A logistics transportation method based on electronic fence

Abstract

This invention relates to a logistics transportation method based on electronic fences, belonging to the field of logistics transportation. The method includes: selecting process DNA matching the logistics task to be processed from a gene bank as parent DNA; performing genetic processing on the parent DNA to obtain multiple offspring DNAs; selecting one of the offspring DNAs as the master DNA based on digital twin technology; generating a dynamic electronic fence based on the master DNA; calculating a deviation vector based on acquired current state data and future state trajectory prediction data; when the deviation vector meets preset conditions, determining the gene segment to be mutated in the master DNA based on the deviation vector and performing mutation processing to obtain multiple variant DNAs; selecting one of the variant DNAs as the optimal variant DNA based on digital twin technology and current state data; updating the master DNA and the dynamic electronic fence based on the optimal variant DNA. This method has the advantages of enabling proactive monitoring and reducing cargo damage and transportation delays.

Description

Technical Field

[0001] This invention relates to the technical field of logistics transportation, and in particular to a logistics transportation method based on electronic fences. Background Technology

[0002] With the rapid development of IoT, big data and digital twin technologies, visualization and real-time monitoring of logistics transportation processes have been widely applied. However, existing logistics management systems mainly adopt a perception-alarm-manual intervention management model, which involves collecting data through sensors and comparing the collected data with preset thresholds, such as fixed paths, fixed time windows or fixed upper and lower temperature limits. When the collected data exceeds the static threshold, an alarm is triggered, and dispatchers manually analyze the data and issue processing instructions.

[0003] The above approach is essentially a passive management model, which has the following drawbacks: The system only issues passive alarms when the collected data exceeds a threshold. By this time, the abnormal event has already occurred, such as temperature deviation, route deviation, or delays. Managers can only learn about the problem and take action after the abnormal event has occurred, missing the best intervention window and causing irreparable losses such as cargo damage and transportation delays. Furthermore, the actual logistics environment is time-varying and uncertain. Factors such as road congestion, sudden weather changes, and temporary traffic control can affect the optimal route and time window. Fixed thresholds cannot adapt to normal dynamic fluctuations during transportation, which can easily lead to false alarms or missed alarms, reducing the effectiveness of the alarm. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a logistics transportation method based on electronic fences, which aims to solve at least one of the above-mentioned technical problems.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: This application provides a logistics transportation method based on electronic fences, which adopts the following technical solution: A logistics transportation method based on electronic fences includes: Get pending logistics tasks; At least two process DNAs that match the logistics task to be processed are selected from the gene bank as parent DNAs. The gene bank includes process DNAs encoded by multiple successfully completed historical logistics tasks, and each process DNA represents a logistics operation sequence. The parental DNA was subjected to genetic processing to obtain multiple offspring DNAs; Based on digital twin technology, each of the daughter DNAs is simulated to obtain a first simulation result, and one of the daughter DNAs is selected as the master DNA based on the first simulation result. A dynamic electronic fence is generated based on the master DNA, and the dynamic electronic fence includes a dynamic path, a dynamic time window, and dynamic driving environment parameter thresholds. Real-time acquisition of current state data and future state trajectory prediction data, and calculation of the deviation vector between the future state trajectory prediction data and the dynamic electronic fence; When the deviation vector meets the preset conditions, the gene fragment to be mutated in the main DNA is determined based on the deviation vector, and the gene fragment to be mutated is subjected to mutation processing to obtain multiple variant DNAs; Based on the digital twin technology and current state data, each of the mutant DNAs is simulated to obtain a second simulation result, and one of the mutant DNAs is selected as the optimal mutant DNA based on the second simulation result; The master DNA is updated based on the optimal variant DNA to obtain the update result, and the dynamic electronic fence is updated based on the update result.

[0006] The beneficial effects of this invention are as follows: Successfully completed historical logistics tasks are encoded as process DNA, and dynamic electronic fences are generated through genetic processing and digital twin technology. This avoids the passive management mode with fixed thresholds, obtains current status data and future status trajectory prediction data in real time, and calculates the deviation vector from the dynamic electronic fence. This allows for early identification of risks in abnormal trend stages, overcoming the shortcomings of passive alarms and missed intervention windows after exceeding thresholds. When the deviation meets the conditions, the gene fragment to be mutated in the mutated master DNA is determined, and the optimal variant is selected through digital twin technology to update the master DNA and the dynamic electronic fence. This allows the dynamic electronic fence to adapt to time-varying factors such as road conditions and weather, reducing false alarm and missed alarm rates, and improving the proactive early warning capability to a certain extent. At the same time, it reduces cargo damage and transportation timeouts, improving transportation efficiency and transportation safety.

[0007] Based on the above technical solution, the present invention can be further improved as follows.

[0008] Furthermore, the construction of the gene bank includes: Acquire historical information of multiple successfully completed historical logistics tasks. The historical information includes cargo type, historical travel route, historical arrival and departure time of each node, historical travel environment parameters of each transportation segment, and historical operation events of each node. Based on the task execution order, each historical logistics task is divided into multiple transportation segments. Each transportation segment includes a starting node, an ending node, historical transportation actions between the starting node and the ending node, and corresponding historical operation events. The historical transportation actions include historical driving routes, historical driving time, and historical driving environment parameters. Based on preset coding rules, all transportation segments of each historical logistics task are sequentially arranged to obtain the process DNA. The process DNA is a transportation segment sequence, each gene fragment in the process DNA corresponds to a transportation segment, and each process DNA corresponds to a cargo type. The gene library is generated based on all process DNA, and a success evaluation index is generated for each process DNA, wherein the success evaluation index includes whether it is on time, whether cargo damage occurs, and whether energy consumption is lower than a preset standard.

[0009] The beneficial effects of adopting the above-mentioned further scheme are: successfully completed historical logistics tasks are transformed into process DNA, so that subsequent genetic selection, mutation and other operations are processed based on success evaluation indicators. By taking the transportation segment as the smallest gene fragment, the causal relationship of historical logistics tasks in the spatiotemporal dimension is preserved, avoiding the information fragmentation problem caused by simple statistics or rule extraction. By identifying the process DNA by cargo type, it is ensured that the cargo type of the selected parent DNA matches the cargo type of the logistics task to be processed, which improves the effectiveness of offspring DNA generation to a certain extent.

[0010] Furthermore, the step of selecting at least two process DNAs from the gene bank that match the logistics task to be processed as parental DNAs includes: Obtain the task start node, task end node, task cargo type, current driving environment parameters, and expected delivery time of the logistics task to be processed; The total mileage of the logistics task to be processed is calculated based on the task start node and the task end node, and the expected total time is calculated based on the expected delivery time and the current time. Select process DNA from the gene bank that matches the task start node, task end node, and cargo type; For each selected process DNA, the similarity between the process DNA and the logistics task to be processed is calculated based on the total mileage, expected total time, current driving environment parameters, historical total mileage, historical total time, and historical driving environment parameters. Select the N process DNAs with the highest similarity as the candidate set, where N≥3; Two process DNAs are randomly selected from the candidate set as parental DNAs, wherein the probability of random selection is positively correlated with the success evaluation index of each candidate DNA.

[0011] The beneficial effects of adopting the above-mentioned further scheme are as follows: the similarity between the logistics task to be processed and the process DNA in the gene bank is calculated by the task start node, task end node, cargo type, current driving environment parameters and expected delivery time. The N most similar ones are selected as the candidate set. Based on the probability of a positive correlation with the success evaluation index, two parent DNAs are randomly selected from the candidate set. This ensures that the selected parent DNA matches the logistics task to be processed. The higher the success evaluation index, the greater the probability of the process DNA being selected. This guides the genetic algorithm to converge towards a better direction, retains a certain degree of randomness to avoid local optima, and improves the quality of offspring DNA to a certain extent.

[0012] Furthermore, the genetic treatment of the parental DNA includes: Obtain the transport segment sequence of the parental DNA; Select at least one cutting position, and cut the transport segment sequences of the two parental DNAs based on the cutting position. Exchange the cut tail transport segment fragments and splice them together to obtain two first daughter DNAs. For each first progeny DNA, at least one transport segment is selected, and the parameters of the selected transport segment are adjusted to obtain multiple second progeny DNAs. All first-generation and second-generation DNA were combined to obtain multiple daughter DNAs; The parameter adjustment includes at least one of the following methods: Replace the historical driving route of the transportation segment with another feasible route on the map that connects the same starting node and ending node; simultaneously advance or postpone the start and end times of the historical time window of the transportation segment by a first preset duration; and adjust the threshold values ​​of the historical driving environment parameters of the transportation segment by a preset value.

[0013] The beneficial effects of adopting the above-mentioned further scheme are as follows: by cutting and exchanging the tail transport segment fragments of the parent DNA to generate the first daughter DNA, adjusting the parameters of the transport segment in the first daughter DNA to generate the second daughter DNA, and summing the first and second daughter DNA to obtain multiple daughter DNAs, the excellent transport segment sequences from successfully completed historical logistics tasks are recombined and locally mutated. This not only retains the core advantages of the parent DNA, but also adds new feasible paths, dynamic time windows, and dynamic driving environment parameter thresholds, thereby overcoming the defects of fixed paths, times, and thresholds in traditional methods. By adjusting parameters to generate diverse daughter DNAs, a rich pool of candidate schemes is provided for digital twin simulation, which to a certain extent increases the probability of screening for better master DNA.

[0014] Furthermore, the generation of a dynamic electronic fence based on the master DNA includes: Extract the historical driving path, historical time window, and historical driving environment parameter thresholds corresponding to each transport segment from the master DNA. Connect the historical travel routes of all transportation segments in the order of execution to obtain the dynamic path of the transportation process; For each transport segment, the start and end times of the historical time window are extended forward and backward by a second preset duration, respectively, to obtain a dynamic time window; For each transportation segment, the lower and upper limits of the historical driving environment parameter thresholds are adjusted downward and upward by a preset tolerance range to obtain the dynamic driving environment parameter thresholds. The dynamic path, dynamic time window, and dynamic driving environment parameter thresholds are used as the dynamic electronic fence.

[0015] The beneficial effects of adopting the above-mentioned further scheme are as follows: extract the historical driving path, historical time window, and historical driving environment parameter threshold of each transportation segment from the master DNA, connect the historical driving paths of each transportation segment in sequence to form a dynamic path, expand the historical time window in both directions to obtain the dynamic time window, and adjust the historical driving environment parameter threshold up and down to obtain the dynamic driving environment parameter threshold, thereby generating a complete dynamic electronic fence. Then, the genetically optimized master DNA is converted into a three-dimensional electronic fence constraint. By expanding the time window and floating the driving environment parameter threshold, a flexible space is reserved for the actual transportation process, avoiding frequent false alarms due to small fluctuations in the fixed threshold. The segmented connection of the dynamic path ensures the continuity and consistency of the fence in the spatiotemporal dimension.

[0016] Furthermore, the calculation of the deviation vector between the future state trajectory prediction data and the dynamic electronic fence includes: The future state trajectory prediction data and the dynamic electronic fence are aligned in the same spatiotemporal coordinate system to obtain the aligned predicted position sequence, the predicted time sequence of arrival at each node, and the predicted driving environment parameter change curve. Calculate the spatial deviation component based on the predicted location sequence and the dynamic path in the dynamic electronic fence; Calculate the time deviation component based on the predicted arrival time series of each node and the dynamic time window in the dynamic electronic fence; Calculate the environmental deviation component based on the predicted driving environment parameter change curve and the dynamic driving environment parameter threshold in the dynamic electronic fence; The spatial deviation component, temporal deviation component, and environmental deviation component are used as the deviation vector of the dynamic electronic fence.

[0017] The beneficial effects of adopting the above-mentioned further scheme are as follows: by aligning the future state trajectory prediction data and the dynamic electronic fence in the same spatiotemporal coordinate system, the spatial deviation component, temporal deviation component, and environmental deviation component are calculated to form a multi-dimensional deviation vector, thereby realizing the multi-dimensional quantification of deviation risk in the transportation process. This overcomes the limitations of traditional schemes that rely solely on a single spatiotemporal threshold or simple over-limit alarms. By separating the spatial deviation component, temporal deviation component, and environmental deviation component, the specific dimension and degree of anomaly occurrence can be located, providing a basis for identifying gene fragments to be mutated and for mutation processing. This improves the effectiveness of mutation operations to a certain extent and enhances the adjustment accuracy of the dynamic electronic fence.

[0018] Furthermore, determining the gene fragment to be mutated in the main DNA based on the deviation vector includes: For the spatial deviation component, identify the spatial deviation point between the predicted position sequence in the future state trajectory prediction data and the dynamic path in the dynamic electronic fence, and take the transport segment corresponding to the spatial deviation point as the first candidate gene fragment. For the time deviation component, identify the time-exceeding nodes in the future state trajectory prediction data whose predicted arrival time series exceeds the dynamic time window in the dynamic electronic fence, and use the transportation segment corresponding to the time-exceeding node as the second candidate gene fragment. For the environmental deviation component, identify the environmental over-limit period when the predicted driving environment parameter change curve in the future state trajectory prediction data exceeds the dynamic driving environment parameter threshold in the dynamic electronic fence, and take the transportation segment corresponding to the environmental over-limit period as the third candidate gene segment. The first candidate gene fragment, the second candidate gene fragment, and the third candidate gene fragment are merged and deduplicated to obtain a set of gene fragments to be mutated.

[0019] The beneficial effects of adopting the above-mentioned further scheme are as follows: For spatial deviation, the transportation segment corresponding to the deviation point between the predicted location and the dynamic path is determined; for temporal deviation, the transportation segment corresponding to the node arrival time exceeding the limit is determined; for environmental deviation, the transportation segment corresponding to the period when environmental parameters exceed the limit is determined. After merging and deduplicating the three, a set of gene fragments to be mutated is obtained, thereby realizing the location of abnormal transportation segments and distinguishing whether the deviation is due to path error, time delay or environmental exceedance. This avoids the defect that a single threshold alarm cannot clearly identify the root cause of the problem. By merging and deduplicating, it is ensured that each abnormal transportation segment is processed only once, which improves the targeting and efficiency of subsequent mutation operations, while avoiding duplicate mutation or missed mutation.

[0020] Furthermore, the mutation treatment of the gene fragment to be mutated yields multiple variant DNAs, including: Based on the type of the gene fragment to be mutated and the deviation direction of the deviation vector, multiple mutation operators are selected from a preset mutation operator library. The mutation operator library includes a path replacement operator for adjusting the path, a time translation or scaling operator for adjusting the time window, and a threshold scaling operator for adjusting the threshold of driving environment parameters. Based on the selected mutation operator, the gene fragment to be mutated is processed sequentially to obtain multiple first-class variant DNAs. Each time the gene fragment to be mutated is processed, only one mutation operator is used. Multiple second-type variant DNAs are generated based on a randomly combined mutation operator and the gene fragment to be mutated; The first type of variant DNA and the second type of variant DNA were combined to obtain multiple variant DNAs.

[0021] The beneficial effects of adopting the above-mentioned further scheme are as follows: Based on the type of gene fragment to be mutated and the deviation direction of the deviation vector, the corresponding mutation operator is selected from the preset mutation operator library. The first type of variant DNA is obtained by processing with a single mutation operator, and the second type of variant DNA is obtained by processing with multiple mutation operators in random combination. The first type of variant DNA and the second type of variant DNA are combined to obtain multiple variant DNA, thereby realizing the mutation of abnormal transport segments. Among them, the mutation of a single mutation operator ensures the targeting of local optimization and avoids excessive mutation that destroys effective gene fragments. The mutation of multiple mutation operators increases the diversity of variants and provides a data foundation for solving multi-dimensional deviation problems at the same time.

[0022] Furthermore, the first simulation results include transportation time, transportation energy consumption, time window deviation, and duration of exceeding driving environment parameters. Based on the first simulation results, one of the offspring DNAs is selected as the master DNA, including: Based on the transportation time, transportation energy consumption, time window deviation, and the duration of exceeding the driving environment parameters, a comprehensive score is calculated for each offspring DNA, and the offspring DNA with the highest score is selected as the candidate master DNA. The candidate master DNA is subjected to feasibility verification. If the candidate master DNA passes the feasibility verification, it is adopted as the master DNA. If the candidate master DNA fails the feasibility verification, the progeny DNAs are verified in descending order of comprehensive score until the master DNA is determined.

[0023] The beneficial effects of adopting the above-mentioned further scheme are as follows: by calculating the comprehensive score of each offspring DNA based on transportation time, transportation energy consumption, time window deviation, and the time exceeding the limit of driving environment parameters, the highest score is selected as the candidate master DNA, and its feasibility is verified. If it passes, it is determined as the master DNA. If it fails, the subsequent offspring DNAs are verified in descending order of score until a feasible solution is found. This avoids the defect that optimization of a single index will lead to the deterioration of other indices. Through feasibility verification, it is ensured that the selected master DNA not only has excellent simulation performance, but is also executable in the actual logistics environment, overcoming the risk that traditional simulation optimization ignores actual constraints and the solution becomes unusable.

[0024] Furthermore, the second simulation results include the predicted travel path for the remaining transport segments, the predicted arrival time at each node, the predicted travel time for each transport segment, the predicted energy consumption, the predicted curves of changes in travel environment parameters, and a conformity indicator for each transport segment regarding whether it meets the dynamic time window and travel environment parameter thresholds. The selection of one variant DNA as the optimal variant DNA based on the second simulation results includes: Based on the predicted travel path of the remaining transport segments, the predicted arrival time at each node, the predicted travel time of each transport segment, the predicted energy consumption, the predicted change curve of the travel environment parameters, and the conformity identifier and corresponding weight value of each transport segment in terms of whether it meets the dynamic time window and the threshold of the travel environment parameters, the corrected fitness of each variant DNA is calculated. The mutant DNA with the highest modified fitness was selected as the optimal mutant DNA.

[0025] The beneficial effects of adopting the above-mentioned further scheme are as follows: by calculating the corrected fitness of each variant DNA through the predicted travel path of the remaining transportation segment, the predicted time to each node, the time spent on each segment, energy consumption, environmental parameter change curves, compliance identifiers, and weight values, and selecting the variant DNA with the highest corrected fitness, the rationality of the route, time compliance, energy consumption economy, and environmental adaptability of the remaining transportation segment are comprehensively evaluated in a weighted manner. This avoids the one-sidedness caused by selecting variants based on a single indicator. The compliance identifier ensures that variants that meet the dynamic time window and driving environment parameter threshold constraints are selected first. At the same time, transportation efficiency and energy consumption are added to achieve the optimal decision under the trade-off of multiple objectives. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating a logistics transportation method based on electronic fences in this embodiment. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0028] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0029] This application provides a logistics transportation method based on electronic fences. This logistics transportation method based on electronic fences can be executed by electronic devices, which can be servers or mobile terminal devices. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services. The mobile terminal device can be a laptop computer, a desktop computer, etc., but is not limited to these.

[0030] like Figure 1 As shown in the figure, the logistics transportation method based on electronic fence provided in this application includes the following steps: S101, Obtain pending logistics tasks; In this embodiment, the user inputs the logistics task into the electronic device via an input device. The electronic device then uses the acquired logistics task as a pending logistics task. The input device includes, but is not limited to, a mouse and keyboard, and the electronic device includes, but is not limited to, a computer and a mobile phone. The pending logistics task includes the task start node, the task end node, the type of goods, the current driving environment parameters, and the expected delivery time. In other optional embodiments, the pending logistics task can also be connected to the upstream business system through a data interface to automatically obtain pending logistics tasks from the order management system or the warehouse management system. For example, when a new order is generated and needs to be shipped, the order management system converts the order into a pending logistics task and sends the pending logistics task to the electronic device.

[0031] S102, at least two process DNAs that match the logistics task to be processed are selected from the gene bank as parent DNAs. The gene bank includes process DNAs encoded by multiple successfully completed historical logistics tasks, and each process DNA represents a logistics operation sequence.

[0032] In this embodiment, a logistics operation sequence is a series of ordered operations required to complete a logistics task, such as a cargo transportation operation sequence from the origin to the destination. The logistics operation sequence includes the origin node, the destination node, historical transportation actions, driving environment parameters, and historical operation events.

[0033] The construction of the gene bank includes: acquiring historical information of multiple successfully completed historical logistics tasks, including cargo type, historical travel route, historical arrival and departure times of each node, historical travel environment parameters of each transport segment, and historical operation events of each node; dividing each historical logistics task into multiple transport segments based on the task execution order, where each transport segment includes the origin node, destination node, historical transport actions between the origin node and destination node, and corresponding historical operation events, with historical transport actions including historical travel route, historical travel time, and historical travel environment parameters; sequentially arranging all transport segments of each historical logistics task according to preset coding rules to obtain process DNA, where process DNA is a transport segment sequence, each gene fragment in process DNA corresponds to one transport segment, and each process DNA corresponds to one cargo type; generating a gene bank based on all process DNA, and generating success evaluation indicators for each process DNA, where success evaluation indicators include whether it is on time, whether cargo damage occurred, and whether energy consumption is lower than preset standards.

[0034] In this embodiment, historical information of successfully completed historical logistics tasks can be obtained through the data interface of the transportation management system and IoT devices. For example, the transportation management system can obtain the starting node, ending node, cargo type, and historical arrival and departure time of each node for historical logistics tasks. The IoT devices can obtain the real-time location, speed, and cargo temperature of logistics vehicles. Weather information and traffic flow information can be obtained through the meteorological service interface, traffic information platform, or vehicle sensors. The IoT devices include, but are not limited to, GPS trackers and temperature sensors. Historical driving environment parameters include, but are not limited to, cargo temperature, weather information, and traffic flow information.

[0035] In this embodiment, when each historical logistics task is divided into multiple transportation segments based on the task execution order, key nodes in the historical logistics task, such as loading nodes, unloading nodes, transfer stations, and inspection stations, are identified as the boundaries of the transportation segments. The historical travel path of each historical transportation task is divided into continuous transportation segments according to the order of the key nodes. Each transportation segment includes a starting node and an ending node, and is associated with the corresponding historical transportation actions and historical operation events. The historical operation events include loading and unloading, inspection, waiting, signing, transfer, and temporary stopping.

[0036] In this embodiment, for sequentially arranging all transportation segments of each historical logistics task based on preset coding rules, the key attributes of each transportation segment, such as the starting node ID, ending node ID, hash value of the historical driving path, historical total time, historical driving environment parameter range, and operation event type, are encoded into a fixed-length string as a gene segment in the process DNA. The gene segments are then connected according to the task execution order to form a complete transportation segment sequence, i.e., the process DNA, and a cargo type identifier is added to each process DNA. In other optional embodiments, each transportation segment can also be mapped to a preset transportation segment type, such as an urban delivery segment, a highway segment, and a port transshipment segment, and encoded in combination with key parameters, such as distance, average speed, and historical operation event number, to obtain a transportation segment sequence.

[0037] In this embodiment, all coded process DNAs are stored in a database to form a gene bank. Success evaluation indicators can be generated by comparing historical arrival times with planned arrival times to determine on-time performance, using sensor data such as impact sensors, temperature sensors, or human feedback to determine if cargo damage has occurred, and comparing vehicle fuel or electricity consumption data with preset benchmarks to determine if energy consumption is below a preset standard. The preset standard is set as needed and is not specifically limited here. In other optional embodiments, machine learning models can be used to predict or evaluate success evaluation indicators. For example, a classification model can be trained, taking historical information and results of historical logistics tasks corresponding to the process DNA as input, and outputting its success probability. The historical results include on-time performance, cargo integrity, and energy consumption.

[0038] The process involves selecting at least two process DNAs from the gene bank that match the logistics task to be processed as parent DNAs. This includes: obtaining the task start node, task end node, cargo type, current driving environment parameters, and expected delivery time of the logistics task; calculating the total mileage of the logistics task based on the task start node and task end node, and calculating the expected total time based on the expected delivery time and the current time; selecting process DNAs from the gene bank that match the task start node, task end node, and cargo type; for each selected process DNA, calculating the similarity between the process DNA and the logistics task to be processed based on the total mileage, expected total time, current driving environment parameters, historical total mileage, historical total time, and historical driving environment parameters; selecting the N process DNAs with the highest similarity as a candidate set, where N≥3; and randomly selecting two process DNAs from the candidate set as parent DNAs, where the probability of random selection is positively correlated with the success evaluation index of each candidate DNA.

[0039] The current driving environment parameters can be obtained in real time from meteorological service interfaces, traffic information platforms, IoT devices, or vehicle sensors.

[0040] In this embodiment, the total mileage is calculated using a geographic information system or electronic map, for example, by calculating the distance of the shortest path or recommended path between the task start node and the task end node. The electronic map includes Amap and Baidu Map. The expected total time is calculated by subtracting the current time from the expected delivery time.

[0041] In this embodiment, the selection of process DNA can be achieved through database query language. For example, the start node, end node and cargo type of process DNA in the gene library can be matched to select process DNA that matches the logistics task to be processed. Alternatively, fuzzy matching or ontology-based matching methods can be used to identify potential matches even when there are subtle differences in node names or cargo types. For example, Huamou City Pumou New District and Huamou Pumou can be identified as a match, and frozen meat and fresh food can be identified as a match.

[0042] In this embodiment, similarity is calculated using a weighted average method or an Euclidean distance method. For example, different weights are assigned to the total mileage, expected total time, current driving environment parameters, historical total mileage, historical total time, and historical driving environment parameters to calculate a weighted distance or weighted similarity score. Alternatively, machine learning models, such as support vector machines or neural networks, can be used to predict the similarity between the logistics task to be processed and the selected process DNA by learning the matching relationship between historical logistics tasks and process DNA.

[0043] In this embodiment, all calculated similarity scores are sorted in descending order, and the top N process DNAs are selected as the candidate set. Two process DNAs are randomly selected from the candidate set as parent DNAs. The roulette wheel selection method can be used to calculate the probability of each candidate DNA being selected based on its success evaluation index. For example, the probability of a candidate DNA being selected can be calculated by combining the scores of on-time rate, cargo integrity rate, and energy consumption compliance rate.

[0044] S103 involves genetically processing the parental DNA to obtain multiple offspring DNAs.

[0045] Specifically, the transport segment sequence of the parent DNA is obtained; at least one cutting position is selected, and the transport segment sequences of the two parent DNAs are cut based on the cutting position. The cut tail transport segment fragments are exchanged and spliced ​​to obtain two first-generation DNAs; for each first-generation DNA, at least one transport segment is selected, and the parameters of the selected transport segment are adjusted to obtain multiple second-generation DNAs; all first-generation DNAs and second-generation DNAs are summarized to obtain multiple daughter DNAs; wherein, the parameter adjustment includes at least one of the following methods: replacing the historical travel path of the transport segment with another feasible path on the map connecting the same starting node and ending node, simultaneously advancing or delaying the start and end times of the historical time window of the transport segment by a first preset duration, and raising or lowering the preset value of the historical travel environment parameter threshold of the transport segment.

[0046] In this embodiment, the cutting position can be selected randomly, for example, one or more positions can be randomly selected between any two transport segments of the transport segment sequence; or it can be selected according to a preset strategy, for example, selecting key nodes, such as transit stations. The cutting and exchange splicing involves breaking the sequences of the two parental DNAs at the selected cutting position, exchanging the broken tail transport segment fragments, and splicing the exchanged tail transport segment fragments with the respective first half of the transport segment fragments to form two new first-generation daughter DNAs.

[0047] For each first-generation DNA, at least one transport segment is selected. The transport segment can be selected randomly, i.e., one or more transport segments are randomly selected for mutation; or it can be selected according to a preset probability distribution or based on certain heuristic rules, such as selecting the transport segment that is expected to cause problems.

[0048] For route replacement, in this embodiment, an electronic map or route planning algorithm is used to search for alternative routes other than historical routes between the starting node and the ending node, and one alternative route is randomly selected for replacement; alternative routes can also be identified based on historical data or real-time traffic information, and route segments with congestion, construction or other conditions can be replaced with detour routes.

[0049] Regarding time window adjustment, in this embodiment, the time window for a transportation segment is selected based on the urgency of the current task or the expected delay. The start and end times are shifted forward or backward by a first preset duration. For example, if the logistics task to be processed is expected to be completed ahead of schedule, the subsequent time windows are advanced as a whole. The first preset duration is set as needed and is not specifically limited. Alternatively, historical data analysis can be combined to model the time fluctuations of a specific transportation segment, and the start and end times of the historical time window can be adjusted according to the fluctuation range predicted by the model.

[0050] Regarding the adjustment of driving environment parameter thresholds, in this embodiment, the thresholds of driving environment parameters for selected transportation segments are adjusted up or down based on the sensitivity level of the cargo to driving environment parameters. For example, for insensitive cargo, the temperature threshold is relaxed. Alternatively, the environmental conditions faced by a specific transportation segment can be predicted based on historical environmental data and future weather forecasts, and the upper and lower limits of the driving environment parameter thresholds can be dynamically adjusted according to the environmental conditions. The preset values ​​are set as needed and are not specifically limited.

[0051] In this embodiment, the first and second progeny DNAs are merged to obtain multiple progeny DNAs. During the merging process, deduplication is performed to avoid duplicate process DNA.

[0052] S104. Simulate each offspring DNA using digital twin technology to obtain the first simulation result, and select one offspring DNA as the master DNA based on the first simulation result.

[0053] In this embodiment, the digital twin model includes a vehicle dynamics model, a road network topology model, a traffic flow dynamics model, an environmental prediction model, and a node operation model. Each offspring DNA is input into the digital twin model for simulation to obtain the first simulation result. The first simulation result includes transportation time, transportation energy consumption, time window deviation, and the duration of exceeding the limit of driving environment parameters.

[0054] The process of selecting one of the offspring DNAs as the master DNA based on the first simulation results includes: calculating a comprehensive score for each offspring DNA based on transportation time, transportation energy consumption, time window deviation, and the duration of exceeding the limits of driving environment parameters, and selecting the offspring DNA with the highest score as the candidate master DNA; performing feasibility verification on the candidate master DNA, and if the candidate master DNA passes the feasibility verification, it is used as the master DNA; if the candidate master DNA fails the feasibility verification, the offspring DNAs are verified in descending order of comprehensive score until the master DNA is determined.

[0055] In this embodiment, the transportation time is the total time spent by a historical logistics task from the starting node to the ending node; the transportation energy consumption is the calculation of the fuel or electricity consumption required to complete the logistics task; the time window deviation is the degree of deviation between the actual arrival or departure time of the logistics task at each key node and the preset time window, and the key nodes include, but are not limited to, loading points, unloading points and transfer stations; the driving environment parameter over-limit time is the total time during which key environmental parameters exceed the preset safety threshold during the logistics process, wherein the preset safety threshold is set as needed and is not specifically limited.

[0056] In this embodiment, a weighted summation method is used to calculate the comprehensive score. That is, a weight is assigned to each of the following: transportation time, transportation energy consumption, time window deviation, and time exceeding the limit of driving environment parameters. The standardized values ​​of transportation time, transportation energy consumption, time window deviation, and time exceeding the limit of driving environment parameters are multiplied by the corresponding weights and summed to obtain the comprehensive score. The weights are set as needed and are not specifically limited.

[0057] In this embodiment, the comprehensive scores of all progeny DNAs are sorted in descending order, and the progeny DNA with the highest score is selected as the candidate master DNA.

[0058] Feasibility verification of candidate master DNA can be performed by using hard constraint checks, such as verifying whether the paths in the candidate master DNA contain prohibited sections, height-restricted or weight-restricted sections, or checking whether their time windows conflict with the availability of logistics vehicles or drivers and whether they meet the specific storage or operation requirements of the goods; resource matching verification can also be performed, such as confirming whether the target node has sufficient loading and unloading capacity, parking spaces or transit storage space, and whether there are suitable vehicle types available at the planned arrival time.

[0059] If the candidate master DNA fails the feasibility check, all progeny DNAs are sorted from highest to lowest according to their comprehensive scores. The next progeny DNA is then taken from the sorted list and its feasibility is checked. When a progeny DNA passes the feasibility check, it is designated as the master DNA. If all progeny DNAs fail the feasibility check, an alert is issued to prompt staff to intervene manually or select a default contingency plan.

[0060] S105 generates a dynamic electronic fence based on the master DNA. The dynamic electronic fence includes dynamic path, dynamic time window and dynamic driving environment parameter threshold.

[0061] Specifically, the historical driving path, historical time window, and historical driving environment parameter thresholds corresponding to each transportation segment are extracted from the master DNA; the historical driving paths of all transportation segments are connected in execution order to obtain the dynamic path of the transportation process; for each transportation segment, the start and end times of the historical time window are extended forward and backward by a second preset duration, respectively, to obtain the dynamic time window; for each transportation segment, the lower and upper limits of the historical driving environment parameter thresholds are floated downward and upward by a preset tolerance range, respectively, to obtain the dynamic driving environment parameter thresholds; the dynamic path, dynamic time window, and dynamic driving environment parameter thresholds are used as a dynamic electronic fence.

[0062] In this embodiment, the master DNA stores only the identifier of the transport segment, and the identifier is associated with an external database or knowledge base. This database or associated knowledge base stores the historical driving path, historical time window, and historical driving environment parameter threshold corresponding to each transport segment. The historical driving path, historical time window, and historical driving environment parameter threshold can be queried and extracted through the identifier.

[0063] In this embodiment, the historical travel path of each transportation segment is converted into a series of geographic coordinate points, such as a GPS point sequence. These geographic coordinate points are connected end to end according to the order of the transportation segments to form a continuous global path, i.e., a dynamic path.

[0064] In this embodiment, the start time of the historical time window is advanced and the end time is postponed to form a broad dynamic time window. This allows logistics vehicles to arrive at / depart from the node earlier or later within a certain range without immediately triggering an abnormal alarm. The second preset duration can be a base duration, combined with the actual length of the transportation segment and the remaining total time of the current logistics task, to calculate the specific preset duration using a function, such as a linear function or a piecewise function. For example, ΔT = 0.5L + 0.1T r +30, where L is the actual length of the transport segment, T r ΔT represents the remaining total time, and ΔT represents the preset duration.

[0065] In this embodiment, the preset tolerance range is the allowable fluctuation range of thresholds such as cargo temperature, weather visibility, and traffic flow. The basic tolerance is preset according to the cargo sensitivity. For example, cargo sensitivity includes: high sensitivity: temperature ±1℃, visibility ±50 meters, traffic flow ±100 vehicles / hour. The basic tolerance is adjusted according to the transportation stage. For example, it is divided according to the proportion of completed mileage to total mileage: 30% in the initial stage, 30% to 70% in the middle stage, and 70% in the final stage. The adjustment coefficients are 100%, 70%, and 50% respectively. That is, the actual tolerance is equal to the basic tolerance multiplied by the corresponding adjustment coefficient. The cargo sensitivity is determined by cargo type mapping or industry standards.

[0066] Dynamic paths, dynamic time windows, and dynamic driving environment parameter thresholds are used as dynamic electronic fences. These dynamic paths, dynamic time windows, and dynamic driving environment parameter thresholds are integrated to form a complete dynamic electronic fence. The boundaries of the electronic fence are then adjusted based on historical experience, cargo type, and real-time status.

[0067] S106: Real-time acquisition of current state data and future state trajectory prediction data, and calculation of the deviation vector between the future state trajectory prediction data and the dynamic electronic fence.

[0068] Current status data can be obtained in various ways. For example, it can be collected in real time by vehicle sensors, GPS modules, and on-board diagnostic systems to collect operational data of logistics vehicles, such as current vehicle location, current time, current speed, and current remaining fuel or battery level. It can also be obtained by connecting with data interfaces of external data sources, such as real-time traffic information systems of traffic management departments and weather forecast systems of meteorological bureaus, to obtain real-time current driving environment parameters, such as traffic congestion, weather conditions, and road construction information. Future trajectory prediction data is calculated by prediction models based on current status data and historical data. For example, based on the vehicle's current speed and direction, its position in the next few minutes can be predicted.

[0069] The calculation of the deviation vector between the future state trajectory prediction data and the dynamic electronic fence includes: aligning the future state trajectory prediction data and the dynamic electronic fence in the same spatiotemporal coordinate system to obtain the aligned predicted position sequence, the predicted time sequence of arrival at each node, and the predicted driving environment parameter change curve; calculating the spatial deviation component based on the predicted position sequence and the dynamic path in the dynamic electronic fence; calculating the temporal deviation component based on the predicted time sequence of arrival at each node and the dynamic time window in the dynamic electronic fence; calculating the environmental deviation component based on the predicted driving environment parameter change curve and the dynamic driving environment parameter threshold in the dynamic electronic fence; and using the spatial deviation component, temporal deviation component, and environmental deviation component as the deviation vector of the dynamic electronic fence.

[0070] In this embodiment, aligning the future trajectory prediction data and the dynamic electronic fence in the same spatiotemporal coordinate system can be achieved in various ways. For example, timestamp alignment and geographic coordinate system transformation can be used, such as converting all geographic location data to WGS84 or UTM coordinate systems to ensure that the future trajectory prediction data and the dynamic electronic fence data have consistent reference points in time and space. For data that is not collected or generated synchronously, time interpolation, such as linear interpolation, spline interpolation, and spatial mapping, can also be used, such as mapping the predicted trajectory points to the nearest points or segments on the dynamic electronic fence path to complete the alignment, thereby obtaining the aligned predicted position sequence, the predicted time sequence of arrival at each node, and the predicted driving environment parameter change curve.

[0071] In this embodiment, for calculating the spatial deviation component based on the predicted location sequence and the dynamic path in the dynamic electronic fence, the Euclidean distance or geodesic distance from each point in the predicted location sequence to the nearest point on the dynamic path is calculated, and these Euclidean distances or geodesic distances are aggregated, for example, by calculating the average, maximum or integral value, and the corresponding average, maximum or integral value is used as the spatial deviation component. Alternatively, the predicted location sequence and the dynamic path can be represented as geographical regions, for example, by using buffer analysis to calculate the overlap or non-overlapping area of ​​the two regions, thereby measuring the spatial deviation.

[0072] In this embodiment, the time deviation component is calculated based on the predicted arrival time series of each node and the dynamic time window in the dynamic electronic fence. The predicted arrival time of each node is compared with the start time or end time of the dynamic time window. If the predicted time exceeds the dynamic time window, the excess amount is calculated, such as the advance amount or the delay amount. These excess amounts are aggregated, such as by summing or finding the maximum value, and the corresponding summation value or maximum value is used as the time deviation component.

[0073] In this embodiment, for calculating the environmental deviation component based on the predicted driving environment parameter change curve and the dynamic driving environment parameter threshold in the dynamic electronic fence, each sampling point on the predicted driving environment parameter change curve is compared with the upper and lower limits of the dynamic driving environment parameter threshold. If the predicted value exceeds the threshold range, the excess amount or the excess duration is recorded, and these excess information are aggregated, for example, by summing or finding the maximum value, and the corresponding summation value or maximum value is used as the environmental deviation component.

[0074] In this embodiment, the spatial deviation component, temporal deviation component, and environmental deviation component are directly combined into a multi-dimensional vector, for example, [spatial deviation value, temporal deviation value, environmental deviation value]. Different weights can also be assigned to the spatial deviation component, temporal deviation component, and environmental deviation component, and the weighted components are combined into a vector, for example, [spatial deviation component weight * spatial deviation value, temporal deviation value weight * temporal deviation value, environmental deviation value weight * environmental deviation value]. The weights are set as needed and are not specifically limited.

[0075] S107, if the deviation vector meets the preset conditions, determine the gene fragment to be mutated in the main DNA based on the deviation vector, and perform mutation processing on the gene fragment to be mutated to obtain multiple variant DNAs.

[0076] In this embodiment, the preset condition is a threshold restriction condition for each component in the deviation vector, wherein the preset condition includes at least one of the following: For spatial conditions, the spatial deviation component exceeds a preset spatial threshold, for example, an average deviation distance greater than 50 meters or a maximum deviation distance greater than 100 meters; for temporal conditions, the temporal deviation component exceeds a preset time threshold, for example, a cumulative delay time exceeding 10 minutes or a single node delay exceeding 5 minutes; for environmental conditions, the environmental deviation component exceeds a preset environmental threshold, for example, cargo temperature exceeding the allowable range for more than 2 minutes cumulatively, or visibility below the lower limit for 30 seconds; for comprehensive conditions: the spatial deviation value, temporal deviation value, and environmental deviation value are weighted and summed to obtain a comprehensive deviation score. When the comprehensive deviation score exceeds a preset comprehensive threshold, it is determined that the condition is met. For example, comprehensive deviation score = 0.4 × spatial deviation + 0.4 × temporal deviation + 0.2 × environmental deviation, and the preset comprehensive threshold is 60 points.

[0077] The process of determining the gene fragments to be mutated in the main DNA based on the deviation vector includes: for the spatial deviation component, identifying the spatial deviation points between the predicted position sequence in the future state trajectory prediction data and the dynamic path in the dynamic electronic fence, and using the transportation segment corresponding to the spatial deviation point as the first candidate gene fragment; for the temporal deviation component, identifying the time-limited nodes in the future state trajectory prediction data where the predicted arrival time of each node exceeds the dynamic time window in the dynamic electronic fence, and using the transportation segment corresponding to the time-limited node as the second candidate gene fragment; for the environmental deviation component, identifying the environmental excess period where the predicted change curve of the driving environment parameter in the future state trajectory prediction data exceeds the threshold of the dynamic driving environment parameter in the dynamic electronic fence, and using the transportation segment corresponding to the environmental excess period as the third candidate gene fragment; and merging and deduplicating the first, second, and third candidate gene fragments to obtain the set of gene fragments to be mutated.

[0078] Among them, the predicted position sequence in the future state trajectory prediction data is a set of a series of geographical coordinates that the logistics vehicle may pass through in the future period of time, determined based on the current state data and the prediction model. The predicted position sequence is the expected driving path of the vehicle in the future period. In this embodiment, it can be generated by combining GPS data, inertial navigation system data, historical driving data and real-time traffic information through prediction algorithms, such as Kalman filtering and deep learning models, or obtained by simulating the motion trajectory of the logistics vehicle in a digital twin ring model. The dynamic path in the dynamic electronic fence is a set of geographical areas or routes that are allowed for vehicles to travel in the logistics transportation process, determined according to the master DNA. The dynamic path is a series of connected geographical coordinates, a network composed of multiple path segments, or a virtual channel with a certain width and height.

[0079] Spatial deviation points are the spatial distances between the predicted position sequences in the future trajectory prediction data and the dynamic paths in the dynamic electronic fences, exceeding a preset tolerance range. These geographical locations represent the positions where logistics vehicles will deviate from or have already deviated from the planned paths in the future. In this embodiment, spatial deviation points are identified by calculating the minimum distance from each point in the predicted position sequence to the dynamic path and comparing it with a preset threshold. The preset threshold is determined as needed and is not specifically limited. The first candidate gene segment is the transportation segment directly associated with the detected spatial deviation point. When a spatial deviation point appears in a transportation segment, that transportation segment is marked as the first candidate gene segment. For example, the transportation segment to which the spatial deviation point belongs can be found by reverse tracing based on its geographical location.

[0080] The predicted arrival time series for each node is based on future state trajectory prediction data, predicting the specific time set when logistics vehicles will arrive at or leave each key node in the future. This time series represents the expected time progress of the vehicle within a future period. In this embodiment, it is generated by combining predicted location sequences with predicted speed, road condition information, and node operation time through a time extrapolation model. It can also be obtained by simulating the vehicle's travel and dwell time between nodes using digital twin simulation. The dynamic time window in the dynamic electronic fence is the time range within which vehicles are allowed to arrive at or leave each node during logistics transportation, determined by the master DNA. It is represented by the start and end time corresponding to each node, or a time interval within which vehicles are allowed to operate. Time-exceeding nodes are future state trajectories. The predicted arrival time series of each node in the prediction data includes nodes whose predicted arrival or departure times exceed the dynamic time window in the dynamic electronic fence. These nodes indicate that vehicles will be delayed or arrive early in the future, which does not conform to the time plan. The nodes that exceed the time limit can be identified by comparing the predicted arrival time with the upper and lower limits of the dynamic time window to determine whether it exceeds the preset range; or by calculating the difference between the predicted time and the time window boundary. If the difference exceeds zero, it is considered to exceed the limit. The second candidate gene segment is the transportation segment directly associated with the detected time-exceeding node. When the end node or start node of a transportation segment is identified as a time-exceeding node, the transportation segment is marked as the second candidate gene segment. For example, the transportation segments that serve as the start or end point can be found in reverse based on the time-exceeding node.

[0081] The predicted driving environment parameter change curve is based on current state data and a prediction model to predict the trend of driving environment parameters experienced by logistics vehicles over a future period. In this embodiment, the predicted driving environment parameter change curve is generated by combining sensor data, weather forecast data, and road condition information with an environmental prediction model. For example, a time series prediction model or machine learning model can be used to generate predicted environmental parameter values ​​for a future period with a fixed time step, thereby generating an original prediction curve with time as the horizontal axis and predicted environmental parameter values ​​as the vertical axis. The time series prediction model can use an autoregressive integral moving average model or a prophetic model, where the fixed time step can be 5 minutes or 10 minutes. The machine learning model can use LSTM or random forest. Alternatively, it can be obtained by simulating environmental changes of vehicles under different road sections and times through digital twin simulation. The dynamic driving environment parameter threshold in the dynamic electronic fence is the range of environmental parameters that vehicles are allowed to be in during logistics transportation, determined based on the master DNA. The upper and lower limits of environmental parameters corresponding to each transport segment, or a safe range that allows fluctuations in driving environmental parameters, are defined as needed and are not specifically limited. Environmental exceedance periods are predicted driving environmental parameter change curves in future state trajectory prediction data. These periods exceed the dynamic driving environmental parameter thresholds in the dynamic electronic fence, indicating that the vehicle may face environmental conditions that do not meet cargo requirements in the future. In this embodiment, environmental exceedance periods are identified by comparing the upper and lower limits of the predicted driving environmental parameters with the dynamic driving environmental parameter thresholds to determine whether they exceed a preset range. The third candidate gene fragment is a transport segment directly associated with the detected environmental exceedance periods. When an environmental exceedance period occurs within a transport segment, that transport segment is marked as the third candidate gene fragment, indicating that the environmental conditions of that transport segment may not meet requirements and need to be adjusted through mutation. For example, the transport segment to which the environmental exceedance period belongs can be found by reverse-engineering the transport time corresponding to the environmental exceedance period.

[0082] In this embodiment, the deduplication process involves integrating the first candidate gene fragment, the second candidate gene fragment, and the third candidate gene fragment identified from the spatial deviation component, the temporal deviation component, and the environmental deviation component, respectively, and eliminating duplicate transport segments. For example, the deduplication operation is automatically completed by putting the unique identifier transport segment ID of all candidate gene fragments into a set, thereby obtaining a set of all transport segments that need to be mutated.

[0083] The process involves: mutating the gene fragment to be mutated to obtain multiple variant DNAs, including: selecting multiple mutation operators from a pre-defined mutation operator library based on the type of the gene fragment to be mutated and the deviation direction of the deviation vector; the mutation operator library includes path replacement operators for adjusting the path, time translation or scaling operators for adjusting the time window, and threshold scaling operators for adjusting the threshold of driving environment parameters; processing the gene fragment to be mutated sequentially based on the selected mutation operators to obtain multiple first-type variant DNAs, wherein only one mutation operator is used for each processing of the gene fragment to be mutated; generating multiple second-type variant DNAs based on randomly combined mutation operators and gene fragments to be mutated; and summing the first-type and second-type variant DNAs to obtain multiple variant DNAs.

[0084] Among them, the type of gene fragment to be mutated indicates whether the current deviation mainly occurs in the dimensions of path, time, or environmental parameters. For example, the transportation segment corresponding to the spatial deviation point, the transportation segment corresponding to the time limit exceedance node, or the transportation segment corresponding to the environmental limit exceedance period. The deviation direction of the deviation vector refines the adjustment requirements. For example, whether the path deviation is to the left or right, whether the time limit exceedance is earlier or later, and whether the environmental parameter exceeds the upper limit or is lower than the lower limit.

[0085] The preset mutation operator library includes various mutation operators for different types of deviations. For example, the path replacement operator is used to adjust the transportation route. In this embodiment, multiple alternative routes are provided between the starting and ending nodes of the current transportation segment by querying map services or historical data, or new feasible routes are generated based on real-time traffic prediction. The time translation or scaling operator is used to adjust the time window. In this embodiment, the time window of the current transportation segment is moved forward or backward by a third preset duration, or the duration of the time window is adjusted to make it longer or shorter. The threshold scaling operator is used to adjust the threshold of driving environment parameters. In this embodiment, the upper limit or lower limit of a certain driving environment parameter, such as temperature or humidity, is increased or decreased. The third preset duration, duration, upper limit, and lower limit are set as needed and are not specifically limited.

[0086] For selecting multiple mutation operators from a preset mutation operator library based on the type of gene fragment to be mutated and the deviation direction of the deviation vector, in this embodiment, when path deviation is detected, the path replacement operator is selected first; when time limit is detected, the time translation or scaling operator is selected. The selection method can be based on preset rule matching, or it can be recommended by a machine learning model based on historical data and deviation characteristics.

[0087] When generating the first type of variant DNA, the gene fragments to be mutated are processed sequentially based on the selected mutation operators, and only one mutation operator is used for each processing. For example, for a transport segment with a deviated path, a variant is first generated using the path replacement operator, and then another variant is generated using the time translation operator, thereby ensuring that each first type of variant DNA represents a single adjustment scheme.

[0088] In this embodiment, two or more mutation operators are randomly selected and combined and applied to the same gene fragment to be mutated, or a random combination of mutation operators is applied to different gene fragments to be mutated. For example, a path substitution operator and a time shift operator are applied simultaneously to generate a mutant DNA.

[0089] It should be noted that when multiple consecutive transport segments are marked as gene fragments to be mutated, they are merged into a combined gene fragment for joint mutation to avoid inconsistencies caused by fragmented mutation.

[0090] S108. Based on digital twin technology and current state data, each variant DNA is simulated to obtain a second simulation result, and one of the variant DNAs is selected as the optimal variant DNA based on the second simulation result.

[0091] The current state data includes the current vehicle location, current time, current speed, current remaining fuel or battery power, and current driving environment parameters. Based on digital twin technology and the current state data, simulation is performed on each variant DNA. This includes: starting from the current state data, using digital twin technology to simulate the remaining logistics operation sequence of each variant DNA from the current time to the end of the task, and obtaining the second simulation result. The second simulation result includes the predicted driving path of the remaining transportation segment, the predicted time to reach each node, the predicted time consumption of each transportation segment, the predicted energy consumption, the predicted curve of the change of driving environment parameters, and the conformity indicator of whether each transportation segment meets the dynamic time window and driving environment parameter threshold.

[0092] In this embodiment, in the digital twin model, the unexecuted transport segment sequence in the mutant DNA is identified based on the current time and the current vehicle position, and this is used as the starting point for simulation. Based on real-time data such as current speed, remaining fuel or battery power, combined with predicted future road conditions and environmental changes, the entire process of the logistics vehicle completing the remaining tasks is simulated, thereby obtaining the second simulation result.

[0093] Among them, one of the variant DNAs is selected as the optimal variant DNA based on the second simulation results. This includes: calculating the corrected fitness of each variant DNA based on the predicted travel path of the remaining transport segments, the predicted arrival time at each node, the predicted travel time of each transport segment, the predicted energy consumption, the predicted change curve of the travel environment parameters, and the conformity identifier and corresponding weight value of each transport segment in meeting the dynamic time window and the threshold of the travel environment parameters; and selecting the variant DNA with the highest corrected fitness as the optimal variant DNA.

[0094] In this embodiment, the predicted driving path, the predicted driving environment parameter change curve, the predicted arrival time at each node, and the compliance flag are quantified. The driving path is converted into the total length of the driving path; the predicted driving environment parameter change curve is converted into a deviation integral value, that is, the cumulative amount of deviation between the predicted driving environment parameter change curve and the corresponding threshold; the compliance flag is converted into a compliance score, where compliance is 1 and non-compliance is 0, and the compliance flag of the dynamic time window of the remaining transportation segment of each variant DNA and the driving environment parameter threshold are summed to obtain the compliance flag sum value; the predicted arrival time at each node is converted into the total deviation time relative to the dynamic time window, that is, the sum of the absolute values ​​of the delay between the actual arrival time of each node and the expected time window.

[0095] In this embodiment, the total length of the travel route, the integral value of the deviation, the sum of the conformity flags, the sum of the absolute values ​​of the delays, the predicted time for each transport segment, the predicted energy consumption, and the corresponding weight values ​​are weighted and summed to obtain the corrected fitness. The weight values ​​are set as needed and are not specifically limited. The weight values ​​can also be adjusted according to the priority of the logistics task. For example, for tasks with high timeliness requirements, the weight of the time window is higher; for tasks with high cargo sensitivity, the weight of the driving environment parameter threshold is higher. Then, the standardized scores of each variable are multiplied by their corresponding weights and summed to obtain the corrected fitness.

[0096] In this embodiment, the corrected fitness values ​​calculated for all mutant DNAs are sorted, and the mutant DNA with the highest fitness value is selected. In some cases, if the corrected fitness values ​​of multiple mutant DNAs are very close, secondary screening conditions can be used, such as selecting the mutant DNA with the lowest energy consumption or the shortest path, to further optimize the selection results and ensure that the optimal mutant DNA finally selected can maximize the efficiency and safety of logistics transportation.

[0097] S109 updates the master DNA based on the optimal variant DNA, obtains the update result, and updates the dynamic electronic fence based on the update result.

[0098] In this embodiment, the optimal mutant DNA is used as the new master DNA, and a new dynamic electronic fence is generated using step S105 and the new master DNA. For example, if the optimal mutant DNA contains a new path, the dynamic path in the dynamic electronic fence will be updated accordingly.

[0099] The terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0100] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing application concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions claimed in this application.

Claims

1. A logistics transportation method based on electronic fences, characterized in that, include: Get pending logistics tasks; At least two process DNAs that match the logistics task to be processed are selected from the gene bank as parent DNAs. The gene bank includes process DNAs encoded by multiple successfully completed historical logistics tasks, and each process DNA represents a logistics operation sequence. The parental DNA was subjected to genetic processing to obtain multiple offspring DNAs; Based on digital twin technology, each of the daughter DNAs is simulated to obtain a first simulation result, and one of the daughter DNAs is selected as the master DNA based on the first simulation result. A dynamic electronic fence is generated based on the master DNA, and the dynamic electronic fence includes a dynamic path, a dynamic time window, and dynamic driving environment parameter thresholds. Real-time acquisition of current state data and future state trajectory prediction data, and calculation of the deviation vector between the future state trajectory prediction data and the dynamic electronic fence; When the deviation vector meets the preset conditions, the gene fragment to be mutated in the main DNA is determined based on the deviation vector, and the gene fragment to be mutated is subjected to mutation processing to obtain multiple variant DNAs; Based on the digital twin technology and current state data, each of the mutant DNAs is simulated to obtain a second simulation result, and one of the mutant DNAs is selected as the optimal mutant DNA based on the second simulation result; The master DNA is updated based on the optimal variant DNA to obtain the update result, and the dynamic electronic fence is updated based on the update result.

2. The logistics transportation method based on electronic fence according to claim 1, characterized in that, The construction of the gene bank includes: Acquire historical information of multiple successfully completed historical logistics tasks. The historical information includes cargo type, historical travel route, historical arrival and departure time of each node, historical travel environment parameters of each transportation segment, and historical operation events of each node. Based on the task execution order, each historical logistics task is divided into multiple transportation segments. Each transportation segment includes a starting node, an ending node, historical transportation actions between the starting node and the ending node, and corresponding historical operation events. The historical transportation actions include historical driving routes, historical driving time, and historical driving environment parameters. Based on preset coding rules, all transportation segments of each historical logistics task are sequentially arranged to obtain the process DNA. The process DNA is a transportation segment sequence, each gene fragment in the process DNA corresponds to a transportation segment, and each process DNA corresponds to a cargo type. The gene library is generated based on all process DNA, and a success evaluation index is generated for each process DNA, wherein the success evaluation index includes whether it is on time, whether cargo damage occurs, and whether energy consumption is lower than a preset standard.

3. A logistics transportation method based on electronic fences according to claim 1 or 2, characterized in that, The step of selecting at least two process DNAs from the gene bank that match the logistics task to be processed as parent DNAs includes: Obtain the task start node, task end node, task cargo type, current driving environment parameters, and expected delivery time of the logistics task to be processed; The total mileage of the logistics task to be processed is calculated based on the task start node and the task end node, and the expected total time is calculated based on the expected delivery time and the current time. Select process DNA from the gene bank that matches the task start node, task end node, and cargo type; For each selected process DNA, the similarity between the process DNA and the logistics task to be processed is calculated based on the total mileage, expected total time, current driving environment parameters, historical total mileage, historical total time, and historical driving environment parameters. Select the N process DNAs with the highest similarity as the candidate set, where N≥3; Two process DNAs are randomly selected from the candidate set as parental DNAs, wherein the probability of random selection is positively correlated with the success evaluation index of each candidate DNA.

4. The logistics transportation method based on electronic fence according to claim 1, characterized in that, The genetic treatment of the parental DNA includes: Obtain the transport segment sequence of the parental DNA; Select at least one cutting position, and cut the transport segment sequences of the two parental DNAs based on the cutting position. Exchange the cut tail transport segment fragments and splice them together to obtain two first daughter DNAs. For each first progeny DNA, at least one transport segment is selected, and the parameters of the selected transport segment are adjusted to obtain multiple second progeny DNAs. All first-generation and second-generation DNA were combined to obtain multiple daughter DNAs; The parameter adjustment includes at least one of the following methods: Replace the historical driving route of the transportation segment with another feasible route on the map that connects the same starting node and ending node; simultaneously advance or postpone the start and end times of the historical time window of the transportation segment by a first preset duration; and adjust the threshold values ​​of the historical driving environment parameters of the transportation segment by a preset value.

5. A logistics transportation method based on electronic fences according to claim 1, characterized in that, The generation of a dynamic electronic fence based on the master DNA includes: Extract the historical driving path, historical time window, and historical driving environment parameter thresholds corresponding to each transport segment from the master DNA. Connect the historical travel routes of all transportation segments in the order of execution to obtain the dynamic path of the transportation process; For each transport segment, the start and end times of the historical time window are extended forward and backward by a second preset duration, respectively, to obtain a dynamic time window; For each transportation segment, the lower and upper limits of the historical driving environment parameter thresholds are adjusted downward and upward by a preset tolerance range to obtain the dynamic driving environment parameter thresholds. The dynamic path, dynamic time window, and dynamic driving environment parameter thresholds are used as the dynamic electronic fence.

6. A logistics transportation method based on electronic fences according to claim 1, characterized in that, The calculation of the deviation vector between the future state trajectory prediction data and the dynamic electronic fence includes: The future state trajectory prediction data and the dynamic electronic fence are aligned in the same spatiotemporal coordinate system to obtain the aligned predicted position sequence, the predicted time sequence of arrival at each node, and the predicted driving environment parameter change curve. Calculate the spatial deviation component based on the predicted location sequence and the dynamic path in the dynamic electronic fence; Calculate the time deviation component based on the predicted arrival time series of each node and the dynamic time window in the dynamic electronic fence; Calculate the environmental deviation component based on the predicted driving environment parameter change curve and the dynamic driving environment parameter threshold in the dynamic electronic fence; The spatial deviation component, temporal deviation component, and environmental deviation component are used as the deviation vector of the dynamic electronic fence.

7. A logistics transportation method based on electronic fences according to claim 6, characterized in that, Determining the gene fragment to be mutated in the main DNA based on the deviation vector includes: For the spatial deviation component, identify the spatial deviation point between the predicted position sequence in the future state trajectory prediction data and the dynamic path in the dynamic electronic fence, and take the transport segment corresponding to the spatial deviation point as the first candidate gene fragment. For the time deviation component, identify the time-exceeding nodes in the future state trajectory prediction data whose predicted arrival time series exceeds the dynamic time window in the dynamic electronic fence, and use the transportation segment corresponding to the time-exceeding node as the second candidate gene fragment. For the environmental deviation component, identify the environmental over-limit period when the predicted driving environment parameter change curve in the future state trajectory prediction data exceeds the dynamic driving environment parameter threshold in the dynamic electronic fence, and take the transportation segment corresponding to the environmental over-limit period as the third candidate gene segment. The first candidate gene fragment, the second candidate gene fragment, and the third candidate gene fragment are merged and deduplicated to obtain a set of gene fragments to be mutated.

8. A logistics transportation method based on electronic fences according to claim 1, characterized in that, The gene fragment to be mutated is subjected to mutation treatment to obtain multiple variant DNAs, including: Based on the type of the gene fragment to be mutated and the deviation direction of the deviation vector, multiple mutation operators are selected from a preset mutation operator library. The mutation operator library includes a path replacement operator for adjusting the path, a time translation or scaling operator for adjusting the time window, and a threshold scaling operator for adjusting the threshold of driving environment parameters. Based on the selected mutation operator, the gene fragment to be mutated is processed sequentially to obtain multiple first-class variant DNAs. Each time the gene fragment to be mutated is processed, only one mutation operator is used. Multiple second-type variant DNAs are generated based on a randomly combined mutation operator and the gene fragment to be mutated; The first type of variant DNA and the second type of variant DNA were combined to obtain multiple variant DNAs.

9. A logistics transportation method based on electronic fences according to claim 1, characterized in that, The first simulation results include transportation time, transportation energy consumption, time window deviation, and duration of exceeding driving environment parameters. Based on the first simulation results, one of the offspring DNAs is selected as the master DNA, including: Based on the transportation time, transportation energy consumption, time window deviation, and the duration of exceeding the driving environment parameters, a comprehensive score is calculated for each offspring DNA, and the offspring DNA with the highest score is selected as the candidate master DNA. The candidate master DNA is subjected to feasibility verification. If the candidate master DNA passes the feasibility verification, it is adopted as the master DNA. If the candidate master DNA fails the feasibility verification, the progeny DNAs are verified in descending order of comprehensive score until the master DNA is determined.

10. A logistics transportation method based on electronic fences according to claim 1, characterized in that, The second simulation results include the predicted travel path for the remaining transport segments, the predicted arrival time at each node, the predicted travel time for each transport segment, the predicted energy consumption, the predicted curves of changes in travel environment parameters, and a conformity indicator for each transport segment to meet the dynamic time window and travel environment parameter thresholds. The selection of one variant DNA as the optimal variant DNA based on the second simulation results includes: Based on the predicted travel path of the remaining transport segments, the predicted arrival time at each node, the predicted travel time of each transport segment, the predicted energy consumption, the predicted change curve of the travel environment parameters, and the conformity identifier and corresponding weight value of each transport segment in terms of whether it meets the dynamic time window and the threshold of the travel environment parameters, the corrected fitness of each variant DNA is calculated. The mutant DNA with the highest modified fitness was selected as the optimal mutant DNA.

Images