A berth allocation method and system for bulk cargo wharf in uncertain environment

Abstract

The present application belongs to the technical field of bulk cargo wharf berth allocation, and particularly relates to a bulk cargo wharf berth allocation method and system under uncertain environment. For the berth allocation problem under uncertain environment such as weather conditions and health degree of loading and unloading equipment, an initial scheduling plan is obtained by using the FCFS method considering berth preference based on artificial scheduling experience, the loading and unloading efficiency of the loading and unloading equipment is calculated and predicted by using the deep learning algorithm of LSTM, the scheduling plan is calculated by using the heuristic method of genetic algorithm, the algorithm parameter selection basis is determined according to the fitness function, the influence of the two uncertain factors of weather conditions and equipment health degree on the berth allocation is analyzed, and finally the optimal solution of the berth allocation under the uncertain environment of the bulk cargo wharf is obtained by analyzing multiple instance scales.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of bulk cargo wharf berthing, and particularly relates to a bulk cargo wharf berthing allocation method and system under uncertain environment. BACKGROUND

[0002] In the bulk cargo wharf, the manual features of the berthing allocation process are relatively obvious, and many operation processes and procedures need to be completed by manual operation. Manual operation needs a long time to complete the arrangement and scheduling of tasks, and is easily affected by human factors, resulting in low efficiency, and can only adapt to tasks and situations within a certain range, and is difficult to cope with complex and changeable scenes. At the same time, it is easy to miss or not fully consider, resulting in low accuracy of task scheduling. The application of new technologies such as heuristic algorithms and machine learning algorithms in scheduling can realize intelligent scheduling, which is different from traditional manual scheduling methods, can greatly improve the utilization rate of berths and reduce resource waste, and has important significance for efficient operation of bulk cargo wharfs.

[0003] Intelligent scheduling of wharf refers to breaking through the traditional method of manual scheduling and static scheduling in the process of loading and unloading of goods, realizing dynamic intelligent scheduling management of complex wharf operation process, and is a dynamic scheduling process. In the field of intelligent scheduling of wharf, the intelligent scheduling of container wharf is relatively more, and the core problem of intelligent scheduling of container wharf is truck scheduling and yard optimization. In the field of intelligent scheduling of bulk cargo wharf, the core problem of scheduling is berth allocation and yard allocation. The intelligent scheduling of bulk cargo wharf in China is still in its infancy, and there are few related literatures, and most of the achievements have not been applied in bulk cargo wharf. At present, domestic bulk cargo wharfs are actively exploring automation and intelligence, providing support for realizing intelligent berth allocation. SUMMARY

[0004] The purpose of the present application is to overcome the deficiencies in the prior art, and to provide a bulk cargo wharf berth allocation optimization method and system, which can improve the utilization efficiency of bulk cargo wharf berths and reduce operation costs.

[0005] To achieve the above purpose, the present application adopts the following technical scheme:

[0006] A bulk cargo wharf berth allocation method under uncertain environment, comprising the following steps:

[0007] According to the pre-constructed berth allocation objective function and constraint conditions, a berth allocation optimization mixed integer linear programming mathematical model is established;

[0008] The genetic algorithm is used to solve the berth allocation mixed integer linear programming model, and a long short-term memory (LSTM) deep learning algorithm is used to solve the loading and unloading efficiency of each loading and unloading device under the uncertain environment of weather and device health degree; and a berth allocation optimization scheme is obtained;

[0009] The berth allocation target function is obtained according to a predetermined berth allocation optimization requirement; and the constraint conditions include a conflict constraint of arrival time of a bulk cargo ship, a conflict constraint of berth and loading sequence, a matching constraint of a ship and a berth, and a ship time constraint.

[0010] The LSTM deep learning algorithm is used to solve the loading and unloading efficiency of the bulk cargo ship loading and unloading device.

[0011] Preferably, the berth allocation target function is obtained according to a predetermined berth allocation optimization requirement, including a delay cost of a ship and a loading and unloading cost of the ship, and a sum of the two is a ship operation cost; a corresponding target function is constructed with a target of minimizing the ship operation cost, and a target function expression is as follows:

[0012]

[0013] In formula (1), represents an average service cost of the bulk cargo ship i per unit time; represents an average delay cost of the bulk cargo ship i per unit time; z itm is a binary variable, and if the ship is served by the loading and unloading device m in a period t, the value of the variable is 1, otherwise 0; ldc i represents a loading and unloading completion time of the bulk cargo ship i; ETD i represents an expected departure time of the bulk cargo ship i.

[0014] Preferably, the berth and loading sequence conflict constraint includes a berth position constraint, a constraint that berthing time is later than arrival time, a total number of loading and unloading devices constraint, a number of allocated loading and unloading devices for a single ship constraint, a constraint that a same loading and unloading device only serves one ship in a same time period, a constraint that only one ship is parked at a same position of a berth in a same time, a constraint that relative positions of two ships do not overlap in a same berthing time, and a constraint that berthing times of two ships do not overlap, and corresponding constraint condition expressions are as follows:

[0015] pos i +L i +s≤Len,i∈V (2)

[0016] In formula (2), pos i represents a berthing position of the bulk cargo ship i, i∈V; L idenotes the length of bulk carrier i, i∈V; s denotes the safety distance between bulk carriers in the berth (constant); Len denotes the total length of bulk terminal berth (constant); V denotes the set of bulk carriers to be berthed; equation (2) ensures that the assigned berth position is within the total length of the berth;

[0017] T max ≥ldc i ≥bt i ≥ETA i ,i∈V (3)

[0018] In equation (3), T max denotes the end time of the planning time period; bt i denotes the berthing time of bulk carrier i; ETA i denotes the estimated time of arrival of bulk carrier i; equation (3) ensures that the assigned time of completion of loading and unloading and berthing time are within the planning time period and that the berthing time is later than the arrival time;

[0019] ∑ i ∑ m z itm ≤M max ,i∈C,t∈T,m∈M (4)

[0020] In equation (4), M max denotes the maximum number of handling equipment; M denotes the set of handling equipment that can be assigned to bulk carriers; equation (4) ensures that the number of handling equipment in operation at any time does not exceed the total number of handling equipment;

[0021]

[0022] In equation (5), min denotes the minimum number of handling equipment assigned to carrier i, i and is greater than or equal to 1; denotes the maximum number of handling equipment assigned to carrier i and is less than or equal to the number of cargo holds of bulk carrier i; equation (5) ensures that the number of handling equipment assigned to a carrier is within the minimum and maximum number;

[0023] ∑ i z itm ≤1,i∈V,t∈T,m∈M (6)

[0024] Equation (6) ensures that the same handling equipment serves only one carrier at the same time period;

[0025] pos i +L i +s≤pos j +N(1-y ij ),i,j∈V (7)

[0026] In formula (7), pos i , pos j respectively represent the berthing positions of bulk carriers i and j; N represents a super large constant; y ij represents a binary variable, which takes the value of 1 when the berthing position of the carrier i∈V is on the left side of the carrier j∈V, and takes the value of 0 otherwise; formula (7) ensures that only one ship is parked at the same position on the berth at the same time;

[0027] y ij +y ji ≤1,i,j∈V (8)

[0028] Formula (8) ensures that the relative positions of two ships at the same berthing time do not overlap;

[0029] bt i +ldc i -lds i ≤bt j +N(1-x ij ),i,j∈V (9)

[0030] In formula (9), lds i represents the loading and unloading start time of the bulk carrier i, i∈V; bt j represents the berthing time of the bulk carrier i, i∈V; x ij represents a binary variable, which takes the value of 1 if the berthing time of i∈V is earlier than the berthing time of j∈V, and takes the value of 0 otherwise; formula (9) ensures that the loading and unloading tasks of two bulk carriers at the same berthing position can be normally completed, and N is a super large constant;

[0031] x ij +x ji ≤1,i,j∈V (10)

[0032] Formula (10) ensures that the berthing times of two ships do not overlap;

[0033] y ij +y ji +x ij +x ji ≤1,i,j∈V (11)

[0034] Formula (11) ensures that there is no overlap of ships at the same wharf position at the same time.

[0035] Preferably, the genetic algorithm is used to solve the bulk cargo wharf berth allocation optimization mixed integer linear programming model, comprising:

[0036] Step (1): The berthing position, berthing time and the number of assigned handling equipment of each bulk carrier are taken as decision variables, and the decision variables of all ships are coded to form a two-dimensional array matrix genetic code;

[0037] Step (2): The genetic code is initialized according to the artificial experience FCFS to generate an initial population;

[0038] Step (3): The fitness of the initial population is calculated, and the wheel method is used to select the population to form a paired parent;

[0039] Step (4): The paired parent is subjected to crossover and mutation operations to generate a new population;

[0040] Step (5): The new population is subjected to gene repair according to the constraint conditions;

[0041] Step (6): Return to step (3) and iterate until the error requirement is met to form an optimal solution.

[0042] Preferably, the gene repair of the new population according to the constraint conditions comprises:

[0043] Step (51): Determine whether the gene value of the berthing position of the bulk carrier in the chromosome is within [0, Len-L i ], if yes, execute step (52); otherwise, a new gene value is randomly generated within the range, and step (52) is executed again;

[0044] Step (52): Determine whether the gene value of the berthing time assigned to the bulk carrier in the chromosome is within [0, T max ], if yes, execute step (53); otherwise, a new gene value is randomly generated within the range, and step (53) is executed again;

[0045] Step (54): Determine whether the gene value of the number of handling equipment assigned to the bulk carrier in the chromosome is within , if yes, no gene repair is performed; otherwise, a new gene value is randomly generated within the range to complete the gene repair process.

[0046] Preferably, the following steps are included:

[0047] Step (1): Select the weather influence factors affecting the efficiency of the bulk carrier handling equipment, including wind speed, wind direction, visibility and precipitation; select the equipment influence factors affecting the efficiency of the bulk carrier handling equipment, including equipment type, bulk type and equipment health degree;

[0048] Step (2): Determine the network structure of the input gate, output gate, and forget gate of the LSTM model, and determine the various hyperparameters of the model through grid search;

[0049] Step (3): Obtain the weather factor forecast information in the to-be-assigned time period through the existing weather information service system of the port; obtain the health degree information of each handling equipment in the to-be-assigned time period through the existing equipment full life cycle management system of the port;

[0050] Step (4): Perform data cleaning, data normalization, and other preprocessing operations on the three categories of historical data, measured data, and predicted data of bulk cargo terminal handling operations, environment, and equipment status;

[0051] Step (5): Train the LSTM model according to the historical data of each handling equipment to realize the prediction of ship handling efficiency under different weather factors and different equipment influence factors;

[0052] Step (6): Calculate the corresponding handling time according to the predicted handling efficiency and the amount of cargo handled by the bulk cargo ship, and then calculate the handling cost to evaluate the objective function value of the entire berth allocation scheme, thereby realizing the optimization of berth allocation.

[0053] A bulk cargo terminal berth allocation system, comprising a ship information acquisition module, a weather forecast analysis module, an equipment health monitoring module, a berth allocation module, and a data storage module; wherein:

[0054] Ship information acquisition module: The ship information acquisition module is used to acquire information such as the time of arrival and the planned time of departure of the bulk cargo ship, the length of the ship, the draft, the type of cargo, the amount of cargo, and the planned handling capacity;

[0055] Weather forecast analysis module: The weather forecast analysis module provides the weather conditions of the terminal within the planning period, and analyzes and predicts the handling efficiency of the terminal equipment according to the arrival and departure times of the bulk cargo ship and the current weather information;

[0056] Equipment health monitoring module: The equipment health monitoring module provides the status of the handling equipment on the terminal side within the planning period, and predicts the handling efficiency of the terminal equipment according to the past handling efficiency and the current health degree of the equipment;

[0057] Berth allocation module: The berth allocation module allocates the berth suitable for the bulk cargo ship to be docked according to the acquired ship information;

[0058] Data storage module: The data storage module stores the data of the bulk cargo ship, weather data, equipment efficiency data, and production operation data during the operation, and interacts with the port production scheduling system.

[0059] Compared with the prior art, the application has the following beneficial effects:

[0060] For the berth allocation problem under uncertain environment such as weather conditions and health degree of loading and unloading equipment, an initial scheduling plan is obtained by using the FCFS method considering berth preference based on artificial scheduling experience, the loading and unloading efficiency of the loading and unloading equipment is calculated and predicted by using the LSTM deep learning algorithm, the scheduling plan is calculated by using the heuristic method of the genetic algorithm, the fitness function is used as the basis for evaluating the selection of algorithm parameters, the influence of the two uncertain factors of weather conditions and equipment health degree on the berth allocation is analyzed, and the optimal solution of the berth allocation under the uncertain environment of the bulk cargo terminal is obtained through the analysis of multiple instance scales. In order to be applied in engineering, a berth allocation system is developed to realize the application verification of the method. The work between various functional modules is coordinated through the berth allocation system software, the unified management and scheduling of ship entering and leaving port, scheduling, production and other tasks are realized, the work difficulty and workload of manual berth allocation and scheduling are greatly reduced, the error rate is low compared with manual scheduling, the efficiency is improved, and the increase of operation cost caused by unreasonable manual scheduling is effectively avoided. BRIEF DESCRIPTION OF DRAWINGS

[0061] Figure 1 is a mathematical model of the bulk cargo terminal berth allocation considering uncertain factors provided by the embodiment of the application;

[0062] Figure 2 is a flowchart of a bulk cargo terminal berth allocation optimization method provided by the embodiment of the application;

[0063] Figure 3 is a flowchart of a method for solving a bulk cargo terminal berth allocation optimization mixed integer linear programming model by using a genetic algorithm provided by the embodiment of the application;

[0064] Figure 4 is a flowchart of a genetic repair method when a berth allocation scheme does not meet the constraint condition provided by the embodiment of the application;

[0065] Figure 5 is a comparison chart of the ship loading and unloading efficiency prediction results of the LSTM and RNN neural networks provided by the embodiment of the application;

[0066] Figure 6 is an iterative convergence situation diagram of the genetic algorithm hyperparameter setting provided by the embodiment of the application;

[0067] Figure 7 is a bulk cargo terminal berth allocation optimization result diagram provided by the embodiment of the application. DETAILED DESCRIPTION

[0068] The technical solutions of the present application will be described in detail below with specific examples and drawings. It should be understood that the examples and specific features in the examples are detailed descriptions of the technical solutions of the present application, but not limitations of the technical solutions of the present application, and the specific technical features can be combined with each other.

[0069] A berth allocation method in an uncertain environment of a bulk cargo terminal, comprising the following steps:

[0070] According to the pre-constructed berth allocation objective function and constraint conditions, a berth allocation optimization mixed integer linear programming mathematical model is established;

[0071] The genetic algorithm is used to solve the berth allocation mixed integer linear programming model, and the LSTM is used to solve the loading and unloading efficiency of each loading and unloading device in the uncertain environment of weather and device health degree; and an optimized berth allocation scheme is obtained;

[0072] The berth allocation objective function is obtained according to the pre-determined berth allocation optimization requirements; the constraint conditions include the conflict constraint of the arrival time of the bulk carrier, the conflict constraint of the berth and the loading sequence, the matching constraint of the ship and the berth, and the time constraint of the ship;

[0073] The LSTM deep learning algorithm is used to solve the loading and unloading efficiency of the bulk carrier loading and unloading device.

[0074] The berth allocation objective function is obtained according to the pre-determined berth allocation optimization requirements, including the delay cost of the ship and the loading and unloading cost of the ship, and the sum of the two is the ship operation cost; the corresponding objective function is constructed with the target of minimizing the bulk carrier operation cost, and the objective function expression is as follows:

[0075]

[0076] In formula (1), represents the average service cost of the bulk carrier i per unit time; represents the average delay cost of the bulk carrier i per unit time; z itm Binary variable, if the ship is served by the loading and unloading door machine m in the period t, the value of the variable is 1, otherwise 0; ldc i represents the loading and unloading completion time of the bulk carrier i; ETD i represents the expected departure time of the bulk carrier i.

[0077] The berth and loading sequence conflict constraints include: a berth position constraint, a berthing time later than an arrival time constraint, a total number of loading and unloading devices constraint, a single-ship allocated loading and unloading device number constraint, a same loading and unloading device serving only one ship in a same time period constraint, a same ship berthing at a same position in a same time constraint, a same ship berthing time not overlapping constraint, and a two-ship relative position not overlapping constraint, and corresponding constraint condition expressions are as follows:

[0078] pos i +L i +s≤Len,i∈V (2)

[0079] In formula (2), pos i represents a berthing position of a bulk carrier i, i∈V; L i represents a length of the bulk carrier i, i∈V; s represents a safety distance between bulk carriers in a berth (a constant); Len represents a total length of a bulk terminal berth (a constant); V represents a set of bulk carriers to be berthed; and formula (2) ensures that the allocated berth position is within the total length of the berth.

[0080] T max ≥ldc i ≥bt i ≥ETA i ,i∈V (3)

[0081] In formula (3), T max represents an end time of a planning time period; bt i represents a berthing time of the bulk carrier i, i∈V; ETA i represents an expected arrival time of the bulk carrier i, i∈V. Formula (3) ensures that the allocated loading and unloading completion time and the berthing time are within the planning time period, and ensures that the berthing time of the ship is later than the arrival time.

[0082] ∑ i ∑ m z itm ≤M max ,i∈V,t∈T,m∈M (4)

[0083] In formula (4), M max represents a maximum number of loading and unloading gantries; and M represents a set of loading and unloading gantries that can be allocated to the bulk carrier. Formula (4) ensures that the number of loading and unloading devices working at any time does not exceed the total number of loading and unloading devices.

[0084]

[0085] In formula (5), z represents a minimum number of loading and unloading gantries allocated to the ship i, i∈V, and is greater than or equal to 1; The maximum number of handling equipment assigned to ship i, i∈V, is less than or equal to the number of cargo holds of bulk carrier i. Equation (5) ensures that the number of handling equipment assigned to a ship is within the minimum and maximum number;

[0086] ∑ i z itm ≤1,i∈V,t∈T,m∈M (6)

[0087] Equation (6) ensures that one ship is served by one handling equipment at the same time period;

[0088] pos i +L i +s≤pos j +N(1-y ij ),i,j∈V (7)

[0089] In equation (7), pos i , pos j represent the berthing position of bulk carrier i and j, respectively; N represents a super large constant; y ij is a binary variable, which takes the value of 1 if the berthing position of ship i∈V is on the left side of ship j∈V, otherwise it takes the value of 0. Equation (7) ensures that only one ship is berthing at the same position at the same time;

[0090] y ij +y ji ≤1,i,j∈V (8)

[0091] Equation (8) ensures that the relative position of two ships at the same berthing time does not overlap;

[0092] bt i +ldc i -lds i ≤bt j +N(1-x ij ),i,j∈V (9)

[0093] In equation (9), lds i represents the handling start time of bulk carrier i, i∈V; bt j represents the berthing time of bulk carrier i, i∈V; x ji is a binary variable, which takes the value of 1 if the berthing time of i∈V is earlier than the berthing time of j∈V, otherwise it takes the value of 0. Equation (9) ensures that the handling task of two ships at the same berthing position can be completed normally, and N is a super large constant;

[0094] x ij +x ji ≤1,x,j∈V (10)

[0095] Formula (10) ensures that the berthing time of two ships does not overlap;

[0096] y ij +y ji +x ji +x ji ≤1,i,j∈V (11)

[0097] Formula (11) ensures that there is no overlap of ships at the same time at the same shoreline position.

[0098] The genetic algorithm is used to solve the bulk cargo terminal berth allocation optimization mixed integer linear programming model, comprising:

[0099] Step (1): taking the berthing position, berthing time and the number of allocated handling equipment of each bulk carrier as the decision variable, encoding the decision variable of all ships to form a two-dimensional array matrix genetic code;

[0100] Step (2): initializing the genetic code according to the artificial experience FCFS to generate an initial population;

[0101] Step (3): calculating the fitness of the initial population and selecting the population by roulette method to form paired parents;

[0102] Step (4): performing crossover and mutation operations on the paired parents to generate a new population;

[0103] Step (5): repairing the genes of the new population according to the constraint conditions;

[0104] Step (6): returning to step (3) for iteration until the error requirement is met to form an optimal solution. The gene repair of the new population according to the constraint conditions comprises:

[0105] Step (51): judging whether the gene value of the bulk carrier berthing position in the chromosome is within [0, Len-L i ], if yes, executing step (52); otherwise, randomly generating a new gene value within the range and executing step (52) again;

[0106] Step (52): judging whether the gene value of the berthing time allocated to the bulk carrier in the chromosome is within [0, T max ], if yes, executing step (53); otherwise, randomly generating a new gene value within the range and executing step (53) again;

[0107] Step (54): judging whether the gene value of the number of handling equipment allocated to the bulk carrier in the chromosome is within If not, a new gene value is randomly generated within the range, and the gene repair process is completed.

[0108] The berth allocation optimization of the LSTM model includes the following steps:

[0109] Step (1): Select the meteorological influence factors affecting the efficiency of bulk cargo ship handling equipment, including wind speed, wind direction, visibility and precipitation; select the equipment influence factors affecting the efficiency of bulk cargo ship handling equipment, including equipment type, bulk cargo type and equipment health degree;

[0110] Step (2): Determine the network structure of the input gate, output gate and forget gate of the LSTM model, and determine the various hyperparameters of the model through grid search;

[0111] Step (3): Obtain the weather factor forecast information in the to-be-allocated time period through the existing weather information service system of the port; obtain the health degree information of each handling equipment in the to-be-allocated time period through the existing equipment life cycle management system of the port;

[0112] Step (4): Perform data cleaning, data normalization and other preprocessing operations on the three categories of historical data, measured data and predicted data of bulk cargo terminal handling operation, environment and equipment state;

[0113] Step (5): According to the historical data of each handling equipment, train the LSTM model to realize the prediction of the ship handling efficiency under different meteorological factors and different equipment influence factors;

[0114] Step (6): According to the predicted handling efficiency and the amount of cargo to be handled by the bulk cargo ship, the corresponding handling time is calculated, and then the handling cost is calculated, and the objective function value of the whole berth allocation scheme is evaluated, so as to realize the optimization of berth allocation.

[0115] A bulk cargo terminal berth allocation system of a method, comprising a ship information acquisition module, a weather forecast analysis module, an equipment health monitoring module, a berth allocation module, and a data storage module; wherein:

[0116] The ship information acquisition module: the ship information acquisition module is used to acquire the information of the bulk cargo ship entering the port, the planned departure time, the ship length, the water depth, the cargo type, the cargo capacity, and the planned handling capacity;

[0117] The weather forecast analysis module: the weather forecast analysis module provides the weather conditions of the terminal within the planning period, and analyzes and predicts the handling efficiency of the terminal equipment according to the ship entering and leaving the port time and the current weather information;

[0118] Device health monitoring module: the device health monitoring module provides the situation of the handling equipment on the wharf side within the planning time, and predicts the handling efficiency of the port machine equipment according to the past handling efficiency and the health degree of the current equipment;

[0119] Berth allocation module: the berth allocation module analyzes the berth suitable for the ship to be docked according to the obtained ship information, and performs allocation;

[0120] Data storage module: the data storage module stores the ship data, weather data, equipment efficiency data and production operation data in the operation process, and interacts with the port production scheduling system.

[0121] Figure 2 is a flowchart of a bulk cargo wharf berth allocation optimization method in embodiment one. The flowchart represents the logical sequence of the method described in this embodiment. The method can be executed by a bulk cargo wharf berth allocation system software platform, which can be installed in servers, workstations and other high-performance computer devices. The method of the embodiment specifically includes the following steps:

[0122] Step (1): according to the constructed bulk cargo wharf berth allocation objective function and constraint conditions, a berth allocation optimization mixed integer linear programming model is established;

[0123] Step (2): the LSTM algorithm is used to predict the efficiency of the handling equipment, and the handling efficiency of each bulk cargo handling equipment serving the bulk cargo ship within the working time is obtained;

[0124] Step (3): the genetic algorithm is used to solve the berth allocation optimization mixed integer linear programming model, and the bulk cargo ship berth allocation scheme is obtained;

[0125] The bulk cargo ship berth allocation objective function is constructed according to actual needs; the constraint conditions mainly include bulk cargo ship arrival time conflict constraint, berth and loading sequence conflict constraint, ship and berth matching constraint and ship time constraint, etc.

[0126] It should be understood that the berth allocation optimization requirements and the constraint conditions can be adjusted according to the actual situation of different bulk cargo wharf business processes, business types, user preferences, etc.

[0127] Embodiment two:

[0128] As shown in Figure 4 , according to the uncertain factors such as weather conditions and equipment health degree of the bulk cargo wharf, the efficiency of the handling equipment is predicted based on LSTM, and the specific implementation steps are as follows:

[0129] Step (1): Select the meteorological factors affecting the efficiency of bulk cargo ship handling equipment; mainly including wind speed, wind direction, visibility, precipitation, a total of 4 elements;

[0130] Step (2): Select the equipment factors affecting the efficiency of bulk cargo ship handling equipment; mainly including equipment type, bulk cargo type, equipment health, a total of 3 elements;

[0131] Step (3): Determine the network structure of the input gate, output gate and forget gate of the LSTM model, and determine the model and other hyperparameters through grid search and other methods;

[0132] Step (4): Obtain the meteorological factor forecast information in the to-be-assigned time period through the existing meteorological information service system of the port.

[0133] Step (5): Obtain the health information of each handling equipment in the to-be-assigned time period through the existing equipment life cycle management system of the port.

[0134] Step (6): Perform data cleaning, data normalization and other preprocessing operations on the three categories of historical data, measured data and predicted data of bulk cargo terminal handling operations, environment and equipment state.

[0135] Step (7): According to the historical data of each handling equipment, train the LSTM model to realize the prediction of the ship handling efficiency under different meteorological factors and different equipment factors.

[0136] Step (8): According to the predicted handling efficiency and the amount of cargo handled by the bulk cargo ship, the corresponding handling time is calculated, and then the handling cost is calculated, and the objective function value of the whole berth allocation scheme is evaluated, so as to realize the optimization of berth allocation.

[0137] Example three

[0138] As shown in Figure 3 , according to the genetic rule processing target function and constraint condition, the genetic algorithm is designed to solve the bulk cargo terminal berth allocation optimization mixed integer linear programming model, and the specific implementation steps are as follows:

[0139] Step (1): The berthing position, berthing time and allocated handling equipment quantity of each bulk cargo ship are taken as decision variables, and the decision variables of all ships are coded to form a two-dimensional array matrix genetic code;

[0140] Step (2): Initialize the genetic code according to the FCFS considering the berth preference of artificial experience to generate an initial population;

[0141] Step (3): Calculate the fitness of the initial population, and select the population by roulette method to form paired parents;

[0142] Step (4): Perform crossover and mutation operations on the paired parents to generate a new population;

[0143] Step (5): Perform gene repair on the new population according to the constraints;

[0144] Step (6): Return to step (3) and iterate continuously until the error requirement is met and the optimal solution is formed.

[0145] Example 4:

[0146] This invention provides a bulk cargo terminal berth allocation software system, the system comprising five modules:

[0147] Module 1: Ship Information Acquisition Module, used to acquire information such as the port entry time and planned departure time of bulk carriers, ship length, draft, cargo type, cargo capacity, and planned loading and unloading volume;

[0148] Module 2: Weather forecast and analysis module, which provides the weather conditions of the terminal during the planning period, and analyzes and predicts the loading and unloading efficiency of port machinery equipment based on the arrival and departure times of ships and the current weather information;

[0149] Module 3: Equipment Health Monitoring Module, which provides information on the status of loading and unloading equipment on the quayside within the planned time period, and predicts the loading and unloading efficiency of port machinery equipment based on past loading and unloading efficiency and current equipment health.

[0150] Module 4: Berth allocation module, which analyzes the acquired vessel information, identifies suitable berths for vessels waiting to enter the port, and allocates berths accordingly;

[0151] Module 5: Data storage module, which stores ship data, meteorological data, equipment efficiency data, and production operation data during the operation process, and exchanges information with the port production scheduling system.

[0152] When the software system program is executed by the processor, it implements the steps of the methods described in Embodiments 1 to 3.

[0153] The following uses specific data from a large bulk cargo terminal to verify the method provided in the invention embodiment. This embodiment is coded using Python and analyzes a scenario with nine bulk cargo vessels at consecutive berths, where the berth water depth meets the vessel draft requirements. Vessel information is shown in Table 1, loading and unloading gantry crane information is shown in Table 2, and the 36-hour weather forecast information is shown in Table 3.

[0154] Table 1

[0155]

[0156]

[0157] Table 2

[0158]

[0159] Table 3

[0160]

[0161] The LSTM and RNN neural network ship loading and unloading efficiency prediction result comparison chart provided by the embodiment of the present application is as shown in Figure 5 .

[0162] The population number of the genetic algorithm is set to 50, the iteration number is 200 times, the crossover probability is set to 0.8, the mutation probability is set to 0.01, and the iteration convergence condition diagram of the genetic algorithm is as shown in Figure 6 . Taking a bulk cargo terminal as an example, the berth length is 1200m, the planning time is 120h, and there are 9 ships, and the bulk cargo terminal berth allocation optimization result diagram is as shown in Figure 7 .

[0163] The above is only the preferred embodiment of the present application, and it should be pointed out that for ordinary skilled in the art, without departing from the technical principles of the present application, a number of improvements and modifications can be made, and these improvements and modifications should be considered as the protection scope of the present application.

Claims

1. A berth allocation method for bulk cargo terminals under uncertain environments, characterized in that, Includes the following steps: Based on the pre-constructed objective function and constraints for berth allocation, a mixed-integer linear programming mathematical model for berth allocation optimization is established. A genetic algorithm is used to solve the mixed-integer linear programming model for berth allocation, and a deep learning algorithm using a long short-term memory network (LSTM) is used to solve the loading and unloading efficiency of each loading and unloading equipment under two uncertain environments: weather and equipment health. An optimal berth allocation scheme is obtained. The objective function for berth allocation is constructed and obtained based on the predetermined berth allocation optimization requirements; the constraints include conflicts in the arrival time of bulk carriers, conflicts in the berth and loading sequence, matching constraints between ships and berths, and ship time constraints. The berth allocation objective function is constructed based on predetermined berth allocation optimization requirements, including: vessel delay costs and vessel loading and unloading costs, the sum of which is the vessel operation cost; the objective function is constructed with the goal of minimizing the operation cost of bulk carriers, and its expression is as follows: （1）； In equation (1), Indicates bulk carrier i Average service cost per unit time; ; A binary variable, if the ship is loaded and unloaded by a gantry crane within a period t. m If the service is active, the variable's value is 1; otherwise, it is 0. Indicates bulk carrier i The loading and unloading completion time; Indicates bulk carrier i The estimated departure time, This indicates the assembly of bulk carriers waiting to be moored. This represents the set of loading and unloading gantry cranes that can be assigned to bulk carriers.

2. The berth allocation method for bulk cargo terminals under uncertain environments according to claim 1, characterized in that, The berth and loading sequence conflict constraints include: berth position constraints, berthing time later than arrival time constraints, total number of loading and unloading equipment constraints, number of loading and unloading equipment allocated to a single vessel constraints, the same loading and unloading equipment serving only one vessel at the same time period constraints, only one vessel berthing at the same berth and position at the same time constraints, the relative positions of two vessels not overlapping at the same time constraints, and the berthing times of two vessels not overlapping constraints. The corresponding constraint expressions are as follows: , （2） In equation (2), Indicates bulk carrier i berthing position ; Indicates bulk carrier i Length, ; Indicates the safe distance between bulk carriers within the berth; Indicates the total length of the bulk cargo terminal berths; This represents the collection of bulk carriers waiting to be moored; Equation (2) ensures that the allocated berth locations are within the total length of the berths; （3） Indicates the end time of the planning period; Indicates bulk carrier i berthing time, ; Indicates bulk carrier i The estimated arrival time, Formula (3) ensures that the time for completing loading and unloading and the berthing time are within the planned time period, and ensures that the berthing time of the ship is later than the arrival time. （4） In equation (4), Indicates the maximum number of gantry cranes; This represents the set of loading and unloading gantry cranes that can be allocated to bulk carriers; Equation (4) ensures that the number of loading and unloading equipment in operation at any given time does not exceed the total number of loading and unloading equipment. （5） In equation (5), Assigned to ships i The minimum number of gantry cranes, And greater than or equal to 1; Assigned to ships i The maximum number of gantry cranes, Less than or equal to bulk carriers i The number of cargo holds; Equation (5) ensures that the number of loading and unloading equipment allocated to the ship is within the minimum and maximum quantities; （6） Equation (6) ensures that the same loading and unloading equipment serves only one vessel at a time; （7） In equation (7), , They represent bulk carriers. i and j The berthing position; This represents a very large constant; Represent a binary variable, when the ship The berthing position of the ship When the left side is , the value is 1; otherwise, the value is 0. Equation (7) ensures that only one ship is moored at the same position on the berth at the same time. （8） Equation (8) ensures that the relative positions of the two ships do not overlap at the same berthing time; （9） In equation (9); Indicates the start time of loading and unloading for bulk carrier i. ; Indicates bulk carrier i berthing time, ; Represents a binary variable, if The berthing time was earlier than If the berthing time is 1, then the value is 1; otherwise, the value is 0. Equation (9) ensures that the loading and unloading tasks of two bulk carriers at the same berthing position can be completed normally. N is a very large constant. （10） Equation (10) ensures that the berthing times of the two ships do not overlap; （11） Equation (11) ensures that there will be no overlap of ships at the same location on the same shoreline at the same time.

3. The berth allocation method for bulk cargo terminals under uncertain environments according to claim 2, characterized in that, The method of using a genetic algorithm to solve the mixed-integer linear programming model for optimizing the allocation of bulk cargo terminal berths includes: Step (1): Use the berthing position, berthing time, and number of loading and unloading equipment allocated to each bulk carrier as decision variables, encode the decision variables of all ships, and form a genetic code of a two-dimensional array matrix; Step (2): Initialize the genetic code according to the artificial empirical FCFS to generate an initial population; Step (3): Calculate the fitness of the initial population and use the roulette wheel method to select the population to form paired parents; Step (4): Perform crossover and mutation operations on the paired parents to generate a new population; Step (5): Perform gene repair on the new population according to the constraints; Step (6): Return to step (3) and iterate continuously until the error requirement is met and the optimal solution is formed.

4. The berth allocation method for bulk cargo terminals under uncertain environments according to claim 3, characterized in that, The gene repair of the new population according to the constraints includes: Step (51): Determine whether the gene value for the berthing position of bulk carriers in the chromosome is within [0, Len- If the value is within the specified range, proceed to step (52); otherwise, randomly generate a new gene value within the specified range and then proceed to step (52). Step (52): Based on whether the gene value in the chromosome that assigns berthing time to bulk carriers is within [0, If the value is within the specified range, proceed to step (53); otherwise, randomly generate a new gene value within the specified range and then proceed to step (53). Step (54): Determine whether the gene value for the number of loading and unloading equipment allocated to bulk carriers in the chromosome is within [ , If the value is within the specified range, gene repair will not be performed; otherwise, a new gene value will be randomly generated within that range to complete the gene repair process.

5. The berth allocation method for bulk cargo terminals under uncertain environments according to claim 1, characterized in that, Includes the following steps: Step (1): Select meteorological factors affecting the efficiency of loading and unloading equipment on bulk carriers, including wind speed, wind direction, visibility and precipitation; select equipment factors affecting the efficiency of loading and unloading equipment on bulk carriers, including equipment type, bulk cargo type and equipment health. Step (2): Determine the network structure of the LSTM model, including the input gate, output gate, and forget gate, and determine the hyperparameters of the model through grid search. Step (3): Obtain meteorological forecast information for the time period to be allocated through the port's existing meteorological information service system; obtain health information of each loading and unloading equipment for the time period to be allocated through the port's existing equipment life cycle management system; Step (4): Perform data cleaning and data normalization preprocessing operations on the three categories of historical data, measured data, and predicted data of bulk cargo terminal loading and unloading operations, environment, and equipment status; Step (5): Based on the historical data of each loading and unloading equipment, train the LSTM model to predict the ship loading and unloading efficiency under different meteorological factors and different equipment influencing factors. Step (6): Based on the predicted loading and unloading efficiency and the cargo volume of bulk carriers, calculate the corresponding loading and unloading time, then calculate the loading and unloading operation cost, and evaluate the objective function value of the entire berth allocation scheme, thereby achieving berth allocation optimization.

6. A bulk cargo terminal berth allocation system based on the method of any one of claims 1 to 5, characterized in that, It includes a ship information acquisition module, a weather forecast analysis module, an equipment health monitoring module, a berth allocation module, and a data storage module; among which: Vessel Information Acquisition Module: This module is used to acquire information such as the port entry time and planned departure time of bulk carriers, vessel length, draft, cargo type, cargo capacity, and planned loading / unloading volume. Weather forecast analysis module: The weather forecast analysis module provides the weather conditions of the terminal during the planning period, and analyzes and predicts the loading and unloading efficiency of port machinery equipment based on the arrival and departure times of bulk cargo ships and the current weather information. Equipment Health Monitoring Module: The equipment health monitoring module provides information on the status of loading and unloading equipment on the quayside within the planned timeframe. Based on past loading and unloading efficiency and the current health status of the equipment, it predicts the loading and unloading efficiency of port machinery equipment. Berth allocation module: The berth allocation module analyzes the acquired vessel information to determine suitable berths for bulk carriers waiting to enter the port, and then allocates berths accordingly. Data storage module: The data storage module stores data on bulk cargo vessels, weather data, equipment efficiency data, and production operations during the operation process, and exchanges information with the port production scheduling system.

Images