Port Project Full-Process Decision Management System Based on Big Data Analytics

Abstract

This invention belongs to the field of port operation and management technology. It provides a port project full-process decision management system based on big data analysis, comprising: verifying and analyzing multiple historical estimated arrival times using a hypothetical moving window; determining the optimal moving window using the moving window averaging method and Euclidean algorithm; obtaining a model correction factor based on historical estimates and actual times; obtaining the estimated arrival time; comparing it with the waiting-to-departure time; if a mismatch is found, removing the original stable connecting vessels; reselecting vessels based on connection matching degree and re-evaluating; until a vessel with a time-dimensional connection match is found within the connecting vessel sequence. By optimizing the moving window and introducing correction factors, the accuracy of arrival time prediction is improved, avoiding delays due to time mismatches and effectively improving port vessel scheduling efficiency. This invention solves the problem of how to accurately predict the arrival time of stable connecting vessels and determine whether it matches their waiting-to-departure time.

Description

Technical Field

[0001] This invention belongs to the field of port operation and management technology, specifically a port project full-process decision management system based on big data analysis. Background Technology

[0002] With the increasing prosperity of global trade, ports, as key hubs for cargo transportation, directly affect the development of regional economies through their operational efficiency. The actual operation of port projects involves multiple complex and interconnected links, including wharf layout planning, wharf layout design and construction, and ship berthing and cargo loading and unloading. Among these, ship berthing and cargo loading and unloading is a crucial step in realizing the core function of ports—cargo transportation and transshipment—and its operational efficiency plays a decisive role in the overall benefits of ports.

[0003] In the existing technology, in terms of ship scheduling, due to the lack of scientific and effective decision-making basis, there are often time conflicts between ships waiting to unload cargo and ships waiting to load cargo, resulting in idle or excessively strained loading and unloading equipment and manpower, which greatly reduces loading and unloading efficiency.

[0004] Therefore, this application obtains the berthing and loading departure times of ships waiting to be unloaded from the port project schedule, calculates the time matching degree based on the loading departure time, screens out connecting ships according to a preset range and constructs a sequence, extracts ship cargo information within the sequence, compares cargo types to obtain cargo type matching degree, compares cargo flow direction to obtain cargo flow matching degree, calculates the connection matching degree through a geometric product model, and selects connecting ship sequences accordingly. This solves the problem of how to efficiently select connecting ships waiting to be unloaded and loaded, and avoids problems such as low loading and unloading efficiency, resource waste and transportation disruptions caused by time conflicts, cargo type mismatches and inconsistent cargo flow directions.

[0005] Select any connecting vessel from the connecting vessel sequence, set a historical period and divide it into time periods, obtain the expected and actual arrival times of the vessel in each time period from the shipping schedule to form an arrival analysis group, input all arrival analysis groups into the Pearson distance calculation model to obtain the stable value of arrival time, and select the vessel corresponding to the maximum stable value as the stable connecting vessel, which helps the port to accurately identify vessels with stable arrival times and avoid difficulties in cargo loading and unloading due to unstable vessel arrival times.

[0006] By assuming a moving window, multiple historical estimated arrival times were validated and analyzed. The optimal moving window was determined using the moving window averaging method and the Euclidean algorithm. The model correction factor was obtained based on the historical estimated and actual times to obtain the estimated arrival time. This time was then compared with the waiting departure time. If there was a mismatch, the original stable connecting vessels were removed, and vessels were reselected and re-evaluated according to the connection matching degree until a vessel with a time dimension connection matching was found in the connecting vessel sequence. This solved the problem of how to accurately predict the arrival time of stable connecting vessels and determine whether it matches the waiting departure time.

[0007] To this end, the present invention provides a port project full-process decision management system based on big data analysis. Summary of the Invention

[0008] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0009] The technical solution adopted by this invention to solve its technical problem is:

[0010] The port project full-process decision management system based on big data analysis includes the following modules:

[0011] Vessel matching and initial screening module: Obtain the estimated arrival time and departure time of each vessel before unloading, evaluate the time matching degree between the estimated arrival time before unloading and the departure time before loading, and construct the initial screening connection sequence;

[0012] Connection matching and screening module: Within the initial connection sequence, the cargo type matching degree and cargo flow matching degree of the initial connection vessels are obtained respectively, and a comprehensive analysis is performed with the vessels to be loaded to screen out the connection vessel sequence.

[0013] Connection Stability Analysis Module: Based on multiple historical estimated arrival times of connecting vessels within the connection vessel sequence over historical periods, an arrival stability analysis is performed to obtain stable connection vessels;

[0014] Connection matching assessment module: Based on multiple historical estimated arrival times of stably connected vessels, predict the estimated arrival time and compare it with the departure time to assess whether the connection time matches. If they do not match, re-filter from the connection vessel sequence until a vessel with a matching connection is found.

[0015] A further aspect of this invention is the construction of a preliminary screening linkage sequence, the process of which is as follows:

[0016] The deviation between the estimated arrival time of the unloading vessel and the departure time of the loading vessel is obtained, and the absolute value is taken to obtain the time matching degree.

[0017] If the time matching degree is within the preset time matching degree range, the corresponding unloading vessel will be marked as the initial screening connecting vessel.

[0018] The vessels selected for initial screening are sorted from smallest to largest according to their corresponding time matching degree to construct an initial screening connection sequence.

[0019] A further aspect of this invention is as follows: the process for initially screening the cargo type matching degree of connecting vessels and obtaining the cargo flow matching degree is as follows:

[0020] If the cargo loaded on the vessel to be loaded is the same type of cargo as the cargo unloaded on the vessel initially screened, it is recorded as cargo of the same type. The proportion of the quantity of cargo of the same type to the total quantity of cargo loaded is calculated to obtain the cargo matching degree.

[0021] If at least one cargo flow direction of the initially screened connecting vessel is consistent with the cargo flow direction of the vessel to be loaded, it is recorded as cargo with the same flow direction. The proportion of the number of cargo with the same flow direction to the total number of loaded cargoes is counted to obtain the cargo flow matching degree.

[0022] As a further aspect of the present invention, the process of screening the sequence of connecting vessels is as follows:

[0023] Input the product category matching degree and the product flow matching degree into the geometric product model, and output the connection matching degree.

[0024] If the connection matching degree is greater than or equal to the connection matching threshold, it is marked as a connecting vessel and sorted in descending order of connection matching degree to obtain the connecting vessel sequence.

[0025] A further aspect of this invention is as follows: based on multiple historical estimated arrival times of connecting vessels within the connecting vessel sequence over a historical period, an arrival stability analysis is performed, and the execution process is as follows:

[0026] The historical period is divided into several historical periods. The estimated arrival time and actual arrival time of connecting ships in each historical period are obtained and combined to obtain the arrival analysis group.

[0027] All arrival analysis groups are first input into the Pearson distance calculation model, and the stable value of arrival time is output.

[0028] As a further aspect of the present invention, the process for obtaining the stably connected vessels is as follows:

[0029] Select the connecting vessel corresponding to the maximum stable arrival time within the connecting vessel sequence as the stable connecting vessel.

[0030] As a further aspect of the present invention, the process of predicting the estimated arrival time is as follows:

[0031] A moving window is set up to input multiple historical estimated arrival times of stably connected vessels in historical cycles and model correction factors into the predicted arrival time model, and the estimated arrival time is output.

[0032] A further aspect of this invention is: setting a moving window, the execution process of which is as follows:

[0033] Multiple historical estimated times are assumed to have multiple moving windows according to the time series before and after. The training sets corresponding to the multiple assumed moving windows are input into the moving window averaging method, and the training estimated time is output.

[0034] The estimated training time and the corresponding estimated validation time are input into the Euclidean computation model, and the set decision value is output.

[0035] Extract the hypothetical moving window corresponding to the minimum window setting value, and use it as the setting moving window.

[0036] As a further aspect of the present invention, the process for obtaining the model correction factor is as follows:

[0037] The difference between the historical estimated time and the historical actual time is calculated, and the ratio of the difference to the historical actual time is used to obtain the single time deviation.

[0038] The model correction factor is obtained by summing and averaging all individual time deviations.

[0039] A further aspect of this invention is as follows: The connection times are evaluated to determine if they match, and vessels with matching connection times are selected. The specific process is as follows:

[0040] If the expected arrival time is later than the departure time, a connection mismatch signal is generated. Stable connecting vessels in the connection vessel sequence are removed, and vessels are reselected according to their connection matching degree from high to low. The reselected vessels are then evaluated for connection matching in the time dimension until a vessel with a matching time dimension is found.

[0041] The beneficial effects of this invention are as follows:

[0042] 1. This invention obtains the berthing and loading / departure times of ships waiting to be unloaded from the port project's ship schedule. It calculates the time matching degree based on the loading / departure time, filters and constructs a sequence of connecting ships according to a preset range, extracts cargo information from the ship within the sequence, compares cargo types to obtain cargo type matching degree, compares cargo flow direction to obtain cargo flow matching degree, calculates the connection matching degree through a geometric product model, and filters the connecting ship sequence accordingly. This solves the problem of how to efficiently filter out connecting ships waiting to be unloaded and loaded, avoiding problems such as low loading / unloading efficiency, resource waste, and transportation disruptions caused by time conflicts, mismatched cargo types, and inconsistent cargo flow directions. This provides ports with a scientific basis for ship scheduling and loading / unloading task arrangement, and improves resource utilization efficiency.

[0043] 2. This invention selects connecting vessels from the connecting vessel sequence, sets a historical period and divides it into time periods, obtains the expected and actual arrival times of the vessels in each time period from the shipping schedule to form an arrival analysis group, inputs all arrival analysis groups into the Pearson distance calculation model to obtain stable values ​​of arrival times, thereby reflecting the relationship between expected and actual arrival times, and selects the vessel corresponding to the maximum stable value as a stable connecting vessel, which helps the port to accurately identify vessels with stable arrival times and avoid difficulties in cargo loading and unloading due to unstable vessel arrival times;

[0044] 3. This invention targets vessels with stable connections. By assuming a moving window, it verifies and analyzes multiple historical estimated arrival times. The optimal moving window is determined using the moving window averaging method and the Euclidean algorithm. A model correction factor is obtained based on historical estimates and actual times to obtain the estimated arrival time. This estimated arrival time is then compared with the departure time to be loaded. If there is a mismatch, the original stable connecting vessel is removed, and vessels are reselected and re-evaluated based on the connection matching degree until a vessel with a time-dimensional connection matching is found from the connecting vessel sequence. By optimizing the moving window and introducing a correction factor, the accuracy of arrival time prediction is improved, avoiding the impact of time mismatch on loading and unloading, effectively improving port vessel scheduling efficiency, and solving the problem of how to accurately predict the arrival time of stable connecting vessels and determine whether it matches the departure time to be loaded. Attached Figure Description

[0045] The invention will now be further described with reference to the accompanying drawings.

[0046] Figure 1 This is a schematic diagram of the port project full-process decision management system based on big data analysis according to the present invention;

[0047] Figure 2 This is a flowchart of the steps of the port project full-process decision management system based on big data analysis of the present invention. Detailed Implementation

[0048] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0049] Example 1

[0050] Please see Figure 1 - Figure 2As shown in the embodiment of the present invention, the port project whole-process decision management system based on big data analysis includes the whole process of wharf layout planning, wharf layout design and construction, and ship berthing and cargo loading / unloading. Since the core purpose of a port is to realize the transportation and transshipment of goods, and ship berthing and cargo loading / unloading are key links in realizing this function, and the value of a port largely depends on its ability to efficiently complete ship loading and unloading operations, which directly affects the port's business volume and economic benefits, ships awaiting unloading can be berthed together with ships awaiting loading / unloading. For example, cargo awaiting unloading can be loaded onto ships awaiting loading / unloading, thereby improving the efficiency of berthing and cargo loading / unloading. The specific execution steps are as follows:

[0051] Vessel matching preliminary screening module: Based on the port project's shipping schedule, obtain the estimated arrival time of each vessel to be unloaded and the departure time to be loaded. Using the departure time to be loaded as a benchmark, analyze the time matching degree between the estimated arrival time to be unloaded and the departure time to be loaded to obtain the preliminary screening vessels and construct the preliminary screening connection sequence.

[0052] It should be noted that the shipping schedule includes the vessel's estimated departure time, estimated arrival time, cargo volume loaded, cargo volume unloaded, type of cargo loaded, and type of cargo unloaded.

[0053] For example, if the shipping schedule is an Excel spreadsheet, the estimated arrival time of the vessel to be unloaded and the departure time of the vessel to be loaded can be filtered using the data filtering function to select the estimated arrival time of the vessel to be unloaded and the departure time of the vessel to be loaded, respectively.

[0054] It should be noted that the criterion for selecting the expected arrival time of the unloading vessel is that the expected arrival time of the unloading vessel should be earlier than the departure time of the loading vessel.

[0055] Based on the loading departure time, the time matching degree is obtained by subtracting the expected arrival time of the vessel to be unloaded from the loading departure time of the vessel to be loaded, and taking the absolute value.

[0056] If the time matching degree is not within the preset time matching degree range, it means that the expected arrival time of the vessel to be unloaded and the departure time of the vessel to be loaded have a low matching degree, and a low matching degree signal is generated.

[0057] If the time matching degree is within the preset time matching degree range, it means that the expected arrival time of the vessel to be unloaded and the departure time of the vessel to be loaded have a high matching degree. A high matching degree signal is generated, and the vessel to be unloaded corresponding to the high matching degree signal is marked as the initial screening connecting vessel.

[0058] It should be noted that the preset time matching range was set by those skilled in the art;

[0059] The time matching degree of the vessels in the initial screening is compared and sorted from smallest to largest to construct the initial screening connection sequence.

[0060] Connection matching and screening module: Obtain the cargo type matching degree and cargo flow matching degree between the initial screening connection vessels and the vessels to be loaded within the initial screening connection sequence, and perform comprehensive analysis to evaluate the connection matching degree between the initial screening connection vessels and the vessels to be loaded, and screen out the connection vessel sequence.

[0061] In some embodiments, one vessel to be connected is arbitrarily selected within the initial screening connection sequence;

[0062] Based on the shipping schedule, the loaded cargo volume, unloaded cargo volume, loaded cargo type, and unloaded cargo type are extracted separately according to the cargo on the vessel.

[0063] The cargo types of the vessels awaiting loading are matched with the cargo types of the vessels initially screened for connection to obtain the cargo matching degree. The process is as follows:

[0064] If the cargo loaded on the vessel awaiting loading is the same type of cargo as the cargo unloaded on the vessel connected to the initial screening, it shall be recorded as the same type of cargo.

[0065] If the cargo loaded on the vessel awaiting loading is not of the same type as the cargo unloaded on the vessel connected to the initial screening, it is recorded as a different type of cargo.

[0066] Count the quantity of similar goods and output the ratio with the total quantity of goods loaded on the ships to be unloaded to obtain the cargo matching degree;

[0067] For example, if the cargo to be unloaded by the initially screened connecting vessel is refrigerated cargo, but the vessel to be loaded is not equipped with refrigeration equipment, then the corresponding cargo loaded on the vessel to be loaded does not contain refrigerated cargo, which means they are different types of cargo.

[0068] Extract the unloading cargo flow direction corresponding to the initially screened connecting vessels, and match the cargo flow with the loading cargo flow direction corresponding to the waiting vessels to obtain the cargo flow matching degree.

[0069] If at least one of the cargo flow directions of the initially screened connecting vessel is consistent with the cargo flow direction of the vessel to be loaded, it is recorded as cargo with the same flow direction;

[0070] If the cargo flow direction of the initially screened connecting vessels is not consistent with the cargo flow direction of the vessels to be loaded, it is recorded as cargo with different flow directions.

[0071] The quantity of goods flowing in the same direction is counted and compared with the total quantity of goods loaded on the ships waiting to be unloaded to obtain the cargo flow matching degree.

[0072] Input the product category matching degree and the product flow matching degree into the geometric product model, and output the connection matching degree.

[0073] As those skilled in the art will understand, the geometric product model serves to reflect the synergistic relationship between cargo type matching degree and cargo flow matching degree. Since cargo type and flow direction are both crucial to the success of vessel connection in actual port operations, and the two are not independent of each other, it helps port managers to more intuitively compare the degree of connection matching between different initial screening vessels and vessels to be loaded, making it easier to make more accurate decisions.

[0074] To further clarify, the meaning of connection matching degree is: it is a quantitative indicator that comprehensively measures the degree of matching between the initially screened connection vessels and the vessels to be loaded in two dimensions: cargo type and cargo flow direction. Specifically, cargo type matching degree reflects the similarity between the cargo type loaded by the vessel to be loaded and the cargo type unloaded by the initially screened connection vessels, that is, the proportion of the same type of cargo to the total amount of cargo unloaded by the vessel to be unloaded. Cargo flow matching degree reflects the consistency between the cargo flow direction unloaded by the initially screened connection vessels and the cargo flow direction loaded by the vessel to be loaded, that is, the proportion of the same flow direction to the total amount of cargo unloaded by the vessel to be unloaded. Specifically, if this value is larger, it indicates that the connection matching degree between the initially screened connection vessels and the vessels to be loaded is higher; if this value is smaller, it indicates that the connection matching degree between the initially screened connection vessels and the vessels to be loaded is lower. This allows for a comprehensive assessment of the feasibility of connection between the initially screened connection vessels and the vessels to be loaded, helping ports or logistics companies to rationally arrange loading and unloading tasks and schedule vessels based on the degree of matching between vessels, so as to make more effective use of resources and improve the utilization efficiency of port equipment.

[0075] The connection matching degree is compared with the connection matching threshold, as follows:

[0076] If the connection matching degree is greater than or equal to the connection matching threshold, it indicates that the connection matching degree between the initially screened connection vessel and the vessel to be loaded is relatively high, and it is marked as a connection vessel.

[0077] If the connection matching degree is less than the connection matching threshold, it means that the connection matching degree between the initially screened connection vessel and the vessel to be loaded is relatively high, and it is marked as a non-connection vessel.

[0078] The connection matching degree of the connecting vessels is compared and sorted from largest to smallest to obtain the connecting vessel sequence;

[0079] The solution in this embodiment can be summarized as follows: Obtain the vessel berthing and loading departure times from the port project's shipping schedule; calculate the time matching degree based on the loading departure time; and initially screen and construct a sequence of vessels according to a preset range.

[0080] The system extracts ship cargo information from the sequence, compares cargo types to obtain cargo type matching degree, compares cargo flow direction to obtain cargo flow matching degree, calculates connection matching degree through geometric product model, and filters connecting ship sequences accordingly. This solves the problem of how to efficiently select connecting ships waiting to be unloaded and loaded, avoiding problems such as low loading and unloading efficiency, resource waste and transportation congestion caused by time conflicts, cargo type mismatch and inconsistent cargo flow direction. This provides ports with a scientific basis for ship scheduling and loading and unloading task arrangement, and improves resource utilization efficiency.

[0081] Example 2

[0082] Please see Figure 1 - Figure 2 The port project full-process decision management system based on big data analysis described in this embodiment of the invention, wherein,

[0083] Connection Stability Analysis Module: Based on multiple historical estimated arrival times of connecting vessels within the connection vessel sequence during historical periods, arrival stability analysis is performed to obtain arrival stability values. Based on these arrival stability values, stable connecting vessels are identified.

[0084] In some embodiments, a connecting vessel is arbitrarily selected from the connecting vessel sequence;

[0085] Define a historical cycle and divide the historical cycle into several historical periods;

[0086] The estimated arrival time and actual arrival time of connecting vessels in each historical period are obtained from the shipping schedule and combined to obtain the arrival analysis group.

[0087] All arrival analysis groups are first input into the Pearson distance calculation model, and the stable value of arrival time is output.

[0088] The specific execution process is as follows:

[0089] A1. The estimated arrival times and actual arrival times of connecting vessels throughout all historical periods are averaged to obtain the average estimated time and the average actual time.

[0090] A2 inputs the estimated arrival time, actual arrival time, average estimated time, and average actual time of all connecting vessels within all historical time periods into the Pearson distance model, and outputs a stable value for arrival time. ;

[0091] The Pearson correlation coefficient model is as follows:

[0092] ;

[0093] To help with understanding, n represents the total number of arrival analysis groups. This is expressed as the expected time average. Expressed as the average over actual time, Represented as the first The estimated arrival time within each arrival analysis group. Represented as the first The actual arrival time within each arrival analysis group;

[0094] It should be noted that the role of the Pearson distance model is as follows: The Pearson distance model is essentially based on the Pearson correlation coefficient, which can be used to measure the degree of linear correlation between two variables. Therefore, it can also be used to analyze the degree of deviation between the expected arrival time and the actual arrival time. Specifically, the expected arrival time is used as one variable and the actual arrival time is used as another variable. If the deviation between the expected arrival time and the actual arrival time is small and stable, it indicates that the arrival time of connecting vessels is relatively stable. Thus, the value output by the Pearson distance model can be used as a basis for predicting future arrival times.

[0095] More specifically, the stability value of arrival time indicates that, from the perspective of deviation, it reflects the magnitude and stability of the deviation between the expected arrival time and the actual arrival time. Specifically, a low stability value indicates that the deviation between the two is large and unstable, and the actual arrival time may deviate significantly from the expected time each time, making it difficult to coordinate cargo loading and unloading with waiting vessels. Conversely, a high stability value indicates that the deviation is small and stable, facilitating cargo loading and unloading coordination with waiting vessels, thereby helping the port to make advance resource allocations based on the expected arrival time.

[0096] From the perspective of linear correlation, the Pearson distance model is based on the Pearson correlation coefficient, which measures the degree of linear correlation between two variables. When calculating the stable value of arrival time, the expected arrival time and the actual arrival time are the two variables. If the linear correlation between the two is strong, it means that the expected arrival time can reflect the actual arrival time well, which means that the expected arrival time and the actual arrival time are very close each time, and the time stability is high. If the correlation between the two is weak, it means that the expected arrival time cannot reflect the actual arrival time, which means that the expected arrival time and the actual arrival time deviate significantly each time.

[0097] The arrival time stability values ​​of all connecting vessels in the connecting vessel sequence are compared, and the connecting vessel corresponding to the largest arrival time stability value is selected as the stable connecting vessel.

[0098] The technical solution of this embodiment is as follows: Select any connecting vessel from the connecting vessel sequence, set a historical period and divide it into time periods, obtain the expected and actual arrival times of the vessel in each time period from the shipping schedule to form an arrival analysis group, input all arrival analysis groups into the Pearson distance calculation model to obtain the stable value of arrival time, and select the vessel corresponding to the maximum stable value as the stable connecting vessel, thereby helping the port to accurately identify vessels with stable arrival times and avoid difficulties in cargo loading and unloading due to unstable vessel arrival times.

[0099] Example 3

[0100] Please see Figure 1 - Figure 2 The port project whole-process decision management system based on big data analysis described in this embodiment of the invention further includes:

[0101] The connection matching assessment module predicts the expected arrival time based on multiple historical estimated arrival times of stably connected vessels in the historical period, obtains the waiting time for loading and departure, and compares it with the predicted expected arrival time to assess whether the connection matches in the time dimension. If they do not match, the connection vessel sequence is re-filtered until a vessel that matches in the time dimension is found.

[0102] In some preferred embodiments, multiple historical estimated arrival times of stably connected vessels in historical cycles and model correction factors are input into the predicted arrival time model, and the predicted arrival time is output.

[0103] The specific execution process is as follows:

[0104] B1. Multiple historical estimated times are shifted into windows according to the preceding and following time series assumptions, and verification analysis is performed based on the shifted windows to complete the setting of the shifted windows;

[0105] It should be noted that the moving window is assumed to be based on the order of historical estimated times in the time dimension. For example, if there are multiple historical estimated times A, B, C, D, and E, and the moving window is assumed to be 2, then A and B are the training set and C is the validation estimated time of the validation set, or B and C are the training set and D is the validation estimated time of the validation set, or C and D are the training set and E is the validation estimated time of the validation set.

[0106] If we assume the moving window is 3, then A, B, and C are the training set, and D is the estimated validation time for the validation set; or B, C, and D are the training set, and E is the estimated validation time for the validation set.

[0107] If we assume the moving window is 4, then A, B, C, and D are the training set, and E is the estimated validation time for the validation set.

[0108] In detail, based on any hypothetical moving window, the training set corresponding to the hypothetical moving window is input into the moving window averaging method, and the output is the estimated training time;

[0109] The estimated training time and the corresponding estimated validation time for each training session are input into the Euclidean algorithm, and the output is the set decision value. ;

[0110] Specifically, the Euclidean computational model: ,in, This represents the total estimated training time. This represents the estimated training time. Verify estimated time;

[0111] It should be noted that the purpose of inputting into the Euclidean computation model is to evaluate different hypothetical moving windows. Specifically, different hypothetical moving window sizes will produce different combinations of training and validation sets, and thus different training estimated times. By comparing the set judgment values ​​under different hypothetical moving windows, the moving window size that makes the prediction results closest to the actual situation can be found, thereby determining the optimal moving window and improving the accuracy and reliability of future time prediction based on historical estimated times.

[0112] It is understandable that the meaning of setting the judgment value is to measure the degree of difference between the training estimated time and the validation estimated time under different hypothetical moving windows, and to reflect the "distance" between the training estimated time obtained based on the moving window averaging method and the corresponding validation estimated time. This helps to select the moving window size that is closest to the actual situation, and then determine an optimal moving window, thereby improving the accuracy and reliability of future time prediction based on historical estimated time.

[0113] Compare the window settings corresponding to different hypothetical moving windows, and extract the hypothetical moving window corresponding to the smallest window setting value as the setting moving window;

[0114] B2, obtain the model correction factor based on multiple historical predicted times and historical actual times;

[0115] In detail, since the Pearson distance model can assess the strong linear correlation between the expected arrival time and the actual arrival time from the perspective of linear correlation, the difference between the historical expected time and the historical actual time is calculated, and the ratio is calculated with the historical actual time to obtain the single time deviation.

[0116] The model correction factor is obtained by summing and averaging all individual time deviations.

[0117] B3, based on the set moving window, inputs multiple historical estimated times and model correction factors into the predicted arrival time model, and outputs the predicted arrival time. ;

[0118] In detail, the model for predicting arrival time is as follows: Where k is the set moving window, and g represents the current position in the order where the moving average is being calculated. Indicates the specific data involved in the calculation. Represented as the model correction factor;

[0119] The process of comparing the estimated arrival time with the departure time is as follows:

[0120] If the expected arrival time in Hong Kong is later than the departure time before loading, a connection mismatch signal will be generated;

[0121] If the expected arrival time is earlier than the departure time, a connection matching signal will be generated;

[0122] If a mismatch signal is generated, the stable connected vessels obtained in the connected vessel sequence are removed, and the vessels are reselected according to the connection matching degree from large to small. The reselected vessels are then evaluated for connection matching in the time dimension until a vessel with a matching connection in the time dimension is selected.

[0123] The corresponding technical solution in this embodiment is as follows: For vessels with stable connections, multiple historical estimated arrival times are verified and analyzed by assuming a moving window. The optimal moving window is determined using the moving window averaging method and the Euclidean algorithm. The model correction factor is obtained based on the historical estimated and actual times to obtain the estimated arrival time, which is then compared with the waiting-to-departure time. If there is a mismatch, the original stable connecting vessel is removed, and vessels are reselected and re-evaluated according to the connection matching degree until a vessel with a time dimension connection matching is found from the connecting vessel sequence. By optimizing the moving window and introducing the correction factor, the accuracy of arrival time prediction is improved, avoiding the impact of time mismatch on loading and unloading, effectively improving the efficiency of port vessel scheduling, and solving the problem of how to accurately predict the arrival time of stable connecting vessels and determine whether it matches the waiting-to-departure time.

[0124] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A port project full-process decision management system based on big data analysis, characterized by: include: Vessel matching preliminary screening module: Evaluate the time matching degree between the expected arrival time of each vessel to be unloaded and the departure time to be loaded, obtain the preliminary screening vessels, and construct the preliminary screening connection sequence; Connection matching and screening module: Based on the cargo type matching degree and cargo flow matching degree of the initially screened connection vessels, comprehensively analyze the connection matching degree between the initially screened connection vessels and the vessels to be loaded, screen out the connection vessels from the initial connection sequence, and construct the connection vessel sequence; Connection stability analysis module: Based on multiple historical estimated arrival times and historical actual arrival times of connecting vessels in historical cycles, a historical arrival time stability analysis is performed to select stable connecting vessels from the connecting vessel sequence; Connection matching assessment module: Based on multiple historical estimated arrival times of stably connected vessels, predict the estimated arrival time and compare it with the waiting departure time to assess whether the connection time matches. If they do not match, re-filter from the connection vessel sequence until a vessel with a matching connection is found. Based on multiple historical estimated arrival times of connecting vessels within the connecting vessel sequence over historical periods, an arrival stability analysis is performed. The analysis process is as follows: The historical period is divided into several historical time periods. The estimated arrival time of connecting ships in each historical time period and the actual historical arrival time are obtained and combined to obtain the arrival analysis group. Input all arrival analysis groups into the Pearson distance calculation model and output the stable value of arrival time. The process of predicting the estimated arrival time is as follows: The estimated arrival times of vessels with stable connections in the historical cycle, along with the obtained model correction factors, are input into the predicted arrival time model after the moving window is set, and the estimated arrival time is output. The process for obtaining the model correction factor is as follows: The difference between the historical estimated time and the historical actual time is calculated, and the ratio of the difference to the historical actual time is used to obtain the single time deviation. The model correction factor is obtained by summing and averaging all individual time deviations. To set up a moving window, the execution process is as follows: Multiple historical estimated times are assumed to have multiple moving windows according to the time series before and after. The training sets corresponding to the multiple assumed moving windows are input into the moving window averaging method, and the training estimated time is output. The estimated training time and the corresponding estimated validation time are input into the Euclidean computation model, and the set decision value is output. Extract the hypothetical moving window corresponding to the minimum window setting value, and use it as the setting moving window.

2. The port project whole-process decision management system based on big data analysis according to claim 1, characterized in that: The initial screening linkage sequence is constructed as follows: The deviation between the estimated arrival time of the unloading vessel and the departure time of the loading vessel is obtained, and the absolute value is taken to obtain the time matching degree. If the time matching degree is within the preset time matching degree range, the corresponding unloading vessel is marked as the initial screening connection vessel, and sorted from smallest to largest according to the corresponding time matching degree to construct the initial screening connection sequence.

3. The port project whole-process decision management system based on big data analysis according to claim 1, characterized in that: The initial screening process for cargo type matching and cargo flow matching of connecting vessels is as follows: If the unloaded cargo of the initially screened connecting vessel and the cargo loaded on the waiting vessel have at least one type of cargo that is the same, they are recorded as the same type of cargo. The proportion of the number of the same type of cargo to the total number of loaded cargoes is calculated to obtain the cargo matching degree. If at least one cargo flow direction of the initially screened connecting vessel is consistent with the cargo flow direction of the vessel to be loaded, it is recorded as cargo with the same flow direction. The proportion of the number of cargo with the same flow direction to the total number of loaded cargoes is counted to obtain the cargo flow matching degree.

4. The port project whole-process decision management system based on big data analysis according to claim 3, characterized in that: The process of constructing a seamless ship sequence is as follows: Input the product category matching degree and the product flow matching degree into the geometric product model, and output the connection matching degree. If the connection matching degree is greater than or equal to the connection matching threshold, it is marked as a connecting vessel and sorted in descending order of connection matching degree to obtain the connecting vessel sequence.

5. The port project whole-process decision management system based on big data analysis according to claim 1, characterized in that: The process for acquiring a stable, connected vessel is as follows: Select the connecting vessel corresponding to the maximum stable arrival time within the connecting vessel sequence as the stable connecting vessel.

6. The port project whole-process decision management system based on big data analysis according to claim 1, characterized in that: The process of assessing whether the connection times match and selecting vessels with matching connections is as follows: If the expected arrival time is later than the departure time, a connection mismatch signal is generated. Stable connecting vessels in the connection vessel sequence are removed, and vessels are reselected according to their connection matching degree from high to low. The reselected vessels are then evaluated for connection matching in the time dimension until a vessel with a matching time dimension is found.

Images