Vehicle scheduling optimization system and method for vehicle renting platform

Abstract

The invention relates to the technical field of scheduling optimization, in particular to a car rental platform car scheduling optimization system and method, and the system comprises a rental saturation recognition module, a scheduling level setting module, a candidate path reconstruction module, an adjustable car screening module and a scheduling path binding module. According to the invention, through analyzing the time and geographic data of the car rental order, accurate determination of the vehicle usage condition is realized, vehicle distribution is optimized, through continuous monitoring and evaluation of the vehicle usage saturation state in the region, a high-demand region can be identified, vehicle distribution is correspondingly adjusted, optimal configuration of resources is ensured, and the vehicle rental efficiency is improved. The scheduling path is intelligently reconstructed according to the actual operation data and the scheduling history of the vehicle, the empty driving time of the vehicle is reduced, the scheduling efficiency is improved, the waiting time of a client is shortened, the operation cost is reduced, and the timeliness of service and the satisfaction of the client are guaranteed by analyzing the emergency degree of a task and preferentially processing an order with high emergency degree.

Description

Technical Field

[0001] The present invention relates to the technical field of scheduling optimization, and in particular to a vehicle scheduling optimization system and method for a car rental platform. Background Art

[0002] The field of scheduling optimization technology involves efficiently matching and allocating resources, tasks, and constraints, with the goal of improving overall system efficiency, reducing resource waste, and lowering costs. This field is widely used in scenarios such as transportation, warehousing and logistics, manufacturing scheduling, and personnel scheduling. Common methods include linear programming, integer programming, heuristic algorithms, multi-objective optimization, graph-based path planning, and reinforcement learning scheduling strategies. In dynamic and complex environments, this technology requires a combination of real-time data, constraint update mechanisms, and feedback-based closed-loop scheduling schemes to support the rapid generation and deployment of optimal or suboptimal decisions.

[0003] The vehicle scheduling optimization system for car rental platforms applies scheduling optimization methods to systematically schedule and allocate vehicle resources within the platform. The system's purpose is to automatically generate reasonable vehicle allocation and scheduling routes based on parameters such as user rental demand, geographic distribution, vehicle availability, return time, and vehicle type. This addresses issues such as high idle vehicle rates, supply-demand mismatches, and response delays, ultimately improving the platform's vehicle utilization and service quality.

[0004] Traditional dispatching systems lack flexibility and have slow response times when dealing with dynamic changes and complex constraints. This is especially true during peak periods of car rental demand, when it is difficult for traditional systems to update vehicle status and customer needs in real time. This leads to irrational resource allocation and high idle vehicle rates. Traditional systems lack refined management of regional vehicle allocation and are unable to effectively identify and resolve supply and demand contradictions, resulting in vehicle surpluses or shortages that affect service quality and customer experience. In terms of emergency order processing, there is a lack of effective forecasting and dispatching mechanisms, resulting in a slow system response that cannot meet customer demands for rapid responses, affecting overall operational efficiency and market competitiveness, limiting the performance of the dispatching system under high-pressure environments, and leading to increased operating costs and reduced customer satisfaction. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a vehicle scheduling optimization system and method for a car rental platform.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solution: a vehicle scheduling optimization system for a car rental platform, the system comprising:

[0007] The rental saturation identification module obtains order records from the car rental platform, maps the rental duration of each order to the time period, evaluates the saturation status of the area within the time period, and generates a set of regional continuous rental saturation identifiers;

[0008] The dispatching level setting module calls the regional continuous rental saturation identification set, extracts the area marked as a high-density rental state, obtains the number of available vehicles at each rental station in the area, combines the vehicle gap and the high-frequency index value, calculates the dispatching level of the station, and obtains the regional overload node dispatching level table;

[0009] The candidate path reconstruction module normalizes the average dispatch response time and the empty distance index of the vehicles based on the regional overload node dispatch level table, performs weighted calculation and sorting, selects the optimal path for executing the dispatch task, and generates a path transferable vehicle dispatch group;

[0010] The adjustable vehicle screening module obtains orders marked as high urgency in the task list, extracts the vehicle's previous task completion time, the average of the last three call intervals, and the standard deviation of the historical response time, and compares them with the maximum response time limit of the order task. If any indicator exceeds the response limit standard, the vehicle number is eliminated to obtain a list of vehicles matching high-urgency tasks.

[0011] As a further solution of the present invention, the regional continuous rental saturation identification set includes a vehicle occupancy ratio label for a time period, an area number index, a continuous saturation state mark and an adjustable vehicle upper limit for the corresponding time period; the regional overload node scheduling level table specifically includes a site vehicle gap value, a historical high occupancy frequency factor, a scheduling level index and a site number label; the path transferable vehicle scheduling group includes a transfer-out node number, a target site path pair, an empty driving distance ratio and a scheduling response time interval; the high-urgency task matching vehicle list specifically refers to a task number label, a vehicle response interval index, a task response limit status and a vehicle scheduling feasibility identifier.

[0012] As a further solution of the present invention, the rental saturation identification module includes:

[0013] The order-period mapping submodule obtains order records from the car rental platform, extracts the rental start and end times, the rental vehicle number, and the geographic region label, maps the duration of each order to the corresponding time period and geographic region, and establishes a dual index structure of region and time period to generate a region-period mapping dataset;

[0014] The occupancy index calculation submodule calls the regional time period mapping dataset, extracts the number of orders and the average usage time of orders in each time period of each region, obtains the vehicle dispatch capacity of the region corresponding to the time period, calculates the occupancy saturation score of the region time period, and constructs a scoring structure based on the region as a unit to obtain the regional occupancy score matrix;

[0015] The formula for calculating the occupancy saturation score of the area time period is:

[0016]

[0017] Among them, S ij represents the occupancy saturation score of region i in time period j, T ijk is the normalized value of the rental duration of the k-th order in area i and time period j, n ij is the total number of orders in region i during time period j, C ij is the normalized value of vehicle dispatch capacity configured for region i in time period j, R ij is the normalized value of the total amount of scheduling resources in region i in time period j, M ij is the normalized value of the average task resource demand of region i in time period j;

[0018] The saturation state determination submodule compares each score value with the scheduling load reference threshold set by the platform based on the area occupancy score matrix, marks the area and time period combinations whose score values exceed the reference threshold for more than three consecutive times, and merges them into interval labels to generate an area continuous rental saturation identification set.

[0019] As a further solution of the present invention, the scheduling level setting module includes:

[0020] The high-density area extraction submodule calls the area continuous lease saturation identification set, filters the area numbers marked as saturated, establishes the area site mapping index, and generates a high-density area label set;

[0021] The resource gap identification submodule extracts the number of available vehicles for each rental station in the corresponding area based on the high-density area tag set and the recorded area number, obtains the minimum number of operating vehicles set by the platform, performs difference calculation, filters the station numbers with insufficient available numbers, and calculates their gap values to obtain the vehicle resource gap set;

[0022] The level weight calculation submodule calls the vehicle resource gap set, extracts the high-density occurrence frequency of the area where the station is located in the last three scheduling cycles, performs superimposed proportion scoring based on the frequency value and the corresponding gap amount, calculates the scheduling level index value of the station, constructs the station level mapping structure, and generates the regional overload node scheduling level table.

[0023] As a further solution of the present invention, the formula for calculating the dispatch level index value of the site is:

[0024]

[0025] Among them, L x represents the dispatching level index value of the x-th station, D x K represents the high-density occurrence frequency of the area where the x-th station is located within three scheduling cycles, xzrepresents the number of vehicle resource gaps at the x-th station in time period z, m x is the number of time periods with gaps recorded at the x-th station, H x is the ranking factor of the x-th site in the region, F x The number of redundancy scheduling failures for the x-th site in the historical period.

[0026] As a further solution of the present invention, the candidate path reconstruction module includes:

[0027] The priority node extraction submodule filters the station numbers marked as priority transfer based on the regional overload node scheduling level table, establishes a transfer target node index, and extracts the adjacent area range of each target node to generate a set of transferable target nodes;

[0028] The vehicle status calculation submodule extracts the vehicle's station number, current location, last return time, and current idle status in the adjacent area based on the set of transferable target nodes, calculates the driving distance from the current location to the target node and the return interval, and uses the station information for normalization to obtain a standardized vehicle status data table;

[0029] The path screening and sorting submodule performs weighted calculation and sorting on the normalized driving distance and normalized response time of the vehicle based on the standardized vehicle status data table, and screens the vehicle numbers and corresponding path records whose scores are within the target range to generate a path-transferable vehicle dispatch group.

[0030] As a further embodiment of the present invention, the adjustable vehicle screening module includes:

[0031] The time limit construction submodule obtains orders marked as high urgency in the task list, extracts the creation time and maximum response time of each order, combines the two data into time limit record entries according to the task number, and generates the task time limit dataset;

[0032] The response index extraction submodule calls the path transferable vehicle dispatch group, extracts the previous task completion time of the corresponding vehicle, the average time interval of the last three task calls, and the standard deviation of the historical dispatch response time, establishes an index correspondence according to the vehicle number, and obtains a vehicle response performance parameter table;

[0033] The task vehicle screening submodule compares the three indicators of task time limit and vehicle response item by item based on the task time limit dataset and the vehicle response performance parameter table, determines whether each indicator is less than the task response limit standard, screens out any vehicle number that exceeds the limit, and marks the scheduling adaptation relationship between the remaining vehicles and the task, thereby obtaining a list of vehicles matching high-urgency tasks.

[0034] As a further embodiment of the present invention, the system further comprises:

[0035] The dispatch path binding module matches the vehicle list of the high-urgency task with the path and station information in the path-transferable vehicle dispatch group, reconstructs the vehicle ID, starting station ID, and target path segment in the dispatch instruction structure, and records the task number and dispatch level label in an entry format to generate task-bound path dispatch information;

[0036] The task-bound path scheduling information includes a vehicle instruction structure item, a task scheduling number, a site path identifier, and a scheduling level item.

[0037] As a further solution of the present invention, the scheduling path binding module includes:

[0038] The path field matching submodule matches the vehicle list based on the high-urgency task, extracts the vehicle number and matches the corresponding starting station ID and target path segment number in the path transferable vehicle dispatch group, and extracts its vehicle ID information, constructs a three-field combination structure, and generates a path field association set;

[0039] The instruction content update submodule calls the path field association set, replaces the corresponding positions in the scheduling instruction structure in the order of vehicle ID, starting station ID, and target path segment according to each group of vehicle field combinations, and records the task number corresponding to the replaced field to obtain a structured scheduling instruction set;

[0040] The entry format generation submodule extracts the task number and the vehicle corresponding scheduling level identifier of each instruction based on the structured scheduling instruction set, integrates them into a parallel structure of task field and scheduling field, reorganizes them into entry-type standardized expression content, and generates task-bound path scheduling information.

[0041] A vehicle scheduling optimization method for a car rental platform, which is used to implement the above-mentioned vehicle scheduling optimization system for the car rental platform, comprises the following steps:

[0042] S1: Obtain order records from the car rental platform, map the rental duration of each order to the time period, evaluate the saturation status of the region within the time period, and generate a regional continuous rental saturation identifier set;

[0043] S2: calling the regional continuous rental saturation identification set, extracting the area marked as a high-density rental state, combining the vehicle gap and the high-frequency index value, calculating the station's dispatch level, and obtaining a regional overload node dispatch level table;

[0044] S3: Based on the regional overload node dispatch level table, the average dispatch response time and the empty distance index of the vehicles are normalized and then weighted and ranked to screen the optimal path for executing the dispatch task and generate a dispatch group of path-transferable vehicles;

[0045] S4: Obtain orders marked as high urgency in the task list, extract the vehicle's previous task completion time, the average of the last three call intervals, and the standard deviation of the historical response time, and compare them with the maximum response time limit of the order task. If any of the indicators exceeds the response limit, the vehicle number is removed to obtain a list of vehicles matching high urgency tasks;

[0046] S5: Based on the high-urgency task matching vehicle list, the vehicle ID, starting station ID, and target path segment in the scheduling instruction structure are reconstructed, and the task number and scheduling level label are combined into an entry format to generate task-bound path scheduling information.

[0047] Compared with the prior art, the advantages and positive effects of the present invention are:

[0048] In the present invention, by analyzing the time and geographical data of car rental orders, accurate judgment of vehicle usage status is achieved, and vehicle allocation is optimized. By continuously monitoring and evaluating the saturation status of regional vehicle usage, high-demand areas can be identified, and vehicle distribution can be adjusted accordingly to ensure the optimal allocation of resources. According to the actual operation data and scheduling history of the vehicle, the scheduling path is intelligently reconstructed, which reduces the idle time of the vehicle, improves scheduling efficiency, shortens customer waiting time, and reduces operating costs. By analyzing the urgency of the task, high-urgency orders are given priority, which ensures the timeliness of the service and customer satisfaction, improves the overall scheduling performance, and brings a double improvement in economic benefits and customer loyalty to the car rental platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0050] Figure 1 is a system flow chart of the present invention;

[0051] Figure 2 Schematic diagram of the system framework of the present invention;

[0052] Figure 3 Flowchart of leasing saturation identification module for the present invention;

[0053] Figure 4 This is a flow chart of the scheduling level setting module of the present invention;

[0054] Figure 5 This is a flow chart of the candidate path reconstruction module of the present invention;

[0055] Figure 6 A flow chart of the adjustable vehicle screening module of the present invention;

[0056] Figure 7 This is a flow chart of the scheduling path binding module of the present invention;

[0057] Figure 8 Schematic diagram of the method steps of the present invention. DETAILED DESCRIPTION

[0058] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0059] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0060] In the embodiments of the present invention, the terms "image" and "picture" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same. The terms "of," "corresponding," and "corresponding" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same.

[0061] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0062] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0063] See also Figure 1 The present invention provides a technical solution: a vehicle scheduling optimization system for a car rental platform, the system comprising:

[0064] The rental saturation identification module obtains order records from the car rental platform, extracts the rental start and end times, the rental vehicle number, and the geographic region tag, divides the rental time of the orders into time periods by region, maps the rental duration of each order to the time period, and performs cross-aggregation statistics by region and time period. It uses the ratio of average usage time to the vehicle dispatch capacity within the time period to calculate the vehicle occupancy saturation score. The occupancy saturation score is used to evaluate the saturation status of the regional time period and generate a regional continuous rental saturation identification set;

[0065] Average usage duration is the sum of all rental durations for orders within a time period divided by the number of orders. Vehicle dispatch capacity refers to the maximum number of vehicles that can be dispatched within the area and time period, as defined by the platform. This data is derived from the platform's vehicle distribution and dispatch system capacity settings. The ratio of the two reflects the vehicle occupancy rate within the time period.

[0066] The dispatching level setting module calls the regional continuous rental saturation identification set, extracts the area marked as high-density rental status, obtains the number of available vehicles at each rental station in the area, and compares it with the minimum number of operating vehicles. If it is less than the standard, the station number is recorded and the number of vehicle gaps is extracted. At the same time, the high occupancy frequency of the area in the past three scheduling cycles is counted. Combining the vehicle gap and the high frequency index value, the dispatching level of the station is calculated by setting weights, and the regional overload node dispatching level table is obtained;

[0067] The minimum number of operating vehicles refers to the minimum number of vehicles configured by the platform to ensure service continuity, generally based on the average daily rental frequency and user density. The high occupancy frequency index value refers to the number of times an area is in a high occupancy state within a unit period, and is usually expressed as the statistical frequency within a sliding time window.

[0068] Based on the regional overload node scheduling level table and the station numbers marked as priority entry nodes, the candidate path reconstruction module extracts the current station numbers and vehicle status information of vehicles in adjacent areas. It then obtains the driving distance from the vehicle's current location to the target node, the current vehicle idle status, and the last return time. It calculates the average scheduling response time and the idle distance index. After standardizing the average scheduling response time and the idle distance index, it performs weighted calculation and sorting to select the optimal path for executing the scheduling task and generate a path transfer vehicle scheduling group.

[0069] The average dispatch response time refers to the average time required for this type of vehicle on the platform to complete the instruction and reach the dispatch target point, which is calculated from historical dispatch records; the idle driving distance index is the ratio of the distance between the vehicle's current position and the dispatch target site to the maximum idle driving distance that the platform can tolerate, reflecting the risk of resource waste in vehicle allocation.

[0070] The adjustable vehicle screening module obtains orders marked as high-urgency in the task list, reads the creation time and maximum response time, forms a time-limited data group, calls the path transfer vehicle dispatch group, extracts the vehicle's previous task completion time, the average of the last three call intervals, and the standard deviation of the historical response time, and compares them with the maximum response time of the order task. If any indicator exceeds the response limit standard, the vehicle number is eliminated, and a list of vehicles matching high-urgency tasks is obtained;

[0071] The maximum response time is the maximum tolerance time from the creation of the task to the start of the scheduling operation, which is usually defined by the user's car rental reservation time and the platform service response agreement;

[0072] The scheduling path binding module matches the vehicle list based on high-urgency tasks, matches the path and station information in the path-transferable vehicle scheduling group, reconstructs the vehicle ID, starting station ID, and target path segment in the scheduling instruction structure, and records the combined task number and scheduling level label in an entry format to generate task-bound path scheduling information.

[0073] The regional continuous rental saturation identification set includes the vehicle occupancy ratio label of the time period, the regional number index, the continuous saturation state mark and the upper limit of the adjustable vehicles in the corresponding time period. The regional overload node scheduling level table specifically includes the site vehicle gap value, the historical high occupancy frequency factor, the scheduling level index and the site number label. The path transferable vehicle scheduling group includes the transfer-out node number, the target site path pair, the empty driving distance ratio and the scheduling response time interval. The high-urgency task matching vehicle list specifically refers to the task number label, the vehicle response interval index, the task response limit status and the vehicle scheduling feasibility mark. The task-bound path scheduling information includes the vehicle instruction structure item, the task scheduling number, the site path identifier and the scheduling level entry.

[0074] See also Figure 2 and Figure 3 ,The rental saturation identification module includes an order period mapping submodule, an occupancy index calculation submodule, and a saturation state determination submodule;

[0075] The order-period mapping submodule obtains order records from the car rental platform, extracts the rental start and end times, the rental vehicle number, and the geographic region label, maps the duration of each order to the corresponding time period and geographic region, and establishes a dual index structure of region and time period to generate a region-period mapping dataset;

[0076] After obtaining the order records from the car rental platform, extract the rental start and end time of each order, and calculate the rental duration corresponding to the order by the difference between the two. If the start time of order A is 8:00 on July 1, 2024, and the end time is 18:00 on July 1, the duration is 10 hours. Then extract the vehicle number corresponding to each order to associate it with the specific vehicle usage, and spatially classify the vehicle usage location in combination with the geographic area label bound to each order. Then divide the duration of all orders into specific time periods according to the hourly granularity to form an aggregated statistical unit of time period-region. If order A is in region X, its 10-hour duration is distributed to the time period of 8:00-18:00 on July 1 and is marked in the region. On this basis, a dual index structure of regions and time periods is established. This structure uses the region number and time period number as the index key, summarizes all order IDs and their corresponding rental duration information under the index, and stores them in a two-layer structure table. At the same time, the number of orders for each time period in the structure is counted. For example, in region X, there are 5 orders mapped in the time period from 8:00 to 9:00 on July 1, 2024, with a cumulative duration of 5.3 hours. The total rental time in region X is 5.3 hours. The statistical structure adopts a two-dimensional table format, with each row representing a different region and each column corresponding to the time period number. The corresponding cell is filled with the total order duration value of the time period area, and the relevant vehicle number and order ID are retained for subsequent processing. Combining the above steps, a region-time period mapping dataset is generated.

[0077] The occupancy index calculation submodule calls the regional time period mapping dataset, extracts the number of orders and the average usage time of each order in each time period of each region, obtains the vehicle dispatch capacity of the corresponding time period of the region, calculates the occupancy saturation score of the region time period, and constructs a scoring structure based on the region as a unit to obtain the regional occupancy score matrix;

[0078] The formula for calculating the occupancy saturation score for a region and time period is:

[0079]

[0080] Among them, S ij represents the occupancy saturation score of region i in time period j, T ijk is the normalized value of the rental duration of the k-th order in area i and time period j, n ij is the total number of orders in region i during time period j, C ij is the normalized value of vehicle dispatch capacity configured for region i in time period j, R ij is the normalized value of the total amount of scheduling resources in region i in time period j, M ij is the normalized value of the average task resource demand of region i in time period j;

[0081] Call the regional time period mapping dataset, extract the number of orders in each time period of each region and the average usage time of all orders in the time period one by one. For example, there are 4 orders in region A and time period T1, and the rental time is 2, 3, 4, and 5 hours respectively. The average usage time is (2+3+4+5) / 4=3.5 hours. Then obtain the vehicle dispatch capacity configured by the platform for region A in time period T1. The dispatch capacity is the maximum number of dispatchable vehicles set by the platform in the system. If the value is 10 vehicles, the capacity value of the participating formula is 10. For this dispatch The capacity value is normalized. Assuming that the platform's full scheduling capacity is 100 vehicles, the normalized capacity value is 10 / 100=0.1. At the same time, the total scheduling resources and task resource requirements in the area and time period are extracted. For example, if the current regional resource volume is 12 vehicles and the average task demand is 9 vehicles, the normalized values are 0.12 and 0.09 respectively. The absolute value of the resource difference is calculated to be 0.03, and then the square root calculation is performed to obtain approximately 0.1732. Finally, the occupancy saturation score of area A in time period T1 is calculated according to the formula: 3.5×4÷

[0082] (0.1 + 0.1732) = 14 / 0.2732 ≈ 51.25. The above calculation results are combined with other regions and time periods to establish a scoring matrix structure. The matrix rows represent the region numbers, and the columns represent the time period numbers. The cells are filled with the saturated score values under this combination to obtain the region occupancy score matrix.

[0083] The saturation state determination submodule compares each score value with the scheduling load reference threshold set by the platform based on the area occupancy score matrix. It marks the area and time period combinations whose score values exceed the reference threshold for more than three consecutive times, merges them into interval labels, and generates the area continuous rental saturation identification set;

[0084] Based on the regional occupancy score matrix, each score is compared with the platform's dispatch load reference threshold. This threshold is preset by the platform and is set based on the 75th percentile of the daily average saturation score data for all regions over the past month. If region X's saturation score is 42 and its 75th percentile score is 40, it is marked as exceeding the threshold. The number of time periods in each region that continuously exceed this threshold is then counted. When a region's score exceeds the threshold for three or more consecutive time periods, it is marked as continuously saturated. The region and time period number are combined to create an interval label. For example, if region X exceeds the threshold in time periods T3, T4, and T5, the region is marked as saturated in the T3-T5 interval. This label information is summarized as a triple structure, including the region number, the starting time period number, and the ending time period number. The output format is {X, T3, T5}. The saturation status labels of all regions are output as a structured set to generate the regional continuous lease saturation identification set.

[0085] See also Figure 2 and Figure 4 ,The scheduling grade setting module includes a high-density area extraction submodule, a resource gap identification submodule, and a grade weight calculation submodule;

[0086] The high-density area extraction submodule calls the area continuous lease saturation identification set, filters the area numbers marked as saturated, and establishes the area site mapping index to generate a high-density area label set;

[0087] The continuous rental saturation identification set is called to identify all regions marked as saturated. The filtering operation is performed based on the flag information for scores that continuously exceed the reference threshold. First, the region number and the consecutive exceeding-threshold time period are extracted from each record to determine whether there are three or more consecutive time periods. If a region is marked as exceeding the threshold continuously at T3, T4, and T5, the region number is determined to be a saturated region that meets the conditions and its number X is included in the saturated region set. Then, each region number in this set is used as an index to retrieve the station distribution records in the car rental platform structure database. The unique IDs of all available stations in the region are extracted based on the region-station affiliation relationship. For example, if stations S1, S2, S3, and S4 exist in region X, an index structure {X: [S1, S2, S3, S4]} is created. This structure is constructed using a hash map, which supports the subsequent rapid retrieval of all station numbers in the region using the region number as the key. This extraction operation completes the data preparation for the spatial clustering index. Finally, all region numbers that meet the continuous saturation state and their corresponding station sets are uniformly output to generate a high-density region label set.

[0088] The resource gap identification submodule is based on a high-density area tag set. Based on the recorded area number, it extracts the number of available vehicles for each rental station in the corresponding area, obtains the minimum number of operating vehicles set by the platform, performs difference calculations, filters out station numbers with insufficient available numbers, and calculates their gap values to obtain the vehicle resource gap set.

[0089] Based on the high-density area tag set, a site-level operation is performed on each area number record. First, all mapped site numbers under the area are extracted, and then the current number of available vehicles data is read for each site number. For example, the number of idle vehicles at site S1 is 4. The difference between this value and the minimum number of operating vehicles set by the platform is judged. If the platform defines the minimum number of operating vehicles as 6, the difference is 4-6=-2. Since the result is a negative number, it means that S1 is a site with insufficient resources and the gap is 2. The reference value is set to refer to the vehicle scheduling pressure balance rules of each area of the platform, which is based on the average number of vehicles in the area over the past seven days. The average task initiation quantity and the average adjustable frequency of a single vehicle are estimated together, and a fixed lower limit is set for each area. An example of the rationality of this value is: if the average daily task in a certain area is 10 orders and the frequency of a single vehicle is 2 times / day, then the benchmark value is 10 / 2=5 vehicles. After the gap site numbers are extracted through the negative value screening mechanism, a one-to-one correspondence between the site numbers and the gap quantities is formed. The result is recorded as a mapping structure of {site number: gap quantity}. For example, {S1: 2, S4: 1} indicates that there is a shortage of vehicles at two sites, and the total gap is 3 vehicles. This set supports grouping by site number or area number to obtain a vehicle resource gap set.

[0090] The level weight calculation submodule calls the vehicle resource gap set, extracts the high-density occurrence frequency of the station area in the last three scheduling cycles, performs a superimposed proportion score based on the frequency value and the corresponding gap amount, calculates the station's scheduling level index value, constructs the station level mapping structure, and generates the regional overload node scheduling level table;

[0091] The formula for calculating the dispatch level index value of a site is:

[0092]

[0093] Among them, L x represents the dispatching level index value of the x-th station, D x K represents the high-density occurrence frequency of the area where the x-th station is located within three scheduling cycles, xz represents the number of vehicle resource gaps at the x-th station in time period z, m x is the number of time periods with gaps recorded at the x-th station, H x is the ranking factor of the x-th site in the region, F x The number of redundancy scheduling failures for the x-th site in the historical period.

[0094] Call each station number in the vehicle resource gap set, extract the corresponding area number to which the station belongs, and further obtain the frequency value of the area being marked as high density in the last three scheduling cycles, recorded as D x , if site S1 belongs to area X, and area X has been judged as high density twice in the scheduling score in the past three scheduling cycles, then Dx =2, then extract the total number of time periods m in which the site appears in the gap records x And the gap value K for each time period xz For example, station S1 has a shortage of 2 and 1 vehicles in the two time periods Z1 and Z2 respectively, indicating that m x =2,K x1 =2,K x2 =1, and then calculate the sum of the absolute values of the gaps Then extract the site number sorting factor H x , this factor is determined according to the sequence number of the site in the region to which it belongs. Suppose S1 is the second site in region X, then H x =2, and then obtain the number of redundant scheduling failures F in the historical period of the site x If the scheduling process is recorded as a failure due to reasons such as path failure or vehicle scheduling failure, then F x =1, substitute into the following formula:

[0095]

[0096] Therefore, the dispatch level index value L of station S1 is x The value of the node is about 7.07. Repeating the above process can perform the corresponding scoring operation on each gap site. Then, the numbers of all sites and their corresponding dispatch level scores are integrated to construct a key-value mapping structure. The key is the site number and the value is the scoring result, for example, {S1: 7.07, S4: 6.50, S8: 8.10}. Finally, the complete structure data is formed to generate the regional overload node dispatch level table.

[0097] See also Figure 2 and Figure 5 ,The candidate path reconstruction module includes a priority node extraction submodule, a vehicle state calculation submodule, and a path screening and sorting submodule;

[0098] The priority node extraction submodule filters the station numbers marked as priority transfer based on the regional overload node scheduling level table, establishes the transfer target node index, and extracts the adjacent area range of each target node to generate a set of transferable target nodes;

[0099] Based on the regional overload node scheduling grade table, all site numbers with grade scores higher than the priority scheduling threshold are identified. The priority scheduling threshold is set by the 90th percentile of the platform according to the regional score. If the threshold is 6.5, all sites with scores greater than or equal to 6.5 are judged as priority transfer sites. For example, site S3 with a score of 7.1 and site S4 with a score of 6.8 both meet the screening conditions. The platform records the score structure as {S1: 4.2, S2: 5.9, S3: 7.1, S4: 6.8}, where S3 and S4 are the screening results. The target node index table is then established for the site numbers screened out above, forming the target node of {S3, S4}. Point index set, with each site in the set as the center, searches for the corresponding adjacent area unit in the map structure of the geographical area where it is located. Adjacent areas are defined as adjacent administrative areas with a straight-line distance of less than 5 kilometers from the site on the map or sharing a border line. If S3 belongs to area A and its adjacent areas are B, C, and D, then the adjacent areas of the target node of the site are recorded as the set {B, C, D}. In this way, a reverse-association area index is established for all target nodes. Each target node is linked to multiple adjacent resource areas that can be overflow-deployed. Finally, a multi-dictionary structure is constructed with the site number as the key value and the adjacent area set as the value to generate a set of transferable target nodes.

[0100] The vehicle status calculation submodule extracts the vehicle's station number, current location, last return time, and current idle status in the adjacent area based on the set of transferable target nodes. It calculates the driving distance from the current location to the target node and the return interval, and uses the station information for normalization to obtain a standardized vehicle status data table.

[0101] According to the set of transferable target nodes, the adjacent area numbers corresponding to each target node are extracted in turn, and all vehicles in the area are screened in the platform vehicle database to obtain their current belonging site numbers and positioning coordinates as basic location information. The vehicle real-time status record table is called to extract whether the current vehicle is in an idle state. If the status value is "idle", it is marked as an idle vehicle. Then the end time field of the vehicle's last completed task is read and compared with the current system time to calculate the time interval. For example, vehicle V201 completed the task at 12:30 on April 1, and the current time is 14:00 on April 1, then the return interval is 1.5 hours. Then, based on the vehicle's current position and the site coordinates of the transferable target node, the straight-line distance formula is used to calculate the current driving distance to the target node. If the current The coordinates are (100.4, 20.6), and the target node coordinates are (101.0, 21.2). The straight-line distance is approximately 0.848 kilometers. The three parameters of each vehicle's idle state, return time interval, and driving distance are normalized in their respective dimensions. The maximum and minimum normalization method is used. The normalized value calculation formula is: (X-Xmin) / (Xmax-Xmin). Taking the driving distance as an example, the maximum distance is 3.0 kilometers and the minimum is 0.4 kilometers. The normalized distance of V201 is (0.848-0.4) / (3.0-0.4)≈0.185. The normalization of all parameters of all vehicles is completed separately. In the result structure, the vehicle number is used as the key and the three standardized parameters are used as the value to form a triple, and a standardized vehicle status data table is obtained.

[0102] The route screening and sorting submodule uses a standardized vehicle status data table to perform weighted calculation and sorting based on the vehicle's normalized driving distance and normalized response time. It then selects the vehicle numbers and corresponding route records that fall within the target range and generates a route-transferable vehicle dispatch group.

[0103] Based on the standardized vehicle status data table, the normalized travel distance and normalized response time are read one by one by vehicle number, and a comprehensive dispatch score is calculated for each vehicle. The score is calculated using a weighted average method. Assuming a weight of 0.6 for travel distance and 0.4 for response time, and a normalized travel distance of 0.185 and a response time of 0.25 for vehicle V201, the comprehensive score is 0.6 × 0.185 + 0.4 × 0.25 = 0.111 + 0.1 = 0.211. All candidate vehicles are comprehensively scored and ranked in this manner. The vehicle numbers and their corresponding target path records with the top 10% scores are selected. If there are 100 candidate vehicles in total, the top 10 vehicles are selected. The vehicle numbers are combined with basic dispatch information records such as the current station number, the target transfer node number, the path distance, and the response time to construct a path dispatch record structure table. The record content structure is unified in the format of {vehicle ID, starting station ID, target node ID, score value}, and a path transferable vehicle dispatch group is generated.

[0104] See also Figure 2 and Figure 6 ,The adjustable vehicle screening module includes a time constraint construction submodule, a response index extraction submodule, and a task vehicle screening submodule;

[0105] The time limit construction submodule obtains orders marked as high urgency in the task list, extracts the creation time and maximum response time of each order, combines the two data into time limit record entries according to the task number, and generates the task time limit dataset;

[0106] Get the orders marked as high in the task list, identify the records with the field "priority_level" as "high", and extract the task number, creation time and maximum response time field in this type of order in turn. The creation time is the timestamp of the order submitted to the system. The maximum response time indicates the longest scheduling response time that the platform can tolerate after the order is generated, in minutes or hours. For example, the creation time of task T01 is 9:00 on April 5, 2024, and the maximum response time is 30 minutes. After extraction, combine the two data according to the task number to form a structured entry format. The entry format is {task number, creation time, maximum response time}. Orders are uniformly constructed as a set of time limit records to provide scheduling time boundaries for subsequent vehicle response determination. During the construction process, the time fields need to be uniformly converted to a timestamp format (such as Unix timestamp), and the response time limit is converted to seconds or minutes to support subsequent comparative calculation operations. For example, the creation time of task T02 is 13:30 on April 5, 2024, and the corresponding timestamp is 1712295000. The maximum response time limit is set to 45 minutes, which is converted to 2700 seconds. The constructed entry is {T02, 1712295000, 2700}. All entries are finally uniformly output as a task key-value pair structure to generate a task time limit dataset.

[0107] The response index extraction submodule calls the path transferable vehicle dispatch group, extracts the previous task completion time of the corresponding vehicle, the average time interval of the last three task calls, and the standard deviation of the historical dispatch response time, establishes the indicator correspondence according to the vehicle number, and obtains the vehicle response performance parameter table;

[0108] The call path can be transferred to the vehicle dispatch group to extract the dispatch information associated with each vehicle number. First, read the previous task completion time of vehicle V number as the end timestamp of the vehicle's most recent valid task. For example, the last task completion time of vehicle V125 is 11:00 on April 5, 2024, and the corresponding timestamp is 1712288400. Then extract the time intervals between the last three task calls recorded in the dispatch log for the vehicle. The time interval is the difference between the task scheduling start time and the last task completion time. , set the three intervals to be 20 minutes, 15 minutes, and 25 minutes respectively, then the average value is (20+15+25) / 3=20 minutes, corresponding to 1200 seconds. Continue to read all response time fields in the historical dispatch records of the vehicle, calculate the standard deviation to measure its response stability. If the response time records are 14, 18, 22, and 16 minutes, then its standard deviation σ=√[(14-17.5)2+(18-17.5)2+(22-17.5)2+(16-17.5)2] / 4=√

[0109] [(12.25+0.25+20.25+2.25) / 4]=√(35 / 4)=√8.75≈2.96 minutes. The above data is recorded in a structure using the vehicle number as the key and the three response indicators as the value. The record structure is {vehicle ID, last completion timestamp, call interval mean, response time standard deviation}. Each indicator value is uniformly converted to a unified unit (seconds) to facilitate alignment with the task response time limit, resulting in a vehicle response performance parameter table.

[0110] The task vehicle screening submodule compares the task time limit and vehicle response performance parameters based on the task time limit dataset and the vehicle response performance parameter table. It determines whether each indicator is less than the task response limit standard, screens out any vehicle numbers that exceed the limit, and marks the scheduling adaptation relationship between the remaining vehicles and the task, thus obtaining a list of vehicles matching high-urgency tasks.

[0111] Based on the task time limit data set and the vehicle response performance parameter table, for each task number, the task creation timestamp and the maximum response time limit are read in the task record, and the task allowed response deadline is calculated by adding the creation timestamp to the response time limit in seconds. For example, the task T03 creation time is 1712292000 seconds, the maximum response time limit is 1800 seconds, and the task response deadline is 1712293800 seconds. Then, for each candidate vehicle number, the corresponding previous task completion timestamp, call interval average value and response time standard deviation are read, and the task deadline is subtracted from the vehicle's previous task completion time. If the result is less than the sum of the call interval average value and the response standard deviation, the vehicle is judged to be If there is a scheduling risk for a vehicle, the vehicle number is eliminated. The judgment formula is: if 1712293800-last completion timestamp < average call interval + response standard deviation, the vehicle does not meet the scheduling conditions. For example, the last completion timestamp of vehicle V208 is 1712292600, the interval mean is 800 seconds, and the response standard deviation is 400 seconds. The remaining response time is 1200 seconds, and the expected response time is 800+400=1200 seconds. If the two are equal, the vehicle is marginally available. The platform sets the judgment threshold to be less than or equal to the threshold for elimination, so V208 is not selected. Finally, the vehicle numbers and their associated task numbers that meet all the judgment conditions are retained to obtain a list of vehicles matching high-urgency tasks.

[0112] See also Figure 2 and Figure 7 ,The scheduling path binding module includes a path field matching submodule, an instruction content updating submodule, and an entry format generating submodule;

[0113] The path field matching submodule matches the vehicle list based on the high-urgency tasks, extracts the vehicle number, and matches the corresponding starting station ID and target path segment number in the path transferable vehicle dispatch group. It also extracts its vehicle ID information, constructs a three-field combination structure, and generates a path field association set.

[0114] Based on the high-urgency task matching vehicle list, extract the vehicle number for each record, traverse the path transfer vehicle dispatch group, match the vehicle number with its associated dispatch field, read the starting station ID and target path segment number fields in the matching item, if the match is successful, keep the record, if the match fails, skip the record, for example, if the vehicle number is V205 and the starting station ID is S102 and the target path segment number is P307 in the dispatch group, then assemble these three pieces of information with the vehicle number into a field combination structure {V205, S102, P307}, and the field order in the structure is fixed to vehicle I D. Start station ID, target path segment number, ensure the consistency of subsequent data binding, repeat the above operation until all vehicle numbers in the list are retrieved, there must be no missing items in the structure and the field content must be kept non-empty. If the data in the dispatch group is missing, it must be supplemented from the real-time vehicle status data, and then the structure of all three fields is stored in the mapping structure, with the vehicle number as the key value and the field combination as the value. For example, the final structure is: {V205:{S102,P307},V210:{S108,P322}}. After the construction is completed, it is archived uniformly to generate a path field association set.

[0115] The instruction content update submodule calls the path field association set. Based on each vehicle field combination, it replaces the corresponding positions in the dispatch instruction structure in the order of vehicle ID, starting station ID, and target path segment, and records the task number corresponding to the replaced field to obtain a structured dispatch instruction set.

[0116] The path field association set is called, and the three fields extracted from each record are sequentially bound to the corresponding positions in the dispatch instruction structure. First, the vehicle ID field in the original dispatch instruction structure is replaced with the vehicle number in the current field combination. Then, the original starting station ID field is replaced with the second station number in the field combination. Finally, the target path segment number field is replaced with the third path number in the field combination. The replacement operation is performed according to the field sequence in the dispatch structure template to ensure that the corresponding positions of the field contents are not misplaced. At the same time, the task number corresponding to the current field combination is bound to the above three fields to form a structured dispatch instruction record. For example, if the task number is T048 and the corresponding field combination is {V205, S102, P307}, the structured dispatch record {T048, V205, S102, P307} is formed. The remaining field combinations are processed in the same way. All records are stored in a unified set in a structured format. The set structure is fixed to four fields: task number, vehicle ID, starting station ID, and target path segment number. The field order is fixed, and the storage format supports retrieval by task number, resulting in a structured dispatch instruction set.

[0117] The entry format generation submodule extracts the task number and the corresponding vehicle scheduling level identifier of each instruction based on the structured scheduling instruction set, integrates them into a parallel structure of task fields and scheduling fields, reorganizes them into entry-style standardized expression content, and generates task-bound path scheduling information;

[0118] Based on the structured scheduling instruction set, the task number field and the corresponding vehicle number field are extracted from each record. The regional overload node scheduling grade table is indexed by the vehicle number to obtain the scheduling grade score corresponding to the station to which the vehicle belongs. The score value is normalized according to a unified interval, ranging from 0 to 10, with 0-3 defined as low grade, 4-7 as medium grade, and 8-10 as high grade. For example, if the scheduling grade of vehicle V205 is 8.4, it is classified as high grade. After extracting the values, a parallel field record of task and grade is constructed. The field content consists of five items: task number, vehicle number, starting station ID, target path segment number, and scheduling grade score, such as {T048, V205, S102, P307, 8.4}. All records in the structured scheduling instruction set are reassembled according to this structure and uniformly converted to an entry-type record format. The field order is consistent with the system scheduling processing field template, and no fields are missing. The generated record format supports fast retrieval by task number or scheduling grade. The unified format is stored and output to generate task-bound path scheduling information.

[0119] See also Figure 8 ,The vehicle scheduling optimization method for a car rental platform includes the following steps:

[0120] S1: Obtain order records from the car rental platform, map the rental duration of each order to the time period, evaluate the saturation status of the region within the time period, and generate a regional continuous rental saturation identifier set;

[0121] S2: Call the regional continuous rental saturation identification set, extract the area marked as high-density rental status, obtain the number of available vehicles at each rental station in the area, combine the vehicle gap and high-frequency index value, calculate the station's dispatch level, and obtain the regional overload node dispatch level table;

[0122] S3: Based on the regional overload node dispatch level table, the average dispatch response time and empty distance index of the vehicles are normalized and then weighted and ranked to screen the optimal path for executing the dispatch task and generate a dispatch group of vehicles with transferable paths;

[0123] S4: Obtain orders marked as high urgency in the task list, extract the vehicle's previous task completion time, the average of the last three call intervals, and the standard deviation of the historical response time, and compare them with the maximum response time limit of the order task. If any of the indicators exceeds the response limit, the vehicle number is removed to obtain a list of vehicles matching high urgency tasks;

[0124] S5: Based on the high-urgency task matching vehicle list, match the path and station information in the path-transferable vehicle scheduling group, reconstruct the vehicle ID, starting station ID, and target path segment in the scheduling instruction structure, and record the combined task number and scheduling level label in an entry format to generate task-bound path scheduling information.

[0125] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0126] In this disclosure, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, "at least one of a, b, or c" can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0127] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0128] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0129] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0130] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of the device or unit, which can be electrical, mechanical or other forms.

[0131] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0132] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0133] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0134] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. The vehicle scheduling optimization system of the car rental platform is characterized by: The system comprises: The rental saturation identification module obtains order records from the car rental platform, maps the rental duration of each order to the time period, evaluates the saturation status of the area within the time period, and generates a set of regional continuous rental saturation identifiers; The dispatching level setting module calls the regional continuous rental saturation identification set, extracts the area marked as a high-density rental state, obtains the number of available vehicles at each rental station in the area, combines the vehicle gap and the high-frequency index value, calculates the dispatching level of the station, and obtains the regional overload node dispatching level table; The candidate path reconstruction module normalizes the average dispatch response time and the empty distance index of the vehicles based on the regional overload node dispatch level table, performs weighted calculation and sorting, selects the optimal path for executing the dispatch task, and generates a path transferable vehicle dispatch group; The adjustable vehicle screening module obtains orders marked as high urgency in the task list, extracts the vehicle's previous task completion time, the average of the last three call intervals, and the standard deviation of the historical response time, and compares them with the maximum response time limit of the order task. If any indicator exceeds the response limit standard, the vehicle number is eliminated to obtain a list of vehicles matching high-urgency tasks.

2. The vehicle scheduling optimization system for the car rental platform according to claim 1 is characterized in that: The regional continuous rental saturation identification set includes a vehicle occupancy ratio label for a time period, an area number index, a continuous saturation state mark, and an adjustable vehicle upper limit for the corresponding time period; the regional overload node scheduling level table specifically includes a site vehicle gap value, a historical high occupancy frequency factor, a scheduling level index, and a site number label; the path transferable vehicle scheduling group includes a transfer-out node number, a target site path pair, an empty driving distance ratio, and a scheduling response time interval; the high-urgency task matching vehicle list specifically includes a task number label, a vehicle response interval index, a task response limit status, and a vehicle scheduling feasibility identifier.

3. The vehicle scheduling optimization system for the car rental platform according to claim 2 is characterized in that: The rental saturation identification module includes: The order-period mapping submodule obtains order records from the car rental platform, extracts the rental start and end times, the rental vehicle number, and the geographic region label, maps the duration of each order to the corresponding time period and geographic region, and establishes a dual index structure of region and time period to generate a region-period mapping dataset; The occupancy index calculation submodule calls the regional time period mapping dataset, extracts the number of orders and the average usage time of orders in each time period of each region, obtains the vehicle dispatch capacity of the region corresponding to the time period, calculates the occupancy saturation score of the region time period, and constructs a scoring structure based on the region as a unit to obtain the regional occupancy score matrix; The formula for calculating the occupancy saturation score of the area time period is: Among them, S ij represents the occupancy saturation score of region i in time period j, T ijk is the normalized value of the rental duration of the k-th order in area i and time period j, n ij is the total number of orders in region i during time period j, C ij is the normalized value of vehicle dispatch capacity configured for region i in time period j, R ij is the normalized value of the total amount of scheduling resources in region i in time period j, M ij is the normalized value of the average task resource demand of region i in time period j; The saturation state determination submodule compares each score value with the scheduling load reference threshold set by the platform based on the area occupancy score matrix, marks the area and time period combinations whose score values exceed the reference threshold for more than three consecutive times, and merges them into interval labels to generate an area continuous rental saturation identification set.

4. The vehicle scheduling optimization system for the car rental platform according to claim 3 is characterized in that: The scheduling level setting module includes: The high-density area extraction submodule calls the area continuous lease saturation identification set, filters the area numbers marked as saturated, establishes the area site mapping index, and generates a high-density area label set; The resource gap identification submodule extracts the number of available vehicles for each rental station in the corresponding area based on the high-density area tag set and the recorded area number, obtains the minimum number of operating vehicles set by the platform, performs difference calculation, filters the station numbers with insufficient available numbers, and calculates their gap values to obtain the vehicle resource gap set; The level weight calculation submodule calls the vehicle resource gap set, extracts the high-density occurrence frequency of the area where the station is located in the last three scheduling cycles, performs superimposed proportion scoring based on the frequency value and the corresponding gap amount, calculates the scheduling level index value of the station, constructs the station level mapping structure, and generates the regional overload node scheduling level table.

5. The vehicle scheduling optimization system for the car rental platform according to claim 4 is characterized in that: The formula for calculating the dispatch level index value of the site is: Among them, L x represents the dispatching level index value of the x-th station, D x K represents the high-density occurrence frequency of the area where the x-th station is located within three scheduling cycles, xz represents the number of vehicle resource gaps at the x-th station in time period z, m x is the number of time periods with gaps recorded at the x-th station, H x is the ranking factor of the x-th site in the region, F x The number of redundancy scheduling failures for the x-th site in the historical period.

6. The vehicle scheduling optimization system for the car rental platform according to claim 5 is characterized in that: The candidate path reconstruction module includes: The priority node extraction submodule filters the station numbers marked as priority transfer based on the regional overload node scheduling level table, establishes a transfer target node index, and extracts the adjacent area range of each target node to generate a set of transferable target nodes; The vehicle status calculation submodule extracts the vehicle's station number, current location, last return time, and current idle status in the adjacent area based on the set of transferable target nodes, calculates the driving distance from the current location to the target node and the return interval, and uses the station information for normalization to obtain a standardized vehicle status data table; The path screening and sorting submodule performs weighted calculation and sorting on the normalized driving distance and normalized response time of the vehicle based on the standardized vehicle status data table, and screens the vehicle numbers and corresponding path records whose scores are within the target range to generate a path-transferable vehicle dispatch group.

7. The vehicle scheduling optimization system for the car rental platform according to claim 6 is characterized in that: The adjustable vehicle screening module includes: The time limit construction submodule obtains orders marked as high urgency in the task list, extracts the creation time and maximum response time of each order, combines the two data into time limit record entries according to the task number, and generates the task time limit dataset; The response index extraction submodule calls the path transferable vehicle dispatch group, extracts the previous task completion time of the corresponding vehicle, the average time interval of the last three task calls, and the standard deviation of the historical dispatch response time, establishes an index correspondence according to the vehicle number, and obtains a vehicle response performance parameter table; The task vehicle screening submodule compares the three indicators of task time limit and vehicle response item by item based on the task time limit dataset and the vehicle response performance parameter table, determines whether each indicator is less than the task response limit standard, screens out any vehicle number that exceeds the limit, and marks the scheduling adaptation relationship between the remaining vehicles and the task, thereby obtaining a list of vehicles matching high-urgency tasks.

8. The vehicle scheduling optimization system for the car rental platform according to claim 7 is characterized in that: The system further comprises: The dispatch path binding module matches the vehicle list of the high-urgency task with the path and station information in the path-transferable vehicle dispatch group, reconstructs the vehicle ID, starting station ID, and target path segment in the dispatch instruction structure, and records the task number and dispatch level label in an entry format to generate task-bound path dispatch information; The task-bound path scheduling information includes a vehicle instruction structure item, a task scheduling number, a site path identifier, and a scheduling level item.

9. The vehicle scheduling optimization system for the car rental platform according to claim 8 is characterized in that: The scheduling path binding module includes: The path field matching submodule matches the vehicle list based on the high-urgency task, extracts the vehicle number and matches the corresponding starting station ID and target path segment number in the path transferable vehicle dispatch group, and extracts its vehicle ID information, constructs a three-field combination structure, and generates a path field association set; The instruction content update submodule calls the path field association set, replaces the corresponding positions in the scheduling instruction structure in the order of vehicle ID, starting station ID, and target path segment according to each group of vehicle field combinations, and records the task number corresponding to the replaced field to obtain a structured scheduling instruction set; The entry format generation submodule extracts the task number and the vehicle corresponding scheduling level identifier of each instruction based on the structured scheduling instruction set, integrates them into a parallel structure of task field and scheduling field, reorganizes them into entry-type standardized expression content, and generates task-bound path scheduling information.

10. A vehicle scheduling optimization method for a car rental platform, characterized in that: The method is used to implement the vehicle scheduling optimization system for a car rental platform according to any one of claims 1 to 9, comprising the following steps: S1: Obtain order records from the car rental platform, map the rental duration of each order to the time period, evaluate the saturation status of the region within the time period, and generate a regional continuous rental saturation identifier set; S2: calling the regional continuous rental saturation identification set, extracting the area marked as a high-density rental state, combining the vehicle gap and the high-frequency index value, calculating the station's dispatch level, and obtaining a regional overload node dispatch level table; S3: Based on the regional overload node dispatch level table, the average dispatch response time and the empty distance index of the vehicles are normalized and then weighted and ranked to screen the optimal path for executing the dispatch task and generate a dispatch group of path-transferable vehicles; S4: Obtain orders marked as high urgency in the task list, extract the vehicle's previous task completion time, the average of the last three call intervals, and the standard deviation of the historical response time, and compare them with the maximum response time limit of the order task. If any of the indicators exceeds the response limit, the vehicle number is removed to obtain a list of vehicles matching high urgency tasks; S5: Based on the high-urgency task matching vehicle list, the vehicle ID, starting station ID, and target path segment in the scheduling instruction structure are reconstructed, and the task number and scheduling level label are combined into an entry format to generate task-bound path scheduling information.

Images