Supply chain scheduling method based on port digital supply chain cloud platform management

Abstract

The present invention relates to the technical field of port dispatching management, and discloses a supply chain dispatching method based on port digital supply chain cloud platform management; the method comprises calculating the loading and unloading cycle of the port, dividing the loading and unloading cycle into continuous sub-periods, obtaining the comprehensive navigation data of the ship in the sub-period, predicting the vehicle demand for the next sub-period through a machine learning model, identifying the available vehicles from the vehicles to be dispatched, and performing supply chain dispatching; compared with the prior art, the present invention can refine and split the overall cargo loading and unloading process of the port terminal, reduce the amount of data in each sub-period, and combine with the prediction method of the machine learning model to evaluate and judge in advance whether the dispatching vehicles in the port terminal meet the supply chain dispatching needs, so that the ship cargo in the port terminal and the dispatching vehicles can achieve an efficient linkage effect, avoiding the phenomenon of accumulation of cargo at the port terminal due to the inability to transfer cargo in time.

Description

Technical Field

[0001] The present invention relates to the technical field of port scheduling management, and more specifically, to a supply chain scheduling method based on port digital supply chain cloud platform management. Background Art

[0002] As an important node in freight trade, the complexity and dynamism of port supply chain management are becoming increasingly prominent. In recent years, the rapid development of digital technology has brought new opportunities for the management of port supply chains, especially the construction of a digital supply chain cloud platform, which can promote collaboration among all participants, break information silos, and achieve optimal allocation of resources. Through data sharing and collaborative decision-making, ports can establish closer cooperative relationships with shipping companies, freight forwarders, logistics companies, etc., thereby improving the efficiency and flexibility of the port's overall supply chain.

[0003] The patent application with reference publication number CN117634832A discloses a supply chain scheduling method and system for coordinated agile production and port distribution. It focuses on the key links, production and distribution in the intelligent supply chain, and establishes a collaborative optimization model of the manufacturing and transportation network of the steel supply chain based on actual conditions. By collecting and analyzing the preset data of the production process and the port distribution stage, it can fully understand the various links and related elements of the supply chain, which is helpful to formulate targeted scheduling strategies; when determining the optimal solution, the preset necessary and sufficient conditions are considered, which ensures that the selected solution meets other important constraints such as cost and time while meeting the needs of production and distribution; the heuristic algorithm based on the dynamic programming algorithm can obtain the optimal solution under the current situation when all the ingots are known to be allocated to the soaking furnace, which is helpful to optimize the production process and improve resource utilization and production efficiency; based on the biased random key genetic algorithm and the flower pollination algorithm, a joint production and distribution plan is generated according to the obtained optimal solution. The application of this combined algorithm helps to optimize the overall scheduling arrangement of the supply chain, improve logistics efficiency and reduce costs; and can help manufacturers and transporters in the same supply chain make better operational decisions, with strong versatility and flexibility;

[0004] The existing port supply chain scheduling method collects the position data and movement data between ships and vehicles in the port in real time, and analyzes the collected position and movement data one by one through the background monitoring center, and adjusts the balance effect between the number of ships and vehicles in real time. Due to the real-time nature of the scheduling operation between ships and vehicles, it is impossible to discover and analyze the imbalance in the number of ships and vehicles that may occur in the future in advance, and it is impossible to perform advance scheduling operations based on the upcoming imbalance in quantity, which makes it easy for goods on the port terminal to pile up and accumulate due to the inability to transfer them in time, causing a lag in the port supply chain scheduling operation.

[0005] In view of this, the present invention proposes a supply chain scheduling method based on the management of the port digital supply chain cloud platform to solve the above problems. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solutions: a supply chain scheduling method based on port digital supply chain cloud platform management, applied to the scheduling cloud platform, comprising:

[0007] S1: Filter out the comprehensive loading and unloading data of the port from the database, and calculate the loading and unloading cycle of the port based on the comprehensive loading and unloading data. The comprehensive loading and unloading data includes the cargo loading and unloading time and the scheduling turnover time;

[0008] S2: Obtain the dispatch unit value of the loading and unloading cycle, and divide the loading and unloading cycle into continuous sub-periods based on the dispatch unit value;

[0009] S3: Obtain the comprehensive navigation data of the ship in the sub-period, which includes the proportion of the docking area, the regional cargo value and the synchronous docking rate;

[0010] S4: Input the comprehensive navigation data into the pre-trained machine learning model to predict the vehicle demand in the next sub-period and determine whether to issue a supply chain scheduling prompt; if a supply chain scheduling prompt is issued, execute S5; if no supply chain scheduling prompt is issued, repeat S3-S4;

[0011] S5: Identify available vehicles from the vehicles to be dispatched, and perform supply chain scheduling on the available vehicles according to supply chain scheduling requirements.

[0012] Further, the screening methods for cargo loading and unloading time include:

[0013] The loading and unloading log records the process of a ship loaded with cargo docking at a port terminal for the first time and loading all cargo on the ship onto vehicles through loading and unloading equipment;

[0014] Taking the time when the loading and unloading log is first generated as the starting time and the current time as the ending time, all loading and unloading logs are marked one by one from the database;

[0015] The detection status of all loading and unloading logs is queried one by one through the log management system, and the loading and unloading logs with the detection status of detected are recorded as valid logs, and i valid logs are obtained;

[0016] Query the time when the ship first docked at the port terminal and the time when all the cargo was loaded onto the vehicle in i valid logs one by one through the timestamp, and obtain i starting times and i ending times;

[0017] The duration between i start times and the corresponding i end times is recorded as the cargo loading and unloading duration, and i cargo loading and unloading durations are obtained.

[0018] Further, the screening method of scheduling turnaround time includes:

[0019] Mark the vehicles in i valid logs one by one and obtain i dispatched vehicles;

[0020] The dispatching management system is used to query the dispatching events of i dispatching vehicles one by one, and the entry time of entering the port terminal and the exit time of leaving the port terminal in the dispatching events are queried to obtain i entry times and i exit times;

[0021] The duration between the i entry times and the corresponding i exit times is recorded as the total dispatch duration, and i total dispatch durations are obtained;

[0022] Within the total dispatching time of i, the driving speed of i dispatched vehicles is detected in real time through the vehicle management system, and the period when the driving speed is 0 is recorded as the parking period, and i parking periods are obtained;

[0023] After subtracting the i total dispatching duration from the duration of the corresponding i parking periods, the i dispatching turnaround durations are obtained;

[0024] The expression of scheduling turnaround time is:

[0025] SC ddi =SC zi -SC sdi ;

[0026] In the formula, SC ddi is the i-th scheduling turnaround time, SC zi is the total scheduling duration of the i-th time, SC sdi is the duration of the i-th parking period;

[0027] The calculation method of loading and unloading cycle includes:

[0028] Remove the maximum and minimum values ​​of cargo loading and unloading time and scheduling turnover time respectively, and add up the remaining i-2 cargo loading and unloading time and the corresponding i-2 scheduling turnover time to get the average, and obtain the loading and unloading cycle;

[0029] The expression of loading and unloading cycle is:

[0030]

[0031] In the formula, ZX zq is the loading and unloading cycle, SC zxa is the loading and unloading time of the ath cargo, SC dda is the a-th scheduling turnaround time.

[0032] Furthermore, the method for obtaining the scheduling unit value includes:

[0033] The weight of i dispatched vehicles before and after loading is queried one by one through the weighing system, and i empty values ​​and i loaded values ​​are obtained;

[0034] Subtract the i loaded values ​​from the i empty values ​​one by one to obtain i loaded values, and then add up the i loaded values ​​and calculate the average to obtain the first unit value;

[0035] The expression for the first unit value is:

[0036]

[0037] Where DW z1 is the first unit value, ZZ zb is the bth loading value, KZ zb is the bth no-load value;

[0038] After comparing the loading and unloading period with the number of dispatched vehicles, a second unit value is obtained;

[0039] The expression for the second unit value is:

[0040]

[0041] Where DW z2 is the second unit value;

[0042] The first unit value and the second unit value are assigned different weight factors, and then added and averaged to obtain a scheduling unit value;

[0043] The expression for the dispatch unit value is:

[0044]

[0045] Where DW dd is the scheduling unit value, σ1 and σ2 are weight factors greater than 0.

[0046] Furthermore, the sub-period division method includes:

[0047] Mark the first and last moments in the loading and unloading cycle;

[0048] Taking the dispatch unit value as the division standard, the first moment as the marking starting point, and the last moment as the marking end point, k time period nodes are marked in the loading and unloading cycle;

[0049] The time period between two adjacent time period nodes is recorded as a sub-time period, and p sub-time periods are obtained.

[0050] Furthermore, the method for obtaining the parking area ratio value includes:

[0051] The electronic navigation chart of the port terminal is queried through the map database, and the location of the port terminal is marked on the electronic navigation chart as the terminal node;

[0052] With the wharf node as the base point and the preset length as the radius, a circular area is drawn on the electronic navigation chart and recorded as the area to be identified;

[0053] The water flow area on the electronic navigation chart is identified by computer vision technology, and the area where the water flow area overlaps with the area to be identified is recorded as the docking area;

[0054] The last moment in each of the p sub-periods is recorded as the detection moment, and the coordinate data of all ships at the detection moment are queried one by one through the positioning system;

[0055] Mark the points corresponding to the coordinate data on the electronic navigation chart one by one, record the coordinate data corresponding to the points located inside the berthing area as valid coordinates, record the ships corresponding to the valid coordinates as valid ships, and count the number of valid ships;

[0056] After comparing the number of valid ships in p sub-periods with the total number of ships in p sub-periods, the proportion values ​​of p berthing areas are obtained;

[0057] The expression of the docking area ratio is:

[0058]

[0059] Where, TK zbp is the proportion of the parking area in the pth sub-period, SL yxp is the number of valid ships in the pth sub-period, ZL cbp is the total number of ships in the pth sub-period.

[0060] Furthermore, the method for obtaining the regional cargo value includes:

[0061] In p sub-periods, the net weights of x valid ships are queried one by one through the technical parameter table to obtain x net weight values;

[0062] Through the weighing system, the real-time total weight of x valid ships is queried one by one to obtain x total load values;

[0063] Subtract the x total load values ​​from the x net weight values ​​one by one to obtain x sub-load values;

[0064] The expression of sub-cargo value is:

[0065] ZH zpx =ZZ zpx -JZzpx ;

[0066] In the formula, ZH zpx is the x-th sub-cargo value in the p-th sub-period, ZZ zpx is the xth total load value in the pth sub-period, JZ zpx is the xth net weight value in the pth sub-period;

[0067] After accumulating x sub-cargo values, we get the regional cargo value;

[0068] The expression of regional cargo value is:

[0069]

[0070] In the formula, QY zhp is the regional cargo value in the pth sub-period, ZH zpc is the cth sub-cargo value in the pth sub-period.

[0071] Furthermore, the method for obtaining the synchronous docking rate includes:

[0072] v speed points of equal duration are divided in p sub-periods respectively, and the real-time speeds of x effective ships at the v speed points are detected by speed sensors respectively to obtain v speed values;

[0073] After removing the maximum and minimum speed values, the remaining v-2 speed values ​​are accumulated and averaged to obtain x speed means;

[0074] The expression of the mean velocity is:

[0075]

[0076] In the formula, SD jzpx is the mean speed of the xth valid ship in the pth sub-period, SD zpxd is the d-th speed value of the x-th valid ship in the p-th sub-period;

[0077] The distances from x effective ships to the dock nodes are measured one by one using a scale to obtain x docking distances. After comparing the x docking distances with the x speed averages, x docking durations are obtained.

[0078] The expression for the stop duration is:

[0079]

[0080] Where, TK scpx is the berthing time of the xth valid ship in the pth sub-period, TK jlpx is the berthing distance of the xth valid ship in the pth sub-period;

[0081] The stop duration that is less than or equal to the dispatch unit value is recorded as the target duration, the number of target durations is counted, the number of p target durations is compared with the number of p stop durations, and p synchronous stop rates are obtained;

[0082] The expression of the synchronous docking rate is:

[0083]

[0084] Where, TK tbp is the synchronous parking rate in the pth sub-period, SL mbp is the number of target durations in the pth sub-period, SL tkp is the number of stop times in the pth sub-period.

[0085] Furthermore, the training method of the machine learning model includes:

[0086] Collecting multiple sets of comprehensive navigation data and vehicle demand quantities corresponding to the comprehensive navigation data in advance;

[0087] The comprehensive navigation data is converted into multiple feature vectors using a sliding window method. The vehicle demand is converted into labels corresponding to the comprehensive navigation data according to the sliding step. One feature vector corresponds to one label and constitutes a set of training data. Multiple sets of training data constitute a training set. The comprehensive navigation data are arranged in the order of collection time. The prediction time step Z, sliding step Q and sliding window length N are preset.

[0088] The feature vector is used as the input of the machine learning model, and the vehicle demand in the next sub-period after the time step Z is predicted as the output. The subsequent vehicle demand of each training set is used as the prediction target. The sum of the minimized prediction errors is used as the training target. The machine learning model is trained to generate a machine learning model that predicts the vehicle demand in the next sub-period based on the comprehensive navigation data of the previous sub-period.

[0089] The determination methods for whether to issue a supply chain scheduling reminder include:

[0090] The port terminal area is identified through computer vision technology, the number of vehicles in the port terminal area is counted, and the vehicle area value is obtained;

[0091] When the predicted vehicle demand in the next sub-period is greater than the vehicle area value, it is determined to issue a supply chain scheduling prompt;

[0092] When the predicted vehicle demand in the next sub-period is less than or equal to the vehicle area value, it is determined that no supply chain scheduling prompt will be issued.

[0093] Further, the supply chain scheduling method includes:

[0094] The coordinate data of s available vehicles are queried one by one through the positioning system to obtain s available coordinate values;

[0095] Mark the points where s available coordinate values ​​are located on the electronic navigation chart one by one to obtain s available points, and measure the distances from the s available points to the terminal nodes one by one to obtain s scheduling distance values;

[0096] Arrange the s dispatching distance values ​​in ascending order from small to large and number them;

[0097] According to the numbering from small to large, the s available vehicles are dispatched one by one to the port terminal area until the vehicle area value is greater than or equal to the predicted vehicle demand in the next sub-period, and the supply chain dispatch of the available vehicles is stopped.

[0098] The technical effects and advantages of the supply chain scheduling method based on the port digital supply chain cloud platform management of the present invention are:

[0099] The present invention screens out the comprehensive loading and unloading data of the port from the database, calculates the loading and unloading cycle of the port based on the comprehensive loading and unloading data, obtains the scheduling unit value of the supply chain scheduling, and divides the loading and unloading cycle into continuous sub-periods based on the scheduling unit value, obtains the comprehensive navigation data of the ship in the sub-period, the comprehensive navigation data includes the proportion of the berthing area, the regional cargo value and the synchronous berthing rate, inputs the comprehensive navigation data into a machine learning model trained in advance, predicts the vehicle demand for the next sub-period, and determines whether to issue a supply chain scheduling prompt, identifies available vehicles from the vehicles to be scheduled, and performs supply chain scheduling on the available vehicles according to the supply chain scheduling demand; compared with the prior art, the loading and unloading cycle can be calculated by collecting comprehensive loading and unloading data, and the loading and unloading cycle with a longer overall duration can be divided into By dividing it into multiple continuous and adjacent sub-periods, the overall cargo loading and unloading process of the port terminal can be refined and split, the amount of data in each sub-period can be reduced, and the accuracy of the comprehensive navigation data collection process can be enhanced. Combined with the prediction method of the machine learning model, it can effectively predict the vehicle demand in the future time period, so as to make an early assessment and judgment on whether the dispatched vehicles in the port terminal meet the supply chain scheduling needs. In the event of a shortage of dispatched vehicles, the supply chain scheduling operations can be carried out on the vehicles in advance, so that efficient linkage can be achieved between the ship cargo and the dispatched vehicles in the port terminal, avoiding the phenomenon of accumulation of goods at the port terminal due to the inability to transfer them in time, thereby effectively improving the digital scheduling effect between ships and vehicles at the port terminal and enhancing the rationality of supply chain scheduling. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] Figure 1A flow chart of a supply chain scheduling method based on port digital supply chain cloud platform management provided in the first embodiment of the present invention;

[0101] Figure 2 A schematic diagram of a supply chain scheduling system managed by a port digital supply chain cloud platform provided in Example 2 of the present invention. DETAILED DESCRIPTION

[0102] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0103] Example 1: Please refer to Figure 1 As shown, the supply chain scheduling method based on the port digital supply chain cloud platform management described in this embodiment is applied to the scheduling cloud platform, including:

[0104] S1: Filter out the comprehensive loading and unloading data of the port from the database, and calculate the loading and unloading cycle of the port based on the comprehensive loading and unloading data;

[0105] Comprehensive loading and unloading data refers to the data generated by the port's loading and unloading equipment when driving the cargo between vehicles and ships in the past period of time. It can represent the loading and unloading process of the port's loading and unloading equipment and serve as the data basis for the subsequent calculation of the port's loading and unloading cycle;

[0106] Comprehensive loading and unloading data includes cargo loading and unloading time and dispatch turnaround time;

[0107] The cargo loading and unloading time refers to the time taken for a ship loaded with cargo to dock at a port loading and unloading terminal and for the cargo on the ship to be loaded onto a vehicle through loading and unloading equipment. This can be used to numerically represent the cargo transfer time between the ship and the vehicle. The longer the cargo loading and unloading time, the longer the cargo transfer time between the ship and the vehicle, and the longer the loading and unloading cycle.

[0108] The screening methods for cargo handling time include:

[0109] The loading and unloading log records the process of a ship loaded with cargo docking at a port terminal for the first time and loading all cargo on the ship onto vehicles through loading and unloading equipment;

[0110] Taking the time when the loading and unloading log is first generated as the starting time and the current time as the ending time, all loading and unloading logs are marked one by one from the database;

[0111] The detection status of all loading and unloading logs is queried one by one through the log management system, and the loading and unloading logs with the detection status of detected are recorded as valid logs, and i valid logs are obtained; the detection status is used to indicate whether the data content in the loading and unloading log has passed the verification test, so as to indicate the rationality of the data content in the loading and unloading log; the detection status includes detected and undetected, detected means that the data content in the loading and unloading log has passed the verification test and can be used directly, and undetected means that the data content in the loading and unloading log has not passed the verification test and cannot be used directly;

[0112] Query the time when the ship first docked at the port terminal and the time when all the cargo was loaded onto the vehicle in i valid logs one by one through the timestamp, and obtain i starting times and i ending times;

[0113] The duration between i start times and the corresponding i end times is recorded as the cargo loading and unloading duration, and i cargo loading and unloading durations are obtained.

[0114] The dispatching turnaround time refers to the time between a vehicle that needs to load cargo and entering the port terminal after receiving an entry instruction and leaving the port terminal after receiving an exit instruction. It can be used to represent the turnaround time of the vehicle dispatching process in the port terminal. The longer the dispatching turnaround time is, the longer the vehicle dispatching turnaround time is in the port terminal, and the longer the loading and unloading cycle is.

[0115] The screening methods for scheduling turnaround time include:

[0116] Mark the vehicles in i valid logs one by one and obtain i dispatched vehicles;

[0117] The dispatching management system is used to query the dispatching events of i dispatching vehicles one by one, and the entry time of entering the port terminal and the exit time of leaving the port terminal in the dispatching event are queried to obtain i entry times and i exit times; the dispatching event is an event used to dispatch vehicles from entering the port terminal to leaving the port terminal, which can represent the driving process of the dispatching vehicle;

[0118] The duration between the i entry times and the corresponding i exit times is recorded as the total dispatch duration, and i total dispatch durations are obtained;

[0119] Within the total dispatching time of i, the driving speed of i dispatched vehicles is detected in real time through the vehicle management system, and the period when the driving speed is 0 is recorded as the parking period, and i parking periods are obtained;

[0120] After subtracting the i total dispatching duration from the duration of the corresponding i parking periods, the i dispatching turnaround durations are obtained;

[0121] The expression of scheduling turnaround time is:

[0122] SC ddi =SC zi -SC sdi ;

[0123] In the formula, SC ddi is the i-th scheduling turnaround time, SC zi is the total scheduling duration of the ith item, SC sdi is the duration of the i-th parking period.

[0124] After the cargo loading and unloading time and the scheduling turnover time are screened out, the loading and unloading cycle can be calculated based on the cargo loading and unloading time and the scheduling turnover time, so that the loading and unloading cycle can be used as the time basis for subsequent port supply chain scheduling;

[0125] The calculation method of loading and unloading cycle includes:

[0126] Remove the maximum and minimum values ​​of cargo loading and unloading time and scheduling turnover time respectively, and add up the remaining i-2 cargo loading and unloading time and the corresponding i-2 scheduling turnover time to get the average, and obtain the loading and unloading cycle;

[0127] The expression of loading and unloading cycle is:

[0128]

[0129] In the formula, ZX zq is the loading and unloading cycle, SC zxa is the loading and unloading time of the ath cargo, SC dda is the a-th scheduling turnaround time.

[0130] S2: Obtain the dispatch unit value of the loading and unloading cycle, and divide the loading and unloading cycle into continuous sub-periods based on the dispatch unit value;

[0131] Since the duration of the loading and unloading cycle is relatively long, the data changes in the supply chain scheduling of ships and vehicles at the port during the loading and unloading cycle are also relatively large, and it is impossible to accurately represent the supply chain scheduling change data within a shorter period of time. In order to achieve data collection and analysis effects within a shorter period of time, it is necessary to divide the loading and unloading cycle into sub-periods, so that each sub-period can represent the supply chain scheduling at the port for a reasonable duration and ensure the integrity of the data within each sub-period;

[0132] The dispatch unit value is used to numerically represent the reasonable duration of the supply chain dispatch at the port, so as to ensure that the dispatch unit value corresponds to the time period in which the relevant data of the supply chain dispatch at the port can be reasonably and completely obtained. The larger the dispatch unit value, the longer the duration corresponding to each sub-period, and vice versa.

[0133] Methods for obtaining the dispatch unit value include:

[0134] The weight of i dispatched vehicles before and after loading is queried one by one through the weighing system, and i empty values ​​and i loaded values ​​are obtained;

[0135] Subtract the i loaded values ​​from the i empty values ​​one by one to obtain i loaded values, and then add up the i loaded values ​​and calculate the average to obtain the first unit value;

[0136] The expression for the first unit value is:

[0137]

[0138] Where DW z1 is the first unit value, ZZ zb is the bth loading value, KZ zb is the bth no-load value;

[0139] After comparing the loading and unloading period with the number of dispatched vehicles, a second unit value is obtained;

[0140] The expression for the second unit value is:

[0141]

[0142] Where DW z2 is the second unit value;

[0143] The first unit value and the second unit value are assigned different weight factors, and then added and averaged to obtain a scheduling unit value;

[0144] The expression for the dispatch unit value is:

[0145]

[0146] Where DW dd is the scheduling unit value, σ1 and σ2 are weight factors greater than 0;

[0147] Among them, σ1+σ2=1, for example, σ1 is 0.32, and σ2 is 0.68.

[0148] When the dispatch unit value is obtained, the dispatch unit value can be used as a standard to divide the loading and unloading cycle into multiple continuous sub-periods, so that each sub-period can play a separate role in the port's supply chain scheduling;

[0149] The sub-period division methods include:

[0150] Mark the first and last moments in the loading and unloading cycle;

[0151] Taking the dispatch unit value as the division standard, the first moment as the marking starting point, and the last moment as the marking end point, k time period nodes are marked in the loading and unloading cycle;

[0152] The time period between two adjacent time period nodes is recorded as a sub-time period, and p sub-time periods are obtained.

[0153] It should be noted that the divided p sub-periods are continuous and uninterrupted on the timeline, and the last moment of the previous sub-period is the first moment of the next sub-period, which can ensure that the two adjacent sub-periods can continuously divide the loading and unloading cycle.

[0154] S3: Obtain the comprehensive navigation data of the ship in the sub-period, which includes the proportion of the docking area, the regional cargo value and the synchronous docking rate;

[0155] Comprehensive navigation data refers to the relevant comprehensive data when a ship loaded with cargo needs vehicles to load cargo when it docks at a port terminal, which can represent the data that affects the number of vehicles, thereby providing data support for the subsequent demand value of the number of vehicles;

[0156] Comprehensive navigation data includes the proportion of docking areas, regional cargo value and simultaneous docking rate;

[0157] The docking area value refers to the ratio of the number of ships entering the docking area of ​​the port terminal in each sub-period to the total number of ships, which can be used to represent the number of ships that will dock at the port terminal. The larger the docking area value, the greater the ratio of the number of ships entering the docking area of ​​the port terminal in the sub-period to the total number of ships, and more vehicles are needed to cooperate in loading and unloading cargo;

[0158] Methods for obtaining the docking area ratio include:

[0159] The electronic navigation chart of the port terminal is queried through the map database, and the location of the port terminal is marked on the electronic navigation chart as the terminal node;

[0160] With the wharf node as the base point and the preset length as the radius, a circular area is drawn on the electronic navigation chart, which is recorded as the area to be identified. The preset length refers to the length between the boundary of the area to be identified on the electronic navigation chart and the wharf node, which can represent the farthest boundary of the area to be identified, thereby providing the accuracy of the area to be identified.

[0161] The water flow area on the electronic navigation chart is identified by computer vision technology, and the area where the water flow area overlaps with the area to be identified is recorded as the docking area;

[0162] The last moment in each of the sub-periods is recorded as the detection moment, and the coordinate data of all ships at the detection moment are queried one by one through the positioning system;

[0163] Mark the points corresponding to the coordinate data on the electronic navigation chart one by one, record the coordinate data corresponding to the points located inside the berthing area as valid coordinates, record the ships corresponding to the valid coordinates as valid ships, and count the number of valid ships;

[0164] After comparing the number of valid ships in p sub-periods with the total number of ships in p sub-periods, the proportion values ​​of p berthing areas are obtained;

[0165] The expression of the docking area ratio is:

[0166]

[0167] Where, TK zbp is the proportion of the parking area in the pth sub-period, SL yxp is the number of valid ships in the pth sub-period, ZL cbp is the total number of ships in the pth sub-period.

[0168] The regional cargo value refers to the total cargo volume of ships entering the port terminal berthing area in each sub-period, which can be used to indicate the cargo volume of ships that are about to dock at the port terminal. The larger the regional cargo value, the larger the total cargo volume of ships entering the port terminal berthing area in the sub-period, and more vehicles are needed to cooperate in loading and unloading cargo;

[0169] Methods for obtaining regional cargo value include:

[0170] In p sub-periods, the net weights of x valid ships are queried one by one through the technical parameter table to obtain x net weight values;

[0171] Through the weighing system, the real-time total weight of x valid ships is queried one by one to obtain x total load values;

[0172] Subtract the x total load values ​​from the x net weight values ​​one by one to obtain x sub-load values;

[0173] The expression of sub-cargo value is:

[0174] ZH zpx =ZZ zpx -JZ zpx ;

[0175] In the formula, ZH zpx is the x-th sub-cargo value in the p-th sub-period, ZZ zpx is the xth total load value in the pth sub-period, JZ zpx is the xth net weight value in the pth sub-period;

[0176] After accumulating x sub-cargo values, we get the regional cargo value;

[0177] The expression of regional cargo value is:

[0178]

[0179] In the formula, QY zhp is the regional cargo value in the pth sub-period, ZH zpc is the cth sub-cargo value in the pth sub-period.

[0180] The synchronous docking rate refers to the ratio of the number of ships docked at the port terminal in the docking area at the same time in each sub-period to the total number of ships in the docking area, which can be used to represent the number of ships docked at the port terminal at the same time. The larger the synchronous docking rate, the larger the ratio of the number of ships docked at the port terminal in the docking area at the same time in the sub-period to the total number of ships in the docking area, and more vehicles are needed to cooperate in loading and unloading cargo;

[0181] Methods for obtaining the synchronous docking rate include:

[0182] v speed points of equal duration are divided in p sub-periods respectively, and the real-time speeds of x effective ships at the v speed points are detected by speed sensors respectively to obtain v speed values;

[0183] After removing the maximum and minimum speed values, the remaining v-2 speed values ​​are accumulated and averaged to obtain x speed means;

[0184] The expression of the mean velocity is:

[0185]

[0186] In the formula, SD jzpx is the mean speed of the xth valid ship in the pth sub-period, SD zpxd is the d-th speed value of the x-th valid ship in the p-th sub-period;

[0187] The distances from x effective ships to the dock nodes are measured one by one using a scale to obtain x docking distances. After comparing the x docking distances with the x speed averages, x docking durations are obtained.

[0188] The expression for the stop duration is:

[0189]

[0190] Where, TK scpx is the berthing time of the xth valid ship in the pth sub-period, TK jlpxis the berthing distance of the xth valid ship in the pth sub-period;

[0191] The stop duration that is less than or equal to the dispatch unit value is recorded as the target duration, the number of target durations is counted, the number of p target durations is compared with the number of p stop durations, and p synchronous stop rates are obtained;

[0192] The expression of the synchronous docking rate is:

[0193]

[0194] Where, TK tbp is the synchronous parking rate in the pth sub-period, SL mbp is the number of target durations in the pth sub-period, SL tkp is the number of stop times in the pth sub-period.

[0195] S4: Input the comprehensive navigation data into the pre-trained machine learning model to predict the vehicle demand in the next sub-period and determine whether to issue a supply chain scheduling prompt;

[0196] After obtaining the comprehensive navigation data, the comprehensive navigation data can be imported into the trained machine learning model, and the vehicle demand in the next sub-period can be predicted through the machine learning model, thereby realizing the prediction effect of the number of vehicles at the port terminal at the future time, so that the cargo on the ships docked at the port terminal can be loaded and transported by sufficient vehicles, so as to achieve the early and accurate scheduling effect of the port supply chain;

[0197] Vehicle demand refers to the number of vehicles actually required based on comprehensive navigation data, and is used to accurately represent the number of vehicles. Vehicle demand is obtained by collecting a large number of vehicle numbers corresponding to different parking area proportions, regional cargo values, and synchronous parking rates.

[0198] Methods for training machine learning models include:

[0199] Collecting multiple sets of comprehensive navigation data and vehicle demand quantities corresponding to the comprehensive navigation data in advance;

[0200] The comprehensive navigation data is converted into multiple feature vectors using a sliding window method. The vehicle demand is converted into labels corresponding to the comprehensive navigation data according to the sliding step. One feature vector corresponds to one label and constitutes a set of training data. Multiple sets of training data constitute a training set. The comprehensive navigation data are arranged in the order of collection time. The prediction time step Z, sliding step Q and sliding window length N are preset.

[0201] The feature vector is used as the input of the machine learning model, and the vehicle demand in the next sub-period after the predicted time step Z is used as the output. The subsequent vehicle demand in each training set is used as the prediction target. The machine learning model is trained with the minimized sum of prediction errors as the training target to generate a machine learning model that predicts the vehicle demand in the next sub-period based on the comprehensive navigation data of the previous sub-period.

[0202] Exemplarily, the machine learning model is any one of CNN or AlexNet;

[0203] The calculation formula for the prediction error is:

[0204] zk=(ak-wk) 2 ;

[0205] Where zk is the prediction error, k is the group number of the feature vector; ak is the predicted state value corresponding to the kth group of feature vectors, and wk is the actual state value corresponding to the kth group of training data;

[0206] In the machine learning model, the feature vector is the comprehensive navigation data, and the state value is the vehicle demand;

[0207] Other model parameters of the machine learning model, target loss value, optimization algorithm, training set test set validation set ratio, and loss function optimization are all achieved through actual engineering implementation and continuous experimental tuning.

[0208] After the collected comprehensive navigation data is imported into the machine learning model, the vehicle demand in the next sub-period can be predicted. The predicted vehicle demand is compared with the number of vehicles in the port terminal to determine whether a supply chain scheduling prompt needs to be issued;

[0209] The determination methods for whether to issue a supply chain scheduling reminder include:

[0210] The port terminal area is identified through computer vision technology, the number of vehicles in the port terminal area is counted, and the vehicle area value is obtained;

[0211] When the predicted vehicle demand for the next sub-period is greater than the vehicle area value, it means that the vehicles in the port terminal cannot meet the vehicle demand for the next sub-period. At this time, there will be a shortage of vehicles, and it is determined to issue a supply chain scheduling prompt;

[0212] When the predicted vehicle demand for the next sub-period is less than or equal to the vehicle area value, it means that the vehicles in the port terminal can meet the vehicle demand for the next sub-period, and there will be no vehicle shortage at this time, so it is determined not to issue a supply chain scheduling prompt.

[0213] S5: Identify available vehicles from the vehicles to be dispatched, and perform supply chain dispatch on the available vehicles according to supply chain dispatch requirements;

[0214] The vehicles to be dispatched refer to all vehicles located outside the port terminal area, so that the vehicles to be dispatched can serve as the vehicle basis for subsequent supply chain scheduling, while the available vehicles are the vehicles among the vehicles to be dispatched that can be used for supply chain scheduling operations, and thus serve as the objects of subsequent scheduling;

[0215] Since the available vehicles are used for the subsequent dispatching and transportation of the cargo on the ship, the available vehicles need to be kept in an empty state. In order to identify the available vehicles from the vehicles to be dispatched, it is necessary to obtain the loading value and empty value of the vehicles to be dispatched in the weighing system. When the loading value and empty value of the vehicles to be dispatched are equal, the vehicles to be dispatched at this time are in a state of not loading cargo. Then the vehicles to be dispatched with equal loading value and empty value are identified as available vehicles, and s available vehicles are obtained;

[0216] After obtaining s available vehicles, since the distances between each available vehicle and the port terminal are inconsistent, the time points at which each available vehicle arrives at the port terminal to load cargo will also be different. Therefore, it is necessary to perform supply chain scheduling operations on the s available vehicles. When performing supply chain scheduling on the available vehicles, it is necessary to schedule them according to the demand for the number of available vehicles corresponding to the supply chain scheduling to ensure that the available vehicles can meet the cargo loading and unloading needs of the ships at the port terminal.

[0217] Supply chain scheduling methods include:

[0218] The coordinate data of s available vehicles are queried one by one through the positioning system to obtain s available coordinate values;

[0219] Mark the points where s available coordinate values ​​are located on the electronic navigation chart one by one to obtain s available points, and measure the distances from the s available points to the terminal nodes one by one to obtain s scheduling distance values;

[0220] Arrange the s dispatching distance values ​​in ascending order from small to large and number them;

[0221] According to the numbering from small to large, the s available vehicles are dispatched one by one to the port terminal area until the vehicle area value is greater than or equal to the predicted vehicle demand in the next sub-period, and the supply chain dispatch of the available vehicles is stopped.

[0222] It should be noted that when s available vehicles are scheduling the supply chain, they are scheduled one by one and independently, so that one scheduling can drive one available vehicle to enter the port terminal area to load and unload the cargo on the ship, and when the available vehicle enters the port terminal area and meets the loading and unloading needs of the cargo on the ship, the available vehicle will no longer perform supply chain scheduling operations.

[0223] In this embodiment, the comprehensive loading and unloading data of the port is screened out from the database, and the loading and unloading cycle of the port is calculated based on the comprehensive loading and unloading data, the scheduling unit value of the supply chain scheduling is obtained, and the loading and unloading cycle is divided into continuous sub-periods based on the scheduling unit value, and the comprehensive navigation data of the ship in the sub-period is obtained. The comprehensive navigation data includes the proportion of the berthing area, the regional cargo value and the synchronous berthing rate. The comprehensive navigation data is input into the machine learning model trained in advance, and the vehicle demand for the next sub-period is predicted, and it is determined whether to issue a supply chain scheduling prompt, and the available vehicles are identified from the vehicles to be scheduled, and the supply chain scheduling is performed on the available vehicles according to the supply chain scheduling requirements. Compared with the prior art, the loading and unloading cycle can be calculated by collecting comprehensive loading and unloading data, and the loading and unloading cycle with a longer overall duration can be calculated by dividing the loading and unloading cycle into a continuous sub-period. By dividing it into multiple continuous and adjacent sub-periods, the overall cargo loading and unloading process of the port terminal can be refined and split, the amount of data in each sub-period can be reduced, and the accuracy of the comprehensive navigation data collection process can be enhanced. Combined with the prediction method of the machine learning model, it can effectively predict the vehicle demand in the future time period, so as to make an early assessment and judgment on whether the dispatched vehicles in the port terminal meet the supply chain scheduling needs. In the event of a shortage of dispatched vehicles, the supply chain scheduling operations can be carried out on the vehicles in advance, so that efficient linkage can be achieved between the ship cargo and the dispatched vehicles in the port terminal, avoiding the phenomenon of accumulation of goods at the port terminal due to the inability to transfer goods in time, thereby effectively improving the digital scheduling effect between ships and vehicles at the port terminal and enhancing the rationality of supply chain scheduling.

[0224] Example 2: Please refer to Figure 2 As shown, the part not described in detail in this embodiment is described in the first embodiment, and a supply chain scheduling system based on the management of the port digital supply chain cloud platform is provided, which is applied to the scheduling cloud platform to implement a supply chain scheduling method based on the management of the port digital supply chain cloud platform, including a loading and unloading cycle module, a time period division module, a data acquisition module, a model prediction module and a supply scheduling module, wherein each module is connected via a wired or wireless network;

[0225] The loading and unloading cycle module is used to filter out the comprehensive loading and unloading data of the port from the database and calculate the loading and unloading cycle of the port based on the comprehensive loading and unloading data. The comprehensive loading and unloading data includes the cargo loading and unloading time and the scheduling turnover time.

[0226] The time period division module is used to obtain the dispatch unit value of the loading and unloading cycle and divide the loading and unloading cycle into continuous sub-periods based on the dispatch unit value;

[0227] The data collection module is used to obtain the comprehensive navigation data of the ship in the sub-period, and the comprehensive navigation data includes the proportion of the docking area, the regional cargo value and the synchronous docking rate;

[0228] The model prediction module is used to input the comprehensive navigation data into the pre-trained machine learning model to predict the vehicle demand in the next sub-period and determine whether to issue a supply chain scheduling prompt;

[0229] The supply scheduling module is used to identify available vehicles from the vehicles to be scheduled and perform supply chain scheduling on the available vehicles according to supply chain scheduling requirements.

[0230] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A supply chain scheduling method based on the management of the port digital supply chain cloud platform is applied to the scheduling cloud platform, which is characterized by: include: S1: Filter out the comprehensive loading and unloading data of the port from the database, and calculate the loading and unloading cycle of the port based on the comprehensive loading and unloading data. The comprehensive loading and unloading data includes the cargo loading and unloading time and the scheduling turnover time; The calculation method of loading and unloading cycle includes: Remove the maximum and minimum values ​​of cargo loading and unloading time and scheduling turnover time respectively, and add up the remaining i-2 cargo loading and unloading time and the corresponding i-2 scheduling turnover time to get the average, and obtain the loading and unloading cycle; S2: Obtain the dispatch unit value of the loading and unloading cycle, and divide the loading and unloading cycle into continuous sub-periods based on the dispatch unit value; Methods for obtaining the dispatch unit value include: The weight of i dispatched vehicles before and after loading is queried one by one through the weighing system, and i empty values ​​and i loaded values ​​are obtained; Subtract the i loaded values ​​from the i empty values ​​one by one to obtain i loaded values, and then add up the i loaded values ​​and calculate the average to obtain the first unit value; After comparing the loading and unloading period with the number of dispatched vehicles, a second unit value is obtained; The first unit value and the second unit value are assigned different weight factors, and then added and averaged to obtain a scheduling unit value; S3: Obtain the comprehensive navigation data of the ship in the sub-period, which includes the proportion of the docking area, the regional cargo value and the synchronous docking rate; Methods for obtaining the docking area ratio include: The electronic navigation chart of the port terminal is queried through the map database, and the location of the port terminal is marked on the electronic navigation chart as the terminal node; With the wharf node as the base point and the preset length as the radius, a circular area is drawn on the electronic navigation chart and recorded as the area to be identified; The water flow area on the electronic navigation chart is identified by computer vision technology, and the area where the water flow area overlaps with the area to be identified is recorded as the docking area; The last moment in each of the p sub-periods is recorded as the detection moment, and the coordinate data of all ships at the detection moment are queried one by one through the positioning system; Mark the points corresponding to the coordinate data on the electronic navigation chart one by one, record the coordinate data corresponding to the points located inside the berthing area as valid coordinates, record the ships corresponding to the valid coordinates as valid ships, and count the number of valid ships; After comparing the number of valid ships in p sub-periods with the total number of ships in p sub-periods, the proportion values ​​of p berthing areas are obtained; Methods for obtaining regional cargo value include: In p sub-periods, the net weights of x valid ships are queried one by one through the technical parameter table to obtain x net weight values; Through the weighing system, the real-time total weight of x valid ships is queried one by one to obtain X total load values; Subtract the X total load values ​​from the X net weight values ​​one by one to obtain the x sub-load values; After accumulating X sub-cargo values, we get the regional cargo value; Methods for obtaining the synchronous docking rate include: v speed points of equal duration are divided in p sub-periods respectively, and the real-time speeds of x effective ships at the v speed points are detected by speed sensors respectively to obtain v speed values; After removing the maximum and minimum speed values, the remaining v-2 speed values ​​are accumulated and averaged to obtain x speed means; The distances from x effective ships to the dock nodes are measured one by one using a scale to obtain x docking distances. After comparing the x docking distances with the x speed averages, x docking durations are obtained. The stop duration that is less than or equal to the dispatch unit value is recorded as the target duration, the number of target durations is counted, the number of p target durations is compared with the number of p stop durations, and p synchronous stop rates are obtained; S4: Input the comprehensive navigation data into the pre-trained machine learning model to predict the vehicle demand in the next sub-period and determine whether to issue a supply chain scheduling prompt; if a supply chain scheduling prompt is issued, execute S5; if no supply chain scheduling prompt is issued, repeat S3-S4; S5: Identify available vehicles from the vehicles to be dispatched, and perform supply chain scheduling on the available vehicles according to supply chain scheduling requirements.

2. The supply chain scheduling method based on port digital supply chain cloud platform management according to claim 1 is characterized in that: The method for screening the cargo loading and unloading time includes: The loading and unloading log records the process of a ship loaded with cargo docking at a port terminal for the first time and loading all cargo on the ship onto vehicles through loading and unloading equipment; Taking the time when the loading and unloading log is first generated as the starting time and the current time as the ending time, all loading and unloading logs are marked one by one from the database; The detection status of all loading and unloading logs is queried one by one through the log management system, and the loading and unloading logs with the detection status of detected are recorded as valid logs, and i valid logs are obtained; Query the time when the ship first docked at the port terminal and the time when all the cargo was loaded onto the vehicle in i valid logs one by one through the timestamp, and obtain i starting times and i ending times; The duration between i start times and the corresponding i end times is recorded as the cargo loading and unloading duration, and i cargo loading and unloading durations are obtained.

3. The supply chain scheduling method based on port digital supply chain cloud platform management according to claim 2 is characterized in that: The screening method for the scheduling turnaround time includes: Mark the vehicles in i valid logs one by one and obtain i dispatched vehicles; The dispatching events of i dispatching vehicles are queried one by one through the dispatching management system, and the entry time of entering the port terminal and the exit time of leaving the port terminal in the dispatching event are queried to obtain i entry times and i exit times; The duration between the i entry times and the corresponding i exit times is recorded as the total dispatch duration, and i total dispatch durations are obtained; Within the total dispatching time of i, the driving speed of i dispatched vehicles is detected in real time through the vehicle management system, and the period when the driving speed is 0 is recorded as the parking period, and i parking periods are obtained; After subtracting the i total dispatching times from the durations of the corresponding i parking periods, we get i dispatching turnaround times.

4. The supply chain scheduling method based on port digital supply chain cloud platform management according to claim 3 is characterized in that: The method for dividing the sub-periods includes: Mark the first and last moments in the loading and unloading cycle; Taking the dispatch unit value as the division standard, the first moment as the marking starting point, and the last moment as the marking end point, k time period nodes are marked in the loading and unloading cycle; The time period between two adjacent time period nodes is recorded as a sub-time period, and p sub-time periods are obtained.

5. The supply chain scheduling method based on port digital supply chain cloud platform management according to claim 4 is characterized in that: The training method of the machine learning model includes: Collecting multiple sets of comprehensive navigation data and vehicle demand quantities corresponding to the comprehensive navigation data in advance; The comprehensive navigation data is converted into multiple feature vectors using a sliding window method. The vehicle demand is converted into labels corresponding to the comprehensive navigation data according to the sliding step. One feature vector corresponds to one label and constitutes a set of training data. Multiple sets of training data constitute a training set. The comprehensive navigation data are arranged in the order of collection time. The prediction time step Z, sliding step Q and sliding window length N are preset. The feature vector is used as the input of the machine learning model, and the vehicle demand in the next sub-period after the time step Z is predicted as the output. The subsequent vehicle demand of each training set is used as the prediction target. The sum of the minimized prediction errors is used as the training target. The machine learning model is trained to generate a machine learning model that predicts the vehicle demand in the next sub-period based on the comprehensive navigation data of the previous sub-period. The determination methods for whether to issue a supply chain scheduling reminder include: The port terminal area is identified through computer vision technology, the number of vehicles in the port terminal area is counted, and the vehicle area value is obtained; When the predicted vehicle demand in the next sub-period is greater than the vehicle area value, it is determined to issue a supply chain scheduling prompt; When the predicted vehicle demand in the next sub-period is less than or equal to the vehicle area value, it is determined that no supply chain scheduling prompt will be issued.

6. The supply chain scheduling method based on port digital supply chain cloud platform management according to claim 5 is characterized in that: The supply chain scheduling method includes: The coordinate data of s available vehicles are queried one by one through the positioning system to obtain s available coordinate values; Mark the points where s available coordinate values ​​are located on the electronic navigation chart one by one to obtain s available points, and measure the distances from the s available points to the terminal nodes one by one to obtain s scheduling distance values; Arrange the s dispatching distance values ​​in ascending order from small to large and number them; According to the numbering from small to large, s available vehicles are dispatched one by one to the port terminal area until the vehicle area value is greater than or equal to the predicted vehicle demand in the next sub-period, and the supply chain dispatch of available vehicles is stopped.

Images