment is recalculated. The conflict weight is used as a negative adjustment factor, weighted and fused with the pheromone concentration and distance factors to obtain a corrected path attractiveness value. The corrected attractiveness values ​​of all path segments are normalized to ensure the sum of probabilities meets standard requirements, forming a path selection probability distribution table that includes conflict weights. This provides a real-time updated probability basis for subsequent dynamic path adjustments.

[0029] The calculation of the comprehensive selection probability extracts the original pheromone concentration, the inverse of the path length, and the real-time conflict weight for each path segment as three independent influencing factors. Importance weight coefficients are assigned to these three factors: the pheromone concentration factor receives a positive promotion weight, the path length factor receives a distance optimization weight, and the conflict weight factor receives a congestion penalty weight. The pheromone concentration value is multiplied by its corresponding positive promotion weight to obtain the pheromone contribution component; the inverse of the path length is multiplied by the distance optimization weight to obtain the distance contribution component; and the conflict weight is multiplied by the congestion penalty weight to obtain the conflict adjustment component. If a path segment experiences real-time conflict, the conflict adjustment component has a negative impact on the overall attractiveness; when there is no conflict, the conflict adjustment component remains neutral and has no additional impact. The pheromone contribution component, distance contribution component, and conflict adjustment component are weighted and summed to obtain the comprehensive attractiveness score of the path segment. The comprehensive attractiveness scores of all path segments are collected, and the sum of these scores is used as the normalized denominator. The comprehensive attractiveness score of each path segment is then divided by the sum of the base scores to obtain the comprehensive selection probability value of that path segment at the current time. The calculation of the comprehensive selection probability value is shown in the following formula: in, Let be the combined selection probability value from device i to region j. The initial pheromone value for the path from device i to region j is used as the pheromone concentration factor. Let be the actual length of the path segment from device i to region j. This represents the real-time conflict weight values ​​for the path segment from device i to region j. The positive promoting weighting coefficient for the pheromone concentration factor is set to 1.5. The distance optimization weight coefficient for the path length factor is set to 2.0. The congestion penalty weight coefficient for the conflict weight factor is set to 1.0. The conflict sensitivity adjustment parameter is set to 3.0. This represents the total number of all possible path segments. The numerical stability protection constant is set to 0.01. Let be a natural constant, and k be all possible regions. This iterates through all possible target regions starting from the current device i. This represents the real-time conflict weight value of the path segment from device i to region k.

[0030] Preferably, considering the original pheromone concentration, path length, and real-time conflict weight when calculating the selection probability aims to construct a comprehensive path evaluation system and overcome the shortcomings of existing scheduling methods' single-dimensional decision-making. Original pheromone concentration reflects the quality of paths in historical experience, and path length reflects the basic efficiency characteristics of paths; however, neither of these traditional factors can reflect the real-time congestion situation. Existing scheduling methods, relying solely on this static information, lack adaptability in the face of dynamic conflicts. The introduction of real-time conflict weight fills the gap in dynamic information, enabling path selection decisions to simultaneously consider information from three dimensions: historical optimization experience, basic path characteristics, and current congestion status. When critical channels become congested, traditional methods relying solely on pheromone concentration and path length will continue to guide goods to already congested high-quality paths, exacerbating congestion. By introducing real-time conflict weight as a negative adjustment factor, while maintaining respect for historically high-quality paths, the probability of selecting congested paths can be dynamically reduced based on the current congestion situation, automatically guiding the flow of goods to relatively idle alternative paths. This multi-dimensional comprehensive decision-making mechanism has the ability to replan the flow of goods in real time, avoiding the problems of resource concentration and congestion exacerbation caused by blindly pursuing a single optimal path.

[0031] For example, suppose there are three possible paths from the cargo storage area to the shipping outlet within a transfer station: Path A goes directly via the main road, Path B detours through the left-hand area, and Path C detours through the right-hand area. Under normal circumstances, Path A is almost always the preferred choice in traditional scheduling methods because it has the shortest distance, a good historical track record, the highest pheromone concentration, and the optimal path length. Paths B and C, due to their longer distances, are rarely used in traditional methods.

[0032] When peak cargo handling season arrives, a large number of handling tasks concentrate on route A, causing congestion on the main road. Under traditional scheduling methods, new tasks continue to be assigned to route A based on historical pheromone concentration and distance advantages, further exacerbating the congestion. At this time, although routes B and C are idle, they are ignored due to low pheromone concentration and distance disadvantage, resulting in wasted resources and decreased overall efficiency.

[0033] After adopting a three-factor comprehensive consideration mechanism, when congestion occurs on path A, the equipment conflict perception mechanism immediately detects multiple virtual ants competing for the same path segment, calculates a higher conflict intensity coefficient, and converts it into a larger conflict weight value. When recalculating the selection probability, although path A maintains its high pheromone concentration and short-distance advantage, its selection probability is significantly reduced due to the negative adjustment of the conflict weight. In contrast, paths B and C, without the negative impact of conflict weight, have correspondingly higher selection probabilities. The system automatically redirects some cargo flow to paths B and C, effectively alleviating the congestion pressure on path A, achieving balanced utilization of the three paths, preventing cargo backlog, and improving overall traffic efficiency.

[0034] S3: The path selection probability distribution is dynamically adjusted according to the real-time channel congestion status, and the attraction weight of each path is updated by using a pheromone decay and enhancement mechanism to form an adaptive path preference sequence. like Figure 2 As shown, real-time congestion data is extracted from the status monitoring devices of each channel, including the number of transport equipment operating in the channel and the remaining capacity of the channel. The current selection probability of each path is compared with the actual congestion level of its corresponding channel. Adjustment needs are identified by calculating the match between the probability value and the congestion level. If a path has a high selection probability but the corresponding channel is also highly congested, a mismatch between probability and reality is identified, requiring negative adjustment to reduce its selection probability. When a path has a low selection probability but the corresponding channel is actually idle or lightly used, it is identified that the path has carrying capacity but the probability configuration is too low, requiring positive adjustment to increase its selection probability. The probability adjustment range for each path is calculated; the greater the deviation between the congestion level and the selection probability, the larger the adjustment range. The identified adjustment range is directly applied to the original selection probability value, and the adjusted probability value is obtained through addition or subtraction. The adjusted probability values ​​of all paths are re-normalized to ensure that the sum of probabilities meets the standard requirements, forming an updated probability distribution that matches the current real-time congestion situation.

[0035] The process involves obtaining the number of currently operating handling equipment from the channel status monitoring device and retrieving the maximum designed capacity of the channel from the channel configuration database. The current utilization rate of the channel is calculated by dividing the current number of operating equipment by the maximum designed capacity. The selection probability value calculated in the previous step is extracted, and the current utilization rate is subtracted from the selection probability value to obtain the probability deviation value. A positive probability deviation value indicates that the selection probability is higher than the actual utilization rate and needs to be reduced; a negative probability deviation value indicates that the selection probability is lower than the actual utilization rate and needs to be increased. The absolute value of the probability deviation value is used as the adjustment demand intensity index. The remaining passage capacity of the channel is obtained from the channel monitoring data, and the remaining passage capacity is divided by the maximum designed capacity to obtain the remaining capacity percentage. The adjustment demand intensity index is multiplied by the remaining capacity percentage to obtain the base value for capacity correction. When the probability deviation value is positive, the base value for adjustment is set to a negative value to reduce the selection probability; when the probability deviation value is negative, the base value for adjustment is set to a positive value to increase the selection probability, forming the final probability adjustment value for the path. This is illustrated in the following formula: in, in, Let $\frac{i}{j}$ be the path probability adjustment value from device $i$ to region $j$. The sign function determines the adjustment direction: when the probability deviation is positive, it outputs a negative sign to decrease the selection probability; when the probability deviation is negative, it outputs a positive sign to increase the selection probability. The probability deviation value is equal to the selection probability value minus the current channel usage value. This is the selection probability value for this path calculated in the previous step. This refers to the number of handling devices currently operating on this path channel, obtained from the channel status monitoring device. The maximum number of devices that can be carried by this path channel is determined by reading the channel configuration database. This refers to the remaining throughput capacity of the channel for this path, obtained from channel monitoring data. The probability deviation response intensity coefficient is set to 2.5 to ensure a reasonable amplification of the adjustment demand intensity index. The probability deviation standardization parameter is set to 4.0 to control the saturation characteristics of the adjustment amplitude. The base weighting coefficient for capacity adjustment, set to 1.5, affects the contribution of the remaining capacity percentage to the adjustment magnitude. The capacity response sensitivity parameter is set to 3.0 to control the response speed when the remaining capacity percentage changes. It is a natural constant.

[0036] Furthermore, the pheromone update direction and intensity for each path are determined based on the probability adjustment amplitude values ​​calculated in the previous step. If the probability adjustment amplitude for a path is negative, the pheromone attenuation mechanism is activated. The current pheromone concentration value for that path is extracted from the path pheromone intensity matrix, and the absolute value of the probability adjustment amplitude is used as an attenuation ratio factor to directly multiply by the current pheromone concentration to obtain the attenuation amount. The attenuation amount is subtracted from the current pheromone concentration to obtain the attenuated new pheromone concentration, and the new value is written back to the corresponding position in the pheromone intensity matrix to complete the attenuation update. When the probability adjustment amplitude for a path is positive, the pheromone enhancement mechanism is activated. The probability adjustment amplitude value is used as an enhancement ratio factor to multiply by the current pheromone concentration to obtain the enhancement amount, and then the enhancement amount is added to the current pheromone concentration to obtain the enhanced new pheromone concentration, which is then updated in the matrix. If the probability adjustment amplitude for a path is zero, the pheromone concentration for that path remains unchanged, and the update operation is skipped. After updating the pheromone concentration for all paths, the updated pheromone concentration value is extracted again and multiplied by the path length factor to calculate the new attraction weight value for each path, forming an updated attraction weight distribution that reflects the actual situation of the current channel.

[0037] It should be noted that using the probability adjustment amplitude to determine the pheromone update direction and intensity is to establish a dynamic feedback mechanism between path selection probability and historical experience records. The probability adjustment amplitude reflects the degree of deviation between the current path selection tendency and the actual channel conditions. If this deviation information is not fed back into the pheromone intensity matrix in a timely manner, decisions will continue to be made based on outdated historical experience. By directly converting the probability adjustment amplitude into pheromone attenuation or enhancement operations, current real-time scheduling experience can be integrated into the accumulation of historical experience, ensuring that the pheromone concentration always reflects the latest path quality and avoiding the disconnect between historical information and current reality. Although the improved ant colony algorithm in step S2 already has the ability to adjust according to real-time conditions, this adjustment is mainly reflected in conflict detection and weight allocation, which is a coarse-grained path selection intervention. The probability adjustment amplitude mechanism, on this basis, provides a deeper level of system optimization function, specifically addressing the numerical matching problem between probability distribution and actual carrying capacity.

[0038] Preferably, the improved real-time adjustment capability of the ant colony algorithm mainly achieves path avoidance through conflict weights. However, this mechanism suffers from insufficient adjustment precision. When multiple paths are congested to varying degrees, the conflict weight mechanism can only simply reduce the probability of selecting congested paths, failing to accurately allocate probabilities based on the specific capacity margin of each path. The probability adjustment amplitude mechanism fills this precision gap by quantitatively analyzing the deviation between the expected probability and the actual capacity of each path, thus achieving fine-grained adjustment of probability allocation.

[0039] More importantly, the probability adjustment amplitude plays a crucial self-correcting role. While the improved ant colony algorithm can respond to current conflicts, it lacks a proactive verification mechanism for the rationality of the probability distribution. The probability adjustment amplitude, by continuously monitoring the match between the probability configuration and the actual carrying capacity, can identify and correct deviations in the probability allocation process, ensuring that the path selection probability always maintains an optimal correspondence with the actual channel capacity. This corrective function is irreplaceable by the improved ant colony algorithm; the two form a complementary dual-guarantee mechanism, jointly ensuring the accuracy and efficiency of scheduling decisions.

[0040] Furthermore, the updated attractiveness scores are sorted, and a path priority order is established according to the scores, forming an adaptive path preference sequence that reflects the current actual channel conditions. This preference sequence dynamically reflects the relative merits of each path at the current moment; higher-priority paths will have more selection opportunities in subsequent cargo allocation decisions, while lower-priority paths will be selected less frequently. The dynamic adjustment process is continuously monitored and repeated to ensure that the path preference sequence reflects the latest congestion distribution within the transfer station in real time, providing accurate path priority guidance for subsequent intelligent scheduling decisions.

[0041] S4: Perform device task allocation according to the adaptive path preference sequence, allocate the handling tasks to the corresponding devices according to the path preference degree, update the pheromone intensity distribution synchronously, and complete the dynamic path optimization scheduling within the transfer station.

[0042] The previous step yields a complete adaptive path preference sequence, arranged by path priority and containing a specific preference value for each path. Simultaneously, the current queue of transport tasks to be assigned is extracted from the task management module. Each transport task includes basic information such as start and end locations, cargo type, urgency, and expected completion time. Based on the start and end coordinates of each transport task, all reachable paths are retrieved from the path database and matched against the adaptive path preference sequence to determine the available paths and their preference ranking for each task.

[0043] Task assignment is performed in descending order of path preference. For the path with the highest preference, the availability of handling equipment resources on that path is checked first, including equipment idle status, remaining power, current location, and equipment type matching. If an idle device matching the task requirements exists on that path, the task is immediately assigned to the device closest to the starting point and with the best performance match, and specific execution instructions are generated in the equipment scheduling. When there is no available equipment on the path with the highest preference, the availability of equipment on the next best path is checked according to the preference sequence, and so on, until a suitable equipment resource is found to complete the task assignment.

[0044] The system continuously monitors the execution progress and path accessibility of all assigned tasks, collecting real-time data on equipment movement, channel occupancy, and task completion. Based on this real-time data, it dynamically adjusts the allocation strategy and path preference sequence for subsequent tasks, forming a closed-loop adaptive scheduling optimization mechanism. When a path is detected to be congested due to continuous task allocation, its preference in subsequent allocations is immediately reduced, and tasks are redirected to other available paths. Through this collaborative mechanism of dynamic task allocation and real-time pheromone updates, intelligent scheduling of handling equipment and optimal utilization of path resources within the transfer station are achieved. This effectively avoids the localized congestion and resource waste problems caused by traditional static scheduling methods, significantly improving overall operational efficiency and cargo turnover speed.

[0045] In summary, the dynamic path optimization scheduling method for transit logistics based on embodiments of the present invention has been clarified. It dynamically collects and analyzes data such as the location of handling equipment and the delivery time window for goods, combining a multi-level ant colony pathfinding algorithm and an equipment conflict perception mechanism to achieve dynamic updates of path pheromone intensity and real-time adjustments to path selection probabilities, thereby effectively optimizing path scheduling efficiency. This significantly reduces the risk of channel congestion and the probability of concurrent equipment conflicts, improves path resource utilization and task allocation rationality, and helps enhance the dynamic response capability and overall operational efficiency of transit station logistics operations. Furthermore, the present invention fully considers the matching relationship between delivery time windows and equipment idle status in task scheduling, which can improve the reliability of timely delivery of goods and reduce sorting and transfer delays.

[0046] Here, those skilled in the art will understand that the specific operations of each step in the above-mentioned dynamic route optimization and scheduling system for transit logistics have been referenced above. Figure 1 and Figure 2 The dynamic path optimization scheduling method for transit logistics has been described in detail, and therefore, its repeated description will be omitted.

[0047] According to another aspect of the present invention, a dynamic route optimization and scheduling system for transit logistics is provided, comprising: The equipment conflict perception module is used to monitor the equipment occupancy status of each path segment in the transfer station in real time, detect path competition conflicts between virtual ants and calculate the conflict weight value. The comprehensive probability calculation module is used to integrate information from three dimensions: pheromone concentration, path length factor, and real-time conflict weight, to calculate the comprehensive selection probability distribution of each path segment. The probability adjustment range calculation module is used to compare the matching degree between the path selection probability and the actual congestion status of the channel, and calculate the probability adjustment range value required for each path. The pheromone dynamic update module is used to attenuate or enhance the pheromone concentration of each path according to the probability adjustment amplitude, forming an adaptive path preference sequence. The intelligent task allocation module is used to perform device task allocation based on the adaptive path preference sequence, synchronously update the pheromone intensity distribution, and complete dynamic path optimization scheduling.

[0048] In summary, the transit logistics dynamic path optimization scheduling system based on the embodiments of the present invention has been clarified. It dynamically collects and analyzes data such as the location of handling equipment and the delivery time window for goods, and combines a multi-level ant colony pathfinding algorithm and an equipment conflict perception mechanism to achieve dynamic updates of path pheromone intensity and real-time adjustments to path selection probabilities, thereby effectively optimizing path scheduling efficiency. This significantly reduces the risk of channel congestion and the probability of concurrent equipment conflicts, improves path resource utilization and task allocation rationality, and helps enhance the dynamic response capability and overall operational efficiency of transit station logistics operations. Furthermore, the present invention fully considers the matching relationship between delivery time windows and equipment idle status in task scheduling, which can improve the reliability of timely delivery of goods and reduce sorting and transfer delays.

Claims

1. A dynamic path optimization scheduling method for transit logistics, characterized in that, include: Receive the location coordinates of each handling equipment in the transfer station and the cargo delivery time window data, set the initial value of pheromone according to the idle time of the equipment and the remaining processing time of the cargo, and generate a path pheromone intensity matrix; Based on the path pheromone intensity matrix, multi-level ant colony pathfinding calculation is performed to construct a device conflict perception mechanism to monitor the concurrent usage status of each path segment in real time and output a path selection probability distribution containing conflict weights. The path selection probability distribution is dynamically adjusted in accordance with the real-time channel congestion status, and the attractiveness weight of each path is updated using a pheromone decay and enhancement mechanism to form an adaptive path preference sequence. Based on the adaptive path preference sequence, the device task allocation is performed, and the handling tasks are assigned to the corresponding devices according to the path preference degree. The pheromone intensity distribution is updated synchronously to complete the dynamic path optimization scheduling within the transfer station.

2. The dynamic path optimization scheduling method for transit logistics according to claim 1, characterized in that, The initial pheromone value is set by matching the idle time of the equipment with the remaining processing time of the goods to obtain a time matching coefficient. The pheromone intensity adjustment factor is determined based on the deviation between the time matching coefficient and the average matching coefficient of all equipment and goods pairings. Finally, the basic pheromone intensity is corrected using the pheromone intensity adjustment factor to obtain the initial pheromone value of the path segment.

3. The dynamic path optimization scheduling method for transit logistics according to claim 2, characterized in that, When the device conflict perception mechanism detects that multiple virtual ants are simultaneously selecting the same path segment, it immediately triggers a conflict marker update operation, calculates the conflict weight value based on the number of virtual ants in the conflict path segment, calculates the path attractiveness score based on the conflict weight value, and then normalizes it to obtain the path selection probability distribution. The device conflict perception mechanism establishes a status monitoring table for each path segment to record the virtual ant identifiers in use. When multiple ants enter the same path segment at the same time, a concurrent conflict is detected and the path segment status is updated. Real-time monitoring of concurrent usage status is achieved by continuously scanning the monitoring table to count the number of occupants in each path segment.

4. The dynamic path optimization scheduling method for transit logistics according to claim 3, characterized in that, The conflict weight value is obtained by comprehensively evaluating the conflict intensity coefficient based on the comparison between the number of virtual ants in the conflict path segment and the carrying capacity of the path segment, as well as the comparison between the conflict duration and the standard processing time. Then, the conflict intensity coefficient is converted into a conflict weight value through an inverse proportional conversion.

5. The dynamic path optimization scheduling method for transit logistics according to claim 4, characterized in that, The path selection probability value in the path selection probability distribution is calculated as follows: in, Let be the combined selection probability value from device i to region j. The initial pheromone value for the path from device i to region j is used as the pheromone concentration factor. Let be the actual length of the path segment from device i to region j. This represents the real-time conflict weight values ​​for the path segment from device i to region j. This represents the positive promoting weighting coefficient of the pheromone concentration factor. Optimize the weighting coefficients for the path length factor based on distance. The congestion penalty weight coefficient is the conflict weight factor. For conflict sensitivity adjustment parameters, This represents the total number of all possible path segments. This is the numerical stability protection constant. Let be a natural constant, and k be all possible regions. This iterates through all possible target regions starting from the current device i. This represents the real-time conflict weight value of the path segment from device i to region k.

6. The dynamic path optimization scheduling method for transit logistics according to claim 5, characterized in that, The path selection probability distribution is dynamically adjusted in accordance with the real-time channel congestion status. The probability adjustment range is determined by calculating the deviation between the path selection probability and the actual congestion level of the channel. The probability adjustment range is then used to directly attenuate or enhance the pheromone concentration, thereby achieving dynamic matching between the path attraction weight and the current channel status.

7. The dynamic path optimization scheduling method for transit logistics according to claim 6, characterized in that, The probability adjustment amplitude is calculated as follows: in, Let $\frac{i}{j}$ be the path probability adjustment value from device $i$ to region $j$. The sign function determines the adjustment direction: when the probability deviation is positive, it outputs a negative sign to decrease the selection probability; when the probability deviation is negative, it outputs a positive sign to increase the selection probability. The probability deviation value is equal to the selection probability value minus the current channel usage value. The maximum number of devices that can be carried by this path channel is determined by reading the channel configuration database. This refers to the remaining throughput capacity of the channel for this path, obtained from channel monitoring data. This is the probability bias response intensity coefficient. For probability bias standardization parameters, Adjust the basic weighting coefficient for capacity. For capacity response sensitivity parameters, It is a natural constant.

8. A dynamic route optimization and scheduling system for transit logistics, characterized in that, include: The equipment conflict perception module is used to monitor the equipment occupancy status of each path segment in the transfer station in real time, detect path competition conflicts between virtual ants and calculate the conflict weight value. The comprehensive probability calculation module is used to integrate information from three dimensions: pheromone concentration, path length factor, and real-time conflict weight, to calculate the comprehensive selection probability distribution of each path segment. The probability adjustment range calculation module is used to compare the matching degree between the path selection probability and the actual congestion status of the channel, and calculate the probability adjustment range value required for each path. The pheromone dynamic update module is used to attenuate or enhance the pheromone concentration of each path according to the probability adjustment amplitude, forming an adaptive path preference sequence. The intelligent task allocation module is used to perform device task allocation based on the adaptive path preference sequence, synchronously update the pheromone intensity distribution, and complete dynamic path optimization scheduling.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the transit logistics dynamic path optimization scheduling method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the transit logistics dynamic path optimization scheduling method according to any one of claims 1 to 7.

Images