An automated container port multi-agent collaborative optimization operation method

Abstract

The present invention discloses a multi-agent collaborative optimization operation method for an automated container port, which relates to the field of port scheduling operations. This method constructs two models with the shortest port operation time and the least operation cost, and considers port fixed costs, carbon emissions, carbon taxes, etc., making assumptions and constraints on the models; uses the Grey Wolf Algorithm for steps such as individual coding and generating an initial population, and tests the performance of the Grey Wolf Algorithm; finally, calculates with an example to verify the effectiveness of the model and the algorithm. This solution uses a mathematical model to simulate the actual container port collaborative operation method, can analyze the relevant factors affecting the container port operation efficiency, and conducts collaborative optimization around the three main bodies of berths, quay cranes, and AGVs in the operation process to transform the multi-objective problem into a single-objective problem, facilitating the simulation and analysis of actual problems. At the same time, the standard deviation of the Grey Wolf Algorithm is lower than that of the other two algorithms, having higher stability.

Description

Technical Field

[0001] The present invention relates to the field of port scheduling operations, and particularly to an automated container port multi-agent collaborative optimization operation method. Background Art

[0002] Some current related technology research considers the carbon emissions during ship berthing activities, and rarely considers the port operation process. In solving port operation scheduling, the currently commonly used technical means include the first-come-first-served berth allocation method, constructing a function with the shortest in-port time of the ship as the goal, and finally solving it through a heuristic algorithm; constructing an optimization model with the shortest in-port time of the ship as the goal through the ship schedule and the preferred position of the ship; restricting the ship's arrival and operation, considering the time window of the ship's operation, and solving it through a heuristic algorithm; and there is also a technology of constructing a multi-Agent joint scheduling optimization model to reduce operation costs and time.

[0003] Most of the current technical means only study a single link separately. In the actual port operation process, the operations of different links should be considered collaboratively, and a production operation plan should be formulated systematically, so as to reduce the operation costs of container ports and improve operation efficiency. In solving the problem of port operation process optimization, this patent discusses the carbon dioxide emission activities of various port operation equipment from the perspective of port enterprise operation management, including loading and unloading operations and auxiliary production operations, and makes up for the technical gaps in related fields through the comprehensive optimization of the entire operation process. Summary of the Invention

[0004] In order to achieve the above invention purpose, the technical solution adopted by the present invention is as follows:

[0005] An automated container port multi-agent collaborative optimization operation method includes the following steps:

[0006] S1. Construct a collaborative optimization model of three agents: port berths, quay cranes, and AGVs according to set conditions;

[0007] S2. Set multi-objective functions and constraint conditions for the constructed collaborative optimization model with the shortest port operation time and the lowest operation cost as the objective functions;

[0008] S3. Set different scenario parameters, encode the constructed model using the grey wolf algorithm, calculate the fitness of each objective function under different scenario parameters respectively, and perform iterative updates;

[0009] S4. When the maximum number of iterative generations is reached, output the optimal collaborative optimization operation plan.

[0010] Further, the objective function in S2 is expressed as:

[0011]

[0012]

[0013] Among them, is the actual departure time of container ship k; is the scheduled arrival time of container ship k; is the time cost of container ship at the port; is the operation time after quay crane c is selected by container ship k; is the operation cost per unit time of each quay crane; is the operation cost per unit time of a single AGV; is the fixed cost of using quay crane; is the number of quay cranes allocated to container ship k; is the fixed cost of using AGV; is the number of AGVs allocated to container ship k; is the delay operation time for quay crane c selected by container ship k waiting for the r-th AGV; is the quay crane delay operation cost; is the delay operation time for the r-th AGV selected by container ship k waiting for quay crane c; is the single AGV delay operation cost; is the latest departure time of container ship k; is the container ship delay operation cost; is the carbon emission carbon tax cost per unit time of container ship at the port; is the carbon emission coefficient per unit time of container ship k; is the time when container ship k actually arrives at the berth; is the cost of using shore power; is the set of quay cranes at the port berth; is the set of sea-side AGVs in the container terminal; is the set of ships arriving at the port for loading and unloading operations within the time period; is the port operation cost minimum objective function; is the port operation time minimum objective function.

[0014] Furthermore, the specific way to set the objective function in S2 is:

[0015] By the method of linear weighting, the multi-objective problem is transformed into a single-objective programming model, expressed as:

[0016]

[0017] Among them, and are the weight coefficients of each objective function.

[0018] Further, the constraint conditions in S2 include:

[0019] (1) The berthing positions and berthing times of container ships do not overlap;

[0020] (2) When a container ship berths, its ship length needs to be less than the length of the idle shoreline;

[0021] (3) The time when a container ship arrives at the berth is later than the time when it arrives at the port;

[0022] (4) The actual departure time of a container ship is equal to the time when the container ship arrives at the berth plus the operation time at the berth;

[0023] (5) The time when the last quay crane finishes its operation is the operation time of the container ship at the berth, and the container ship is released after the quay crane operation ends;

[0024] (6) Each quay crane can only serve one container ship at the same time;

[0025] (7) The number of quay cranes selected for operation by a container ship cannot be greater than the total number of quay cranes in the port;

[0026] (8) The number of AGVs selected for operation by a container ship cannot be greater than the total number of quay cranes in the port;

[0027] (9) The sum of the operation task amounts assigned to each quay crane is equal to the total task amount of the container ship;

[0028] (10) The number of AGVs in service cannot exceed the total number of equipment in the terminal;

[0029] (11) The number of AGVs operating at any time is equal to the number assigned to the container ship;

[0030] (12) The operation time of quay crane c and the service of container ship k satisfy the time relationship:

[0031]

[0032]

[0033] Wherein, is the operation time of quay crane c; is the quay crane with the smallest operation amount among the quay cranes c selected by container ship k; is the operation efficiency of the quay crane for double - way loading of containers; is the quay crane with the largest operation amount among the quay cranes c selected by container ship k; For the single-trip container loading operation efficiency of the quay crane; The quantity of container tasks to be loaded onto the quay crane c assigned to the container ship k; The quantity of container tasks to be unloaded from the ship onto the quay crane c assigned to the container ship k;

[0034] (13) The delay operation time for the quay crane c selected by the container ship k waiting for the r-th AGV satisfies the time relationship:

[0035] Among them, For the single-trip container loading operation efficiency of the AGV; For the double-trip container loading operation efficiency of the AGV;

[0036] (14) The delay operation time for the r-th AGV selected by the container ship k waiting for the quay crane c satisfies the time relationship:

[0037]

[0038] Furthermore, the specific content in S3 includes:

[0039] S31. Set the basic parameters of the grey wolf algorithm, including the population size, the number of iterations, and the weight coefficient of the objective function;

[0040] S32. Encode the collaborative optimization model and randomly generate an initial population for the container port operation process;

[0041] S33. Calculate the fitness of the randomly generated initial population and select the optimal individual in the current population;

[0042] S34. Use the optimal individual selected in the current population for cyclic iteration to obtain the fitness function curves of the objective functions with the lowest port operation cost and the least port operation time, and obtain the collaborative operation plan through the obtained fitness function curves, including the ship berthing position, the number of quay cranes and AGVs assigned, and the operation time.

[0043] This solution has the following beneficial effects:

[0044] 1. Using a mathematical model to simulate the actual container port collaborative operation method can analyze the relevant factors affecting the container port operation efficiency, and perform collaborative optimization around the three main bodies of the berth, quay crane, and AGV in the operation process to transform the multi-objective problem into a single-objective problem, which is convenient for simulating and analyzing practical problems.

[0045] 2. The optimal value and average value of the Grey Wolf algorithm are significantly lower than those of the Particle Swarm Optimization algorithm and the Genetic algorithm, indicating that the Grey Wolf algorithm has higher accuracy in the multi-agent collaborative optimization of container ports. At the same time, its standard deviation is lower than that of the other two algorithms, showing higher stability. Description of the Drawings

[0046] Figure 1 It is a schematic flow chart of an automated container port multi-agent collaborative optimization operation method of the present invention.

[0047] Figure 2 is a schematic diagram of the operation results of the benchmark test function in the embodiment of the present invention. Among them, (a) is the operation result of Sphere, (b) is the operation result of Schwefel, (c) is the operation result of Rastrigin, (d) is the operation result of Ackley, (e) is the operation result of Kowalik, and (f) is the operation result of hekel.

[0048] Figure 3 It is a schematic diagram of the operation results of different algorithms in the embodiment of the present invention.

[0049] Figure 4 is a fitness function curve graph of the objective function in the embodiment of the present invention. Among them, (a) is the minimum cost and (b) is the shortest time. Detailed Embodiments

[0050] The following describes the detailed embodiments of the present invention to facilitate those skilled in the art of the present technology to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the detailed embodiments. For those of ordinary skill in the art of the present technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions and creations using the concept of the present invention are within the scope of protection.

[0051] An automated container port multi-agent collaborative optimization operation method, as Figure 1 shown, includes the following steps:

[0052] S1. Construct a collaborative optimization model for the three main bodies of port berths, quay cranes, and AGVs according to the set conditions;

[0053] In this embodiment, the collaborative optimization model includes the following contents:

[0054] Berth

[0055] When ships are queuing at the ship anchorage, auxiliary fuel power generation is required to maintain the normal operation of the ships. The carbon emissions generated by the combustion of its fuel for power generation are all considered as the carbon emissions generated by the port side, and the carbon tax for this part is levied on the port side;

[0056] All arriving container ships are equipped with shore power facilities and can use shore power at the berth. In this model, the port does not provide subsidies for the cost of ships using shore power, and the port costs are only the construction cost and maintenance cost of shore power;

[0057] The shoreline is a straight line and consists of continuous berths. When container ships dock at the berths, the safety distance between each container ship needs to be considered, which is known in the ship schedule;

[0058] The impact of special events is not considered, such as the impact of tide and water depth on berthing, equipment maintenance, weather, etc. on operations;

[0059] The arrival time in the ship schedule is regarded as the arrival time of the ship. If the departure time exceeds the ship schedule, a penalty cost will be calculated;

[0060] The moving speed of the ship from the anchorage to the berth is the same, and there is no berth shifting until the loading and unloading operations are completed after arrival at the berth;

[0061] The operation time of the container ship in the port is approximately equal to the loading and unloading operation time of the quay crane and AGV. That is, when the quay crane operation ends, the container ship leaves the berth to make room for the next ship to berth;

[0062] Quay crane

[0063] The number of quay cranes is fixed and can move along the track, and the number of quay cranes is allocated according to the container length and the amount of loading and unloading operations;

[0064] Since the port is a fully automated container port with no manual operation, the moving speed and operation efficiency of the quay crane are constant, not affected by external factors, and the time difference of container operations caused by different positions inside the container is ignored;

[0065] The quay crane can only serve the designated container ship at the same time. Once the container ship's operation is completed and the quay crane is released, it can serve other ships;

[0066] There is a certain safe operation distance between quay cranes;

[0067] The moving time of the quay crane accounts for a relatively small proportion of the overall operation time. Therefore, when allocating quay cranes to container ships, the quay cranes can be allocated across berths without calculating the time cost;

[0068] AGV

[0069] The moving speed of the AGV is constant, and once the corresponding operation area is allocated, it needs to wait until all operations are completed before it can operate across areas;

[0070] The charging time of the AGV is not considered;

[0071] The lateral span of the port yard is relatively large. Here, the average value of the AGV operation efficiency is taken, so the efficiency differences in different yards are not considered;

[0072] The AGV and the quay crane are powered by electricity and are not included in the carbon emissions calculation.

[0073] For the above collaborative optimization model, the following relevant symbols are set for subsequent calculations:

[0074] (1) Container terminal resources

[0075] Operation time, combined with the port operation time, with the unit in minutes here;

[0076] The shoreline length of the port, which is 980m for the case study port;

[0077] The total number of quay cranes at the port berths. There are 16 quay cranes in total at this port;

[0078] The set of quay cranes at the port berths, ;

[0079] The total number of AGVs at the port berths. There are 83 AGVs in total at this port;

[0080] The set of sea - side AGVs in the container terminal, ;

[0081] The total number of container ships expected to arrive at the container port within the time period. 28 container ships are selected in the case study;

[0082] The set of ships arriving at the port for loading and unloading operations within the time period, ;

[0083] G: A maximum value;

[0084] (2) Information related to the ship schedule

[0085] : The length of container ship k (including the safety operation spacing), ;

[0086] The arrival time of container ship k at the port (unit: min), ;

[0087] : The latest departure time of container ship k. If the departure time is later than this time, a delay fee will be calculated (unit: min) ;

[0088] : The number of containers that need to be loaded onto the container ship k, ;

[0089] : The number of containers that need to be unloaded from the container ship k, ;

[0090] The minimum number of quay cranes required for the container ship k to complete the operation, ;

[0091] : The maximum number of quay cranes that the container ship k can use, ;

[0092] : The minimum number of AGVs required for the container ship k to complete the operation, ;

[0093] (3) Equipment operation efficiency and cost:

[0094] The operation efficiency of the quay crane for one-way container loading, that is, during the loading and unloading process, only one side is loading containers, and the same applies to AGVs (unit: TUE / min);

[0095] The operation efficiency of the quay crane for two-way container loading (unit: TUE / min);

[0096] The operation efficiency of the AGV for two-way container loading (unit: TUE / min);

[0097] The operation efficiency of the AGV for one-way container loading (unit: TUE / min);

[0098] : The time cost of the container ship in the port (unit: yuan / min);

[0099] : The carbon tax cost of the container ship's carbon emissions per unit time in the port (unit: yuan / min);

[0100] : The operation cost per unit time of each quay crane (unit: yuan / min);

[0101] : The operation cost per unit time of a single AGV (unit: yuan / min);

[0102] : The delay operation cost of the container ship (unit: yuan / min);

[0103] : Quayside crane delay operation cost (unit: yuan / min);

[0104] : Single AGV delay operation cost (unit: yuan / min);

[0105] Shore power usage cost (construction cost and maintenance cost) (unit: yuan / min);

[0106] Quayside crane fixed cost (unit: yuan);

[0107] AGV fixed cost (unit: yuan);

[0108] : Carbon emission coefficient per unit time of container ship k, ;

[0109] (4) Decision variables

[0110] : The arrival time of container ship k at the berth, which is not the schedule time but the actual arrival time (unit: min), ;

[0111] : The berthing position of container ship k. Here, the center point of the container ship is selected as the berthing point, ;

[0112] : The number of AGVs allocated to container ship k, ;

[0113] : The number of quayside cranes allocated to container ship k, ;

[0114] The quantity of container tasks to be loaded onto the ship by the quayside crane c allocated to container ship k, ;

[0115] : The quantity of container tasks to be unloaded from the ship by the quayside crane c allocated to container ship k, ;

[0116] : 0-1 variable, when ship i berths to the left of ship k, then = 1, otherwise 0, ;

[0117] : 0-1 variable, when ship i berths before ship k, then = 1, otherwise 0, ;

[0118] : A 0-1 variable indicating whether quay crane c is operating on container ship k at time t. If so, it is equal to 1; otherwise, it is 0. ;

[0119] : A 0-1 variable indicating whether the r-th AGV is operating on container ship k at time t. If so, it is equal to 1; otherwise, it is 0. ;

[0120] (5) Dependent variables

[0121] : The actual departure time of container ship k;

[0122] : The operation time of container ship k calculated based on the operation volume and the number of allocated equipment;

[0123] : The operation time of quay crane c after being selected by container ship k;

[0124] : The operation time of the r-th AGV after being selected by container ship k;

[0125] : The time of delayed operation of container ship k due to insufficient allocation of operation equipment;

[0126] : The delayed operation time of quay crane c selected by container ship k waiting for the r-th AGV;

[0127] : The delayed operation time of the r-th AGV selected by container ship k waiting for quay crane c;

[0128] : The quay crane with the largest operation volume among the quay cranes c selected by container ship k;

[0129] : The quay crane with the smallest operation volume among the quay cranes c selected by container ship k.

[0130] S2. Set multi-objective functions and constraint conditions for the constructed collaborative optimization model with the shortest port operation time and the lowest operation cost as the objective functions;

[0131] This model constructs a two-objective optimization model with the lowest port cost and the shortest total time in port for container ships from the perspective of the port side.

[0132] (1) Objective function for the lowest port cost

[0133] With the goal of minimizing the cost of container ships in port, their costs are divided into the following parts: 1. The time cost of container ships in port, 2. The production operation cost of container ports, 3. The overtime operation cost of container ships, 4. Carbon tax and shore power cost.

[0134] The time cost of container ships in port

[0135] The stay of container ships in port mainly consists of the following stages: waiting at the berthing anchorage, arriving at the berth from the anchorage, loading and unloading operations at the berth, and leaving the berth. Since container ships can only use shore power during berthing operations, the time in port is divided into two parts: one is the queuing time, that is, the time from when the container ship arrives at the anchorage to when the ship arrives at the berth and starts operations, and the other is the loading and unloading operation time, that is, the time from when the container ship arrives at the berth and starts operations to when the operations are completed and the ship leaves the port.

[0136] Anchorage queuing time:

[0137]

[0138] (3-1)

[0139] Loading and unloading operation time:

[0140]

[0141] (3-2)

[0142] The total time cost of the ship in port is:

[0143]

[0144] (3-3)

[0145] 2) The production operation cost of container ports

[0146] The production operation cost includes the operation time cost and the fixed cost of equipment use. The operation time cost refers to the operation time of quay cranes and AGVs multiplied by the operation cost per unit time, which are respectively . Equipment use also requires a certain fixed cost, which mainly consists of the number of selected equipment multiplied by the fixed depreciation and maintenance cost of the equipment. Among them, the loading and unloading operation time of quay cranes is determined by the number of containers loaded and unloaded by container k , , the number of quay cranes allocated to container ship k , the number of AGVs and their equipment operation efficiency together.

[0147] In summary, the operation cost of container ports is:

[0148]

[0149] (3 - 4)

[0150] 3) Delayed operation cost of container ships

[0151] There are mainly two types of delays in container ports. One is that container ships cannot berth in time according to the shipping schedule. The other is that when there are too many ships queuing at the anchorage or the loading and unloading volume of ships is too large, the limited resources of the port cannot be satisfied in time, resulting in delays. Here we assume that all container ships arrive at the port at the time specified in the shipping schedule, and only consider the delays of ships due to queuing and loading and unloading operations. Therefore, the delayed operation cost is as follows:

[0152]

[0153] (3 - 5)

[0154] 4) Carbon tax and onshore power equipment cost

[0155] Compared with the operation equipment of traditional container ports, all the operation equipment of this port uses electric energy supply. Without considering the carbon emissions of port operation equipment, only the carbon emissions of container ships are considered. Container ships not only consume fuel for energy supply during navigation, but also consume fuel to supply energy to auxiliary motors when they are berthed. Here, the carbon tax cost is for the period from when the container ship arrives at the anchorage to when it leaves the port. Since it is considered that the container ship can use onshore power when it arrives at the berth, this part of the cost is divided into two parts. The first part is the carbon tax cost of the container ship using fuel to supply energy to the auxiliary motor at the anchorage, and the second part is the onshore power cost when the container ship is berthed. Here, the main cost of the port is the onshore power construction cost and onshore power maintenance cost evenly distributed to each container.

[0156] The carbon tax for container ships berthing at the anchorage is:

[0157]

[0158] (3 - 6)

[0159] The onshore power cost when container ships are berthed:

[0160]

[0161] (3 - 7)

[0162] In summary, the minimum objective function of the above four parts of the cost is:

[0163]

[0164] (3 - 8)

[0165] (2)Objective function for the shortest time

[0166]

[0167] (3 - 9)

[0168] In the formula, represents the time cost of the container ship in the port, represents the operation cost of the container ship at the quay crane and AGV, represents the delay cost of the quay crane waiting for the AGV, represents the delay cost of the AGV waiting for the quay crane, represents the delay cost of the container ship not leaving the port on time. The number of containers added is mainly considered because there is a certain relationship between the delay cost of different ships and the number of containers to be loaded and unloaded.

[0169] (3)Multi - objective model

[0170]

[0171] (3 - 10)

[0172]

[0173] (3 - 11)

[0174] This paper uses the weight value conversion method to solve the multi - objective problem. Through linear weighting, the multi - objective problem is transformed into a single - objective programming model:

[0175]

[0176] Among them and are the weight coefficients of each objective, satisfying . The two objective functions are cost and time respectively. Considering two different operation scenarios in actual operations, namely the lowest operation cost and the shortest operation time, in the scenario of the minimum operation cost ; in the scenario of the shortest operation time .

[0177] Constraint conditions:

[0178] The following three formulas indicate that the berthing position and berthing time of container ships cannot overlap, that is, at the same time and at the same berth, only one container ship can be served:

[0179]

[0180]

[0181] (3 - 14)

[0182]

[0183] (3 - 15)

[0184] The following formula indicates that when a container ship berths, its ship length needs to be less than the length of the idle shoreline:

[0185]

[0186] (3-16)

[0187] The following formula indicates that the arrival time of a container ship at the berth should be later than the arrival time at the port:

[0188]

[0189] (3-17)

[0190] The following formula indicates that the actual departure time of a container ship is equal to the arrival time of the container ship at the berth plus the operation time at the berth:

[0191]

[0192] (3-18)

[0193] The following formula indicates that the time when the last quay crane completes the operation is the operation time of the container ship at the berth, and the container ship is released after the quay crane operation ends:

[0194]

[0195] (3-19)

[0196] The following 3 formulas indicate the relationship between the operation time of quay crane c and the service time of container ship k:

[0197]

[0198] (3-20)

[0199]

[0200] (3-21)

[0201]

[0202] (3-22)

[0203] The following formula indicates that each quay crane can only serve one container ship at the same time:

[0204]

[0205] (3-23)

[0206] The following formula indicates the constraint on the number of quay cranes, that is, the number of quay cranes selected for operation by container ships cannot exceed the total number of quay cranes in the port:

[0207]

[0208] (3-24)

[0209] Similarly, the following formula represents the constraint on the number of AGVs, that is, the number of AGVs selected for the container ship operation cannot be greater than the total number of quay cranes in the port:

[0210]

[0211] (3-25)

[0212] The following formula represents the constraint on the number of quay cranes serving the container ship, that is, the allocated number of quay cranes needs to meet the maximum and minimum quantities:

[0213]

[0214] (3-26)

[0215] The following formula represents the operation volume constraint, that is, the sum of the operation tasks assigned to each quay crane is equal to the total task volume of the container ship:

[0216]

[0217] (3-27)

[0218] The following formula represents the constraint on the number of AGVs, that is, the number of AGVs in service cannot exceed the total number of equipment in the terminal:

[0219]

[0220] (3-28)

[0221] The following formula represents that at any time, the number of operating AGVs is equal to the number assigned to the container ship, and this constraint requires that the AGVs cannot be released until the container operation is completed:

[0222]

[0223] (3-29)

[0224] The following formula represents the quay crane operation time. Similarly, the operating quay cranes cannot operate on other ships before the container ship completes the operation:

[0225]

[0226] (3-30)

[0227] The following formula represents the time for the quay crane to wait for the AGV:

[0228]

[0229] The following formula represents the time for the AGV to wait for the quay crane:

[0230]

[0231] S3. Set different scenario parameters, use the Grey Wolf algorithm to encode the constructed model, calculate the fitness of each objective function under different scenario parameters respectively, and perform iterative updates;

[0232] In this embodiment, in order to facilitate the comparison of the advantages and disadvantages of each algorithm, the population size is set to 50, and the maximum number of iterations is 50. The main parameters set for the genetic algorithm are the crossover probability and the mutation probability. The crossover probability is usually 0.4 - 0.99, and the mutation probability is usually 0.0001 - 0.1. If the crossover probability is too small, it is not conducive to updating the population, so it is set to 0.9 here. If the mutation probability is too small, the diversity of the population will decline too fast, which is not conducive to solving the optimal individual. Here, the mutation probability is selected as 0.1. The main parameters set for the particle swarm algorithm are the inertia weight, the velocity range, and the learning constant. Usually, the inertia coefficient is 0.7 - 0.9, which is set to 0.8 here, the velocity range is 0.5, and the learning constant is set to 1. The Grey Wolf algorithm has a simple structure and the coefficient vector needs to be set to 1. When evaluating the algorithm performance, a benchmark test function is usually used. The benchmark test function is a standard special function used to test the optimization performance of the algorithm. In this paper, two test functions are selected from each of the three types of unimodal test functions, multimodal test functions, and fixed-dimensional multimodal test functions for testing.

[0233] Each benchmark test function was operated 50 times, and the output of each operation of the three types of benchmark test functions under different algorithm operations is shown in Figure 2. The abscissa represents the number of operations, and the ordinate represents the fitness value of each operation. A total of 50 operations were performed. From the operation results, the optimal values and average values of the fitness of the Grey Wolf algorithm and the particle swarm algorithm are closer to the theoretical optimal value compared with the genetic algorithm.

[0234] The solution process of the Grey Wolf algorithm includes the following steps:

[0235] S31. Set the basic parameters of the Grey Wolf algorithm, including the population size, the number of iterations, and the weight coefficient of the objective function. In this embodiment, in order to obtain a better optimization scheme and within an acceptable solution time range, the following values are selected as the basic parameters of the Grey Wolf algorithm: the population size is 100, the maximum number of iterations is 500, and the linear weighted combination method is used for the objective function. The importance of each objective in different scenarios is calculated by multiplying the corresponding weight coefficient. In the scenario of minimizing the operation cost ; in the scenario of minimizing the operation time , thus forming the objective function.

[0236] S32. Encode the collaborative optimization model and randomly generate an initial population for the container port operation process. Specifically, initialize the container port berths, quay cranes, and AGVs. The initial population is randomly generated. The first part is the arrangement order of container ships; the second part is the berthing position of container ships, with the center point of the ship as the berthing position, and its initial berthing position is randomly generated; the third part represents the number of quay cranes, and its value is between the minimum and maximum number of quay cranes that the container ship can carry. Considering that the moving time of quay cranes can be ignored compared to the operation duration, the distribution position and moving time of quay cranes are not considered in the model; the fourth part represents the number of AGVs, and it is required that the sum of the number of all AGVs in the service state is less than the total number of AGVs.

[0237] S33. Calculate the fitness of the randomly generated initial population, select the optimal individual in the current population. By calculating the fitness of the initialized container port operation-related processes, select the optimal individual in the current population, and then through continuous cyclic iteration, generate the optimal individual.

[0238] S34. Use the optimal individual selected in the current population for cyclic iteration to obtain the fitness function curves of the objective functions with the lowest port operation cost and the shortest port operation time, as shown in Figure 4, and obtain the collaborative operation plan through the obtained fitness function curves. Through iteration, obtain the berthing position of the ship, allocate the number of quay cranes and AGVs, operation time and other results.

[0239] S4. When the maximum number of iteration generations is reached, output the optimal collaborative optimization operation plan.

[0240] Solution and effectiveness verification.

[0241] This paper analyzes two situations. One is for the same batch of arriving ships, that is, under the same arrival intensity, different operation strategies are adopted; the other is for the same strategy, under different arrival intensities, the allocation of port operation equipment. To control variables, when comparing the multi-agent collaboration in container ports under different scenarios later, the port needs to select different operation strategies for equipment allocation and control the same arrival intensity to judge the impact of different strategies on the allocation of port operation equipment under this arrival intensity. Here, with λ = 1.16, that is, when there are 28 arriving ships during the operation cycle, it is taken as a case for analysis; when comparing the equipment allocation under different arrival intensities, three groups of data are selected for comparative analysis. To test the performance of the grey wolf algorithm in the multi-agent collaborative optimization of container ports, based on the above parameter settings, the grey wolf algorithm, particle swarm algorithm, and genetic algorithm are respectively used to solve the multi-agent collaborative optimization model mentioned in Chapter 3. To facilitate the comparison of the performance of each algorithm in the multi-agent collaborative optimization of the port, the minimum total operating cost of the container port is taken as the objective for solution. λ = 1.16 is selected, that is, 28 arriving ships during the operation cycle are used for testing. Here, the population size is set to 50. Due to the large amount of calculation, the number of iterations is set to 100 times, and ten experiments are carried out.

[0242] Through the solution, the fitness values of different algorithms for ten operations are obtained as Figure 3 shown. The abscissa represents the number of operations, and the ordinate represents the fitness value output for each operation. To more intuitively compare the advantages and disadvantages between algorithms, the optimal value, average value, and standard deviation of the ten results are selected for auxiliary judgment. The optimal value and average value of the grey wolf algorithm are significantly lower than those of the particle swarm algorithm and genetic algorithm, indicating that the grey wolf algorithm has higher accuracy in the multi-agent collaborative optimization of container ports. At the same time, its standard deviation is lower than the other two algorithms, indicating that the grey wolf algorithm has higher stability.

[0243] After calculation, for 500 cycles, by setting the weights of cost and time in the objective function, the fitness function curve graphs for the two scenarios of minimum cost and shortest time are obtained, as shown in Figure 4.

[0244] Based on the above process, through iteration, the berthing positions of ships, the allocation of quay cranes and AGV quantities, operation time and other results are obtained.

[0245] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices produce a means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or a means for implementing the functions specified in multiple blocks.

[0246] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including an instruction means that implements the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or a means for implementing the functions specified in multiple blocks.

[0247] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or a means for implementing the functions specified in multiple blocks.

[0248] Specific embodiments are applied in the present invention to elaborate on the principles and implementation manners of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

[0249] Those of ordinary skill in the art will realize that the embodiments described herein are for helping readers understand the principles of the present invention, and it should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific deformations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed in the present invention, and these deformations and combinations are still within the protection scope of the present invention.

Claims

1. An automated container port multi-agent collaborative optimization operation method, characterized in that, It includes the following steps: S1. Construct a collaborative optimization model for the three main bodies of port berths, quay cranes, and AGVs according to the set conditions; S2. Set multi-objective functions and constraint conditions for the constructed collaborative optimization model with the shortest port operation time and the lowest operation cost as the objective functions. The multi-objective functions are expressed as: Wherein, is the actual departure time of container ship k; is the scheduled arrival time of container ship k; is the time cost of container ship at the port; is the operation time after quay crane c is selected by container ship k; is the operation cost per unit time of each quay crane; is the operation cost per unit time of a single AGV; is the fixed cost of using quay crane; is the number of quay cranes allocated to container ship k; is the fixed cost of using AGV; is the number of AGVs allocated to container ship k; is the delay operation time for quay crane c selected by container ship k waiting for the r-th AGV; is the quay crane delay operation cost; is the delay operation time for the r-th AGV selected by container ship k waiting for quay crane c; is the delay operation cost of a single AGV; is the latest departure time of container ship k; is the container ship delay operation cost; is the carbon tax cost of carbon emissions per unit time of container ship at the port; is the carbon emission coefficient per unit time of container ship k; is the time when container ship k actually arrives at the berth; is the cost of using shore power; is the set of quay cranes at the port berth; is the set of AGVs on the seaside of the container terminal; is the set of ships arriving at the port for loading and unloading operations during the time period; is the minimum objective function of port operation cost; is the minimum objective function of port operation time; S3. Set different scenario parameters, encode the constructed model using the Grey Wolf algorithm, calculate the fitness of the objective functions under different scenario parameters respectively, and perform iterative updates; S4. When the maximum number of iterative generations is reached, output the optimal collaborative optimization operation plan.

2. The automated container port multi-agent collaborative optimization operation method according to claim 1, wherein The specific method for setting the multi-objective functions in S2 is: By means of linear weighting, transform the multi-objective problem into a single-objective programming model, which is expressed as: wherein, and are the weight coefficients of each objective function, Z is the port operation cost, and T is the port operation time.

3. An automated container port multi-agent collaborative optimization operation method according to claim 1, characterized in that The constraint conditions in S2 include: (1) The berthing positions and berthing times of container ships do not overlap; (2) When a container ship berths, its ship length needs to be less than the length of the idle shoreline; (3) The time when a container ship arrives at the berth should be later than the time when it arrives at the port; (4) The actual departure time of a container ship is equal to the time when the container ship arrives at the berth plus the operation time at the berth; (5) The time when the last quay crane finishes the operation is the operation time of the container ship at the berth, and the quay crane releases the container ship after the operation ends; (6) Each quay crane can only serve one container ship at the same time; (7) The number of quay cranes selected for operation by a container ship cannot be greater than the total number of quay cranes in the port; (8) The number of AGVs selected for operation by a container ship cannot be greater than the total number of quay cranes in the port; (9) The sum of the operation task amounts assigned to each quay crane is equal to the total task amount of the container ship; (10) The number of AGVs for service cannot exceed the total number of equipment in the terminal; (11) The number of AGVs operating at any time is equal to the number assigned to the container ship; (12)The working time of quay crane c and the service satisfaction time of container ship k satisfy the time relationship: Among them, is the working time of quay crane c; is the quay crane with the smallest workload among the quay cranes c selected for container ship k; is the double-trip container loading efficiency of quay crane; is the quay crane with the largest workload among the quay cranes c selected for container ship k; is the single-trip container loading efficiency of quay crane; is the container task volume to be loaded onto the ship by quay crane c assigned to container ship k; is the container task volume to be unloaded from the ship by quay crane c assigned to container ship k; (13) The delay operation time of the quay crane c waiting for the rth AGV selected by the container ship k satisfies the time relationship: Among them, is the operation efficiency of the AGV for single-trip container loading; is the operation efficiency of the AGV for double-trip container loading; (14) The delay operation time of the rth AGV waiting for the quay crane c selected by the container ship k satisfies the time relationship: 。 4. An automated container port multi-agent collaborative optimization operation method according to claim 1, characterized in that What is specifically included in S3 is: S31. Set the basic parameters of the Grey Wolf algorithm, including the population size, the number of iterations, and the weight coefficient of the objective function; S32. Encode the collaborative optimization model and randomly generate an initial population for the container port operation process; S33. Calculate the fitness of the randomly generated initial population and select the optimal individual in the current population; S34. Use the selected optimal individual in the current population for cyclic iteration to obtain the fitness function curves of the objective functions with the lowest port operation cost and the least port operation time, and obtain the collaborative operation plan through the obtained fitness function curves, including the ship berthing position, the number of quay cranes and AGVs assigned, and the operation time.

Images