A port production operation scheduling optimization method considering refrigerated container operation demand

Abstract

A method for optimizing the scheduling of port production operations that takes into account the needs of refrigerated container operations. First, a mathematical model is constructed with the objective function of minimizing the total time to complete all operation tasks; second, an operation time window constraint is set for refrigerated containers; virtual tasks and sequence constraints are established for quay crane operations, a sharing mode is adopted for container truck operations, and allocation and sequence constraints are set; yard crane operations allocate storage locations and set sequence constraints according to the attributes of unloading tasks; the travel time of container trucks is considered to ensure that the travel time between tasks is reasonable; in terms of equipment operation time, the start time of subsequent tasks is set to be later than the completion time of the previous tasks; finally, the optimal configuration of resources and the order of operations are solved through the above constraints and objective functions. This invention can ensure efficient collaboration of each link and minimization of operation time, can provide theoretical support and optimization solutions for the scheduling of port refrigerated container operations, can better ensure the quality of refrigerated container operations in ports, and is crucial to improving the quality and efficiency of port refrigerated container transportation.

Description

Technical Field

[0001] The invention belongs to the field of port production operation scheduling optimization, and relates to a port production operation scheduling optimization method taking into account refrigerated container operation requirements. Background Art

[0002] Ports are the core distribution points for international refrigerated container transportation, and are also key operation nodes with intensive loading and unloading activities of refrigerated containers. Formulating a reasonable port operation scheduling plan is another core task to improve the efficiency of port refrigerated container transportation at the level of transportation optimization. Inside the container port, there are many large-scale operating equipment and complex operating processes. When conducting port operation scheduling in the scenario of mixed operation of refrigerated containers and ordinary containers, we should comprehensively consider the operating tasks of ordinary containers and refrigerated containers in the port on the basis of meeting the unique technical requirements of refrigerated containers and ensuring the quality of refrigerated container operations in the port, coordinate the scheduling of various operating equipment in the port, and improve the overall operating efficiency and quality of the port.

[0003] Therefore, the present invention comprehensively considers the pre-assisted operation requirements of refrigerated container operations in port and the power outage duration limit of refrigerated containers caused by the perishability of products, and extracts the key constraints that form refrigerated container operations in port. Aiming at large container ports with mixed operations of ordinary containers and refrigerated containers, from the perspective of taking into account the operation requirements of refrigerated containers and the overall operation efficiency of container ports, while considering the loading and unloading operations of ordinary containers and refrigerated containers in the port, a container port quay crane, horizontal transportation equipment, and yard crane integrated scheduling optimization method (TW-CTISM) is proposed. Summary of the invention

[0004] In view of the problems existing in the prior art, the present invention provides a port operation system scheduling method, which focuses on the scheduling optimization problem of the internal operation system of a container port with mixed operations of ordinary containers and refrigerated containers, pays attention to the three key operation equipment in the port, namely, quay cranes, container trucks and yard cranes, and considers the operation process of picking up containers for loading and unloading and stacking containers in the port (see Figure 1 ), by introducing power outage restrictions for refrigerated container operations (see Figure 2 ), taking into account the operational requirements of refrigerated containers and ordinary containers (see Figure 3 )-based container port production operation scheduling optimization method (TW-CTISM).

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] A method for optimizing the scheduling of port production operations that takes into account the operational needs of refrigerated containers. The method first constructs a mathematical model with the objective function of minimizing the total time to complete all operational tasks. Secondly, an operation time window constraint is set for refrigerated containers. For quay crane operations, virtual tasks and sequence constraints are established. For container truck operations, a sharing mode is adopted and allocation and sequence constraints are set. For yard crane operations, storage locations are allocated according to the attributes of unloading tasks and sequence constraints are set. Consider the travel time of container trucks to ensure that the travel time between tasks is reasonable. In terms of equipment operation time, the start time of subsequent tasks must be set later than the completion time of previous tasks. Finally, through the above constraints and objective functions, the optimal configuration of resources and the order of operations are solved. Specifically, the following steps are included:

[0007] S1: For a given work shift, a mathematical model with minimizing the total time to complete all work tasks as the objective function.

[0008] Furthermore, the mathematical model constructed in step S1 is:

[0009] Z MAX #(1)

[0010] If the start time of the work shift is set to 0, the container loading task will be scheduled at the same time and unloading and stacking tasks The total time to complete the job Z MAX It can be calculated by formula (2) and (3):

[0011]

[0012] In the formula, Assemble for the unloading task; Assemble for the task of loading boxes on board; The moment of releasing the yard bridge for the port operation task u; The moment to release the quay crane for port operations.

[0013] S2: According to the demand for refrigerated container operations in the port, the operation time window constraints are imposed on the refrigerated container operation tasks in the port. The start time of the container pickup operation at the yard for loading refrigerated containers shall not be earlier than the start time of the refrigerated container operation time window, as expressed in formula (4). The time when the quay crane for loading refrigerated containers completes the operation shall not be later than the end time of the refrigerated container operation time window, as expressed in formula (5). Correspondingly, formulas (6) and (7) indicate that the start time of the unloading container quay crane is not earlier than the start time of the refrigerated container operation time window, and the completion time of the yard crane is not later than the end time of the refrigerated container operation time window. That is, according to the loading and unloading attributes of the refrigerated container operation tasks, the completion time of the pre-auxiliary operation time after the refrigerated container task arrives at the port, and the allowable power outage time of the refrigerated container, the constraint condition formula (4-7) is established:

[0014]

[0015] In the formula, Assemble for the refrigerated container mission; is the start time of the operation window of refrigerated container u, which is determined by the completion time of the preceding auxiliary operation of task u; is the end time of the operation time window of the refrigerated container task u. The duration of the operation time window The allowed power outage duration for refrigerated container tasks. Indicates the time when the field crane starts operation; Indicates the moment when the quay crane is released by the port operation task u; Indicates the time when the port operation task u occupies the quay crane; Indicates the moment when the dock crane is released for the port operation task u.

[0016] S3: Establish constraints based on the characteristics of quay crane operations.

[0017] Let u DS and u DE They are virtual start and end tasks respectively. Gathering for the quay cranes in the port, Represents the set of all refrigerated container and ordinary container tasks that the quay crane n should perform to ensure the orderly progress of the quay crane operation, avoid conflicts and confusion, and improve operational efficiency and safety. Establish constraint formula (8-15): Formula (8-9) jointly constrains the quay crane operation to start and end with a virtual task, ensuring that the operation process has a clear starting point and end point. Formula (10-13) constrains tasks to only operate on specified quay cranes, and there must be immediate preceding and immediate succeeding tasks to ensure that the execution order of tasks on the quay crane is reasonable. Formula (14) restricts the same task from having a sequential relationship to avoid circular dependencies and contradictions. Formula (15) ensures that there is at most one sequential relationship between any two tasks, ensuring the uniqueness and certainty of the operation order.

[0018]

[0019] Where u, u, and u represent a refrigerated container task, and u DS and u DE It is a virtual start task and end task. Decision variables for the operation sequence of port production scheduling: On quay crane n, if task uu is executed before task u, it is 1, otherwise it is 0; 1, indicating that task u is executed before task u. Represents task u DS If task u is executed before, it is 1; Indicates that task u is in task u DE If the task is executed before, it is 1; Formula (15) indicates that there can only be one sequential relationship between tasks u and u.

[0020] S4: Establish constraints based on the characteristics of truck operations.

[0021] In order to improve the flexibility of port scheduling, a large operation surface mode is adopted to share container trucks among all quay cranes in the port. In order to ensure the order and independence of container truck operations, improve the flexibility of scheduling and the overall efficiency of port operations, the constraint conditions (16-24) are established: Formula (16) constrains that tasks can only and must be assigned to one container truck to ensure clear task allocation and avoid duplication and omission. Formula (17) constrains the relationship between resource allocation variables and sequence variables to ensure the rationality of allocation and the continuity of sequence. Formulas (18) to (24) constrain the relationship between the sequence variables of container truck operations. Formulas (18) and (19) show that if a container truck is assigned an operation task, the container truck must have a virtual task start and end to clarify the starting and ending points of the container truck operation. Formulas (20) and (21) constrain that any task must have a preceding and succeeding task to ensure the continuity of task execution. Formula (22) constrains that there is no sequence variable for the same task. Formulas (23) and (24) restrict the existence of only one immediate relationship between any two tasks, thus avoiding task order conflicts and confusion.

[0022]

[0023] Where: Represents a set of tasks; represents a collection of container trucks; If task u is assigned to set card v, it is 1, otherwise it is 0; If v first executes task u and then executes task u, it is 1, otherwise it is 0; otherwise, It means that the card v first executes task u, and then executes task u, which is 1, otherwise it is 0; Indicates that the set card v first executes task u DS , and then execute task u, then it is 1; Indicates that card set v first executes task u, and then executes task u DS Then it is 1; It means that the card set v first executes task u, and then executes task u,, then it is 1; If the same task u has a sequence relationship on the set card v, it is 1, otherwise it is 0.

[0024] S5: Establish constraints based on the characteristics of the field crane operation.

[0025] The storage location of the shipping container has been allocated when it arrives at the port, so its storage location is a known quantity. According to the task attributes, the alternative container storage sub-box areas and corresponding yard cranes can be found. In order to ensure the sequence, rationality and efficiency of the yard crane operation, ensure that the unloading task can be carried out in an orderly manner within the limited storage location, and improve the overall efficiency and safety of port operations, the constraint conditions (25-35) are established: Equations (25) and (26) constrain that the unloading containers must be assigned corresponding locations to ensure that the containers have a clear storage location. Equation (27) is the constraint relationship between the allocation decision and the sequence decision, which ensures the rationality of the storage location allocation and the consistency of the operation sequence. Similar to the container truck constraint set, Equations (28) to (34) jointly constrain the operation task sequence variables to avoid conflicts and confusion in the operation process. In addition, due to the capacity constraint of storage, Equation (35) constrains that the total number of tasks assigned to the yard crane d cannot exceed the storage capacity cap of the corresponding storage block of the yard crane. d .

[0026]

[0027] In the formula, Represents a field bridge set; If task u is stored in the block to which the field bridge d belongs, it is 1, otherwise it is 0; It means that the field bridge d first executes task u and then executes task u, then it is 1, otherwise it is 0; Indicates that field bridge d executes task u first DS , and then execute task u, then it is 1; It means that the field bridge d first executes task u, and then executes task u DE Then it is 1; If the same task u has a sequential relationship on the field bridge d, it is 1, otherwise it is 0; It means that the field bridge d first executes task u, and then executes task u, which is 1, otherwise it is 0; If the bridge d executes task u, first and then executes task u,, it is 1, otherwise it is 0; cap d Indicates the storage capacity of the block to which the yard crane belongs;

[0028] S6: Set constraints related to the truck travel time.

[0029] Empty truck travel means driving to the starting point of the next task after completing one task. Empty travel time is related to the location where the previous task ends and the location where the subsequent task starts. The travel time of the truck is determined by the starting and ending points of the previous task and the subsequent task. Since the model considers both loading and unloading tasks, the starting and ending points vary with the type of task. If the loading container has been allocated a storage location when it is collected at the port, its storage location is a known quantity. For each unloading task According to the task attributes, the alternative container storage sub-box areas and corresponding yard cranes can be found. The paths of container trucks in the port are different when the types of previous and subsequent tasks are different. There are four main types of container truck driving paths. The first is from the quay crane to the storage yard to complete the task of unloading and stacking containers; the second is from the storage yard to the quay crane to complete the task of picking up containers and loading them onto the ship; the third is from the storage yard to the storage yard, that is, after stacking the containers, go to another container area to pick up the containers; the fourth is from the quay crane to the quay crane, that is, after completing the task of loading and delivering containers on a quay crane, go to another quay crane to start the task of unloading and picking up containers. Figure 4 As shown. Figure 4 As shown in the figure, the container truck path within the port is concisely represented when the types of the predecessor task and the subsequent task are different.

[0030] make is the time required for the container truck to travel empty to task u, after completing task u. Then, equations (36-39) respectively represent the travel time of the container truck to travel empty to the subsequent task u, after completing the previous task u. Equation (36) shows that if the previous task is loading and the subsequent task is unloading, the empty travel distance is the distance from the quay crane of the previous task to the quay crane of the previous task. To subsequent tasks Formula (37) shows that if the previous task is a loading task and the subsequent task is also a loading task, the empty travel distance is the distance from the quay crane of the previous task to the quay crane of the previous task. To the yard sub-block where the subsequent tasks are located Formulas (38) and (39) show that when the current task is the unloading task, the allocation decision of idle time and its yard resources is When the subsequent task is a loading task, the idle time is the yard sub-block of the previous task to the yard sub-block of the subsequent task. When the subsequent task is a ship unloading task, the idle time is the time from the yard sub-block of the previous task to the quay crane of the subsequent task.

[0031]

[0032] In the formula, are the positions of the quay cranes for tasks u and u, respectively. is the field bridge position of task u. It represents the time required for the vehicle to travel from the end point of task u to the starting point of task u; From the shore crane arrive the time required; From the shore crane Arrival Bridge the time required; If task u is stored in the block to which field bridge d belongs, it is 1, otherwise it is 0; Indicates from field bridge d to Time required; Indicates from yard crane d to shore crane Time required.

[0033] Formula (40) represents the start time of the next task of the container truck: Must be greater than the time when the set card completes the previous task Plus idle time Where M is a positive integer large enough to be 10 times the length of the planning period, preferably 9. 99 .

[0034]

[0035] make is the travel time required for the container truck to complete task u, then equations (41) and (42) respectively represent the constraints of the heavy-load travel time when the task is a loading task and an unloading task. Equation (41) indicates that the horizontal transportation time of the loading task is the block where the yard bridge is located. To the location of its quay crane Formula (42) shows that the horizontal transportation time of the unloading task and its storage allocation decision Correspondingly, formula (43) gives the starting and ending times of the horizontal transportation of task u and The relationship between the two, that is, the time of transportation completion must be greater than the time of transportation start plus the heavy load transportation time.

[0036]

[0037] In the formula, Indicates the bridge from the field to the required time; If task u is stored in the block to which field bridge d belongs, it is 1, otherwise it is 0; From the shore crane The time required to reach the field bridge d.

[0038] S7: Set equipment operation related constraints.

[0039] Formula (44) and Formula (45) indicate that when the quay crane and yard crane are operating, on the same operating equipment, the start time of the subsequent task must be later than the completion time of its immediate predecessor task plus the buffer time of the equipment between tasks. QCI and t YI are the buffer times required for the quay crane and yard crane to continuously perform the operation tasks. Formula (46) represents the completion time of the task on the quay crane Greater than the time it starts working Add the quay crane operation time t QCO Formula (47) constrains the completion time of the task on the field bridge Greater than the time it starts working Add the bridge operation time t YO .

[0040]

[0041] In the formula, Indicates the time when the port operation task u occupies the quay crane; Indicates the time when the port operation task u occupies the yard bridge; If task u on quay crane n is the predecessor task of task u, it is 1, otherwise it is 0; It means that task u on field bridge d is the predecessor task of task u, and its value is 1.

[0042] When integrating and scheduling multiple operating equipment, attention must be paid to the connection between different equipment. Only when the equipment in the next stage extracts the task container can the operating equipment in the previous stage be released. Formula (48) constrains that the container truck can be released only after it connects to the container. Formula (49) constrains that the container truck can be released only after the quay crane grabs the container. Formula (50) stipulates that the quay crane can be released only after the container truck is connected to the container. Formula (51) stipulates that the quay crane can be released only after the terminal crane grabs the container.

[0043]

[0044] S8: The above optimization problem is a short-term scheduling problem involving multiple devices and bidirectional processes, and has high requirements for solution efficiency. The constructed model is a mixed integer linear programming model, and the solver can solve small-scale problems. However, the actual engineering problem is large in scale and involves many tasks, and it is difficult for the solver to complete the solution within a limited time. Therefore, in order to meet the needs of solving actual scale problems, the present invention designs a heuristic algorithm based on the problem structure and verifies the algorithm performance.

[0045] According to the problem structure, a hybrid heuristic algorithm SA-ALNS based on simulated annealing algorithm (SA) and adaptive large neighborhood search algorithm (ALNS) is designed to solve the problem. The Metropolis criterion of the simulated annealing algorithm is used as the acceptance criterion of ALNS. The algorithm flow is as follows: Figure 5 As shown in the figure. The algorithm starts with the initial solution, and the ALNS algorithm searches in the neighborhood space of the solution. The ALNS algorithm constructs a new solution by adaptively selecting the destruction operator and the repair operator. When the new solution is better than the existing solution, the new solution is accepted and the search is performed in the neighborhood space of the new solution. If the new solution is not better than the current solution, the probability of retaining the inferior solution is calculated according to the Metropolis criterion to increase the search space.

[0046] The initial solution is constructed using heuristic rules. The algorithm assigns operating machinery to refrigerated container tasks and arranges the order of operations according to the order of the start time of the operation time window, and then inserts ordinary container tasks at feasible positions. The execution time of the operation task is based on the principle of starting as early as possible, and the execution time of each task is calculated in sequence.

[0047] Since the optimization problem involves the scheduling of three types of equipment, the algorithm coding adopts a three-layer coding method. Figure 5 A simple example is used to illustrate the algorithm coding method. Take the operation system with 3 quay cranes, 2 container trucks and 4 yard cranes as an example. Figure 6 A feasible solution for scheduling 7 tasks is shown. A three-layer real number encoding is used, and the real number represents the container task number. Each layer of encoding represents a different type of operating equipment. Tuples are used to distinguish different equipment in one layer of encoding. For example, the quay crane encoding layer contains three tuples, which represent the operating plans of the three quay cranes. The tasks in the tuple are arranged in the order of operation. For example, container truck No. 1 will transport task 1, task 4 and task 6 in turn. The three-layer real number encoding with tuples intuitively expresses the resource allocation and operation sequence decision of multi-equipment, multi-task and multi-stage scheduling problems.

[0048] In addition, in order to improve the solution efficiency and avoid deadlock problems during the solution process, the algorithm adds a decoding order table. Figure 6 As shown in the figure, the sequence table records the predecessor and successor tasks of each container task. When a task is deleted due to damage to the solution, the sequence table synchronously extracts and deletes the task. When repairing the damaged solution, the repair operator reads the sequence table, obtains the feasible insertion position of the current task to be inserted, and then selects an insertion position from the feasible insertion positions according to the repair rule of the repair operator to perform the repair. Each time a task is inserted after repair, the sequence table is updated synchronously, and then the next repair and insertion is performed. The dynamic sequence table effectively avoids invalid insertions during repair and improves the efficiency of the algorithm.

[0049] Furthermore, the hypothetical scenarios of this scheduling method include:

[0050] Scenario 1: The cross-berth scheduling of the quay crane is not considered, that is, the quay crane operation area is known;

[0051] Scenario 2: During the work shift, the remaining capacity of the yard is known;

[0052] Scenario 3: The shipping containers have been assembled at the port in advance, and the assembly method and arrival time are determined by the route optimization model;

[0053] Scenario 4: Container trucks in the port follow the established driving rules without considering the path conflicts between container trucks;

[0054] Scenario 5: Only the internal container truck scheduling is considered, and the impact of external container trucks is not considered;

[0055] Scenario 6: The yard crane moves within a fixed container area, and cross-regional scheduling is not considered.

[0056] The beneficial effects of the present invention are:

[0057] This scheduling method comprehensively considers the pre-operational needs of refrigerated containers in port and the power outage duration limit of refrigerated containers due to the perishable nature of products; it is aimed at large container ports with mixed operations of ordinary containers and refrigerated containers. It takes into account the operation needs of refrigerated containers and the overall operation efficiency of container ports, and considers the loading and unloading operations of ordinary containers and refrigerated containers in the port at the same time, ensuring efficient collaboration of all links and minimization of operation time. This method can provide theoretical support and optimization solutions for the scheduling of refrigerated container operations in ports, and can better guarantee the quality of refrigerated container operations in ports, which is crucial to improving the quality and efficiency of refrigerated container transportation in ports. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a sea-side operation system of a container port; (a) is a schematic diagram of the refrigerated container unloading process; (b) is a schematic diagram of the port sea-side loading and unloading system;

[0059] Figure 2 Schematic diagram of the relationship between the temperature inside a refrigerated container, the quality of the goods, and the duration of power outage; (a) is for frozen goods; (b) is for fresh goods;

[0060] Figure 3 Producing an operations scheduling framework for container ports;

[0061] Figure 4 It is the travel path of container truck;

[0062] Figure 5 It is the flow chart of the hybrid heuristic algorithm SA-ALNS;

[0063] Figure 6 This is the SA-ALNS encoding and decoding sequence table. DETAILED DESCRIPTION

[0064] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

[0065] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0066] The specific implementation method of the present invention comprises the following steps:

[0067] S1: A mathematical model that takes minimizing the total time to complete all work tasks as the objective function for a specific work shift.

[0068] Furthermore, the mathematical model constructed in step S1 is:

[0069] Z MAX #(1)

[0070] If the start time of the work shift is set to 0, the container loading task will be scheduled at the same time and unloading and stacking tasks The total time to complete the job Z MAX It can be calculated by formula (2) and (3):

[0071]

[0072] In the formula, Assemble for the unloading task; Assemble for the task of loading boxes on board; The moment of releasing the yard bridge for the port operation task u; The moment to release the quay crane for port operations.

[0073] S2: According to the demand for refrigerated container operations in the port, the operation time window constraints are imposed on the refrigerated container operation tasks in the port. According to the loading and unloading attributes of the refrigerated container operation tasks, the completion time of the pre-assisting operation time after the refrigerated container task arrives at the port, and the allowable power outage time of the refrigerated container, the constraint condition formula (4-7) is established:

[0074]

[0075] In the formula, Assemble for the refrigerated container mission; is the start time of the operation window of refrigerated container u, which is determined by the completion time of the preceding auxiliary operation of task u; is the end time of the operation time window of the refrigerated container task u. The duration of the operation time window The allowed power outage duration for refrigerated container tasks. Indicates the time when the field crane starts operation; Indicates the moment when the quay crane is released by the port operation task u; Indicates the time when the port operation task u occupies the quay crane; Indicates the moment when the dock crane is released for the port operation task u.

[0076] The start time of the container pickup operation at the yard for loading refrigerated containers shall not be earlier than the start time of the refrigerated container operation time window, as expressed in formula (4). The time when the quay crane for loading refrigerated containers completes the operation shall not be later than the end time of the refrigerated container operation time window, as expressed in formula (5). Correspondingly, formulas (6) and (7) indicate that the start time of the unloading container quay crane operation shall not be earlier than the start time of the refrigerated container operation time window, and the time when the yard crane completes the operation shall not be later than the end time of the refrigerated container operation time window. S3: Establish constraints based on the characteristics of the quay crane operation.

[0077] Let u DS and u DE They are virtual start and end tasks respectively. Gathering for the quay cranes in the port, Represents the set of all refrigerated container and ordinary container tasks that the quay crane n should perform. Establish constraint formula (8-15): Formula (8-9) jointly constrains the quay crane operation to start and end with a virtual task. Formula (10-13) constrains the task to only operate on the specified quay crane, and there must be a preceding and succeeding task. Formula (14) restricts the same task from having a sequential relationship. Formula (15) ensures that there is at most one sequential relationship between any two tasks.

[0078]

[0079] Where u, u, and u represent a refrigerated container task, and u DS and u DE It is a virtual start task and end task. Decision variables for the operation sequence of port production scheduling: On quay crane n, if task u is executed before task u, it is 1, otherwise it is 0; 1, indicating that task u is executed before task u. Represents task u DS If task u is executed before, it is 1; Indicates that task u is in task u DE If the task is executed before, it is 1; Formula (15) indicates that there can only be one sequential relationship between tasks u and u.

[0080] S4: Establish constraints based on the characteristics of truck operations.

[0081] In order to improve the flexibility of port scheduling, a large operating surface mode is adopted to share container trucks among all quay cranes in the port. Constraints (16-24) are established: Formula (16) constrains that tasks can and must be assigned to only one container truck. Formula (17) constrains the relationship between resource allocation variables and sequence variables. Formulas (18) to (24) constrain the relationship between container truck operation sequence variables. Formulas (18) and (19) show that if a container truck is assigned an operation task, the container truck must have a virtual task start and end. Formulas (20) and (21) constrain that any task must have a preceding and succeeding task. Formula (22) constrains that there is no sequence variable for the same task. Formulas (23) and (24) restrict that there is only one close-before-close-after relationship between any two tasks.

[0082]

[0083] Where: Represents a set of tasks; represents a collection of container trucks; If task u is assigned to set card v, it is 1, otherwise it is 0; If v first executes task u and then executes task u, it is 1, otherwise it is 0; otherwise, It means that the card v first executes task u, and then executes task u, which is 1, otherwise it is 0; Indicates that the set card v first executes task u DS , and then execute task u, then it is 1; Indicates that card set v first executes task u, and then executes task u DS Then it is 1; It means that the card set v first executes task u, and then executes task u,, then it is 1; If the same task u has a sequence relationship on the set card v, it is 1, otherwise it is 0.

[0084] S5: Establish constraints based on the characteristics of the field crane operation.

[0085] The storage location of the shipping container has been allocated when it arrives at the port, so its storage location is a known quantity. According to the task attributes, the alternative container storage sub-box areas and corresponding yard cranes can be found. Establish constraint conditions (25-35): Equations (25) and (26) constrain that the corresponding positions must be assigned to the unloading containers. Equation (27) is the constraint relationship between the allocation decision and the sequence decision. Similar to the constraint set of the container truck, equations (28) to (34) jointly constrain the operation task sequence variables. In addition, due to the capacity constraint of the storage, equation (35) constrains that the total number of tasks assigned to the yard bridge d cannot exceed the storage capacity cap of the corresponding storage block of the yard bridge d .

[0086]

[0087] In the formula, Represents a field bridge set; If task u is stored in the block to which the field bridge d belongs, it is 1, otherwise it is 0; It means that the field bridge d first executes task u and then executes task u, then it is 1, otherwise it is 0; Indicates that field bridge d executes task u first DS , and then execute task u, then it is 1; It means that the field bridge d first executes task u, and then executes task u DE Then it is 1; If the same task u has a sequential relationship on the field bridge d, it is 1, otherwise it is 0; It means that the field bridge d first executes task u, and then executes task u, which is 1, otherwise it is 0; If the bridge d executes task u, first and then executes task u,, it is 1, otherwise it is 0; cap d Indicates the storage capacity of the block to which the yard crane belongs;

[0088] S6: Set constraints related to the truck travel time.

[0089] Empty truck travel means driving to the starting point of the next task after completing one task. Empty travel time is related to the location where the previous task ends and the location where the subsequent task starts. The travel time of the truck is determined by the starting and ending points of the previous task and the subsequent task. Since the model considers both loading and unloading tasks, the starting and ending points vary with the type of task. If the loading container has been allocated a storage location when it is collected at the port, its storage location is a known quantity. For each unloading task According to the task attributes, the alternative container storage sub-box areas and corresponding yard cranes can be found. Figure 4 As shown in the figure, the container truck path within the port is concisely represented when the types of the predecessor task and the subsequent task are different.

[0090] make is the time required for the container truck to travel empty to task u, after completing task u. Then, equations (36-39) respectively represent the travel time of the container truck to travel empty to the subsequent task u, after completing the previous task u. Equation (36) shows that if the previous task is loading and the subsequent task is unloading, the empty travel distance is the distance from the quay crane of the previous task to the quay crane of the previous task. To subsequent tasks Formula (37) shows that if the previous task is a loading task and the subsequent task is also a loading task, the empty travel distance is the distance from the quay crane of the previous task to the quay crane of the previous task. To the yard sub-block where the subsequent tasks are located Formulas (38) and (39) show that when the current task is the unloading task, the allocation decision of idle time and its yard resources is When the subsequent task is a loading task, the idle time is the yard sub-block of the previous task to the yard sub-block of the subsequent task. When the subsequent task is a ship unloading task, the idle time is the time from the yard sub-block of the previous task to the quay crane of the subsequent task.

[0091]

[0092] In the formula, are the positions of the quay cranes for tasks u and u, respectively. is the field bridge position of task u. It represents the time required for the vehicle to travel from the end point of task u to the starting point of task u; From the shore crane arrive the time required; From the shore crane Arrival Bridge the time required; If task u is stored in the block to which field bridge d belongs, it is 1, otherwise it is 0; Indicates from field bridge d to Time required; Indicates from yard crane d to shore crane Time required.

[0093] Formula (40) represents the start time of the next task of the container truck: Must be greater than the time when the set card completes the previous task Plus idle time Where M is 9 99 .

[0094]

[0095] make is the travel time required for the container truck to complete task u, then equations (41) and (42) respectively represent the constraints of the heavy-load travel time when the task is a loading task and an unloading task. Equation (41) indicates that the horizontal transportation time of the loading task is the block where the yard bridge is located. To the location of its quay crane Formula (42) shows that the horizontal transportation time of the unloading task and its storage allocation decision Correspondingly, formula (43) gives the starting and ending times of the horizontal transportation of task u and The relationship between the two, that is, the time of transportation completion must be greater than the time of transportation start plus the heavy load transportation time.

[0096]

[0097] In the formula, Indicates the bridge from the field to the required time; If task u is stored in the block to which field bridge d belongs, it is 1, otherwise it is 0; From the shore crane The time required to reach the field bridge d.

[0098] S7: Set equipment operation related constraints.

[0099] Formula (44) and Formula (45) indicate that when the quay crane and yard crane are operating, on the same operating equipment, the start time of the subsequent task must be later than the completion time of its immediate predecessor task plus the buffer time of the equipment between tasks. QCI and t YI are the buffer times required for the quay crane and yard crane to continuously perform the operation tasks. Formula (46) represents the completion time of the task on the quay crane Greater than the time it starts working Add the quay crane operation time t QCO Formula (47) constrains the completion time of the task on the field bridge Greater than the time it starts working Add the bridge operation time t YO .

[0100]

[0101] In the formula, Indicates the time when the port operation task u occupies the quay crane; Indicates the time when the port operation task u occupies the yard bridge; If task u on quay crane n is the predecessor task of task u, it is 1, otherwise it is 0; It means that task u on field bridge d is the predecessor task of task u, and its value is 1.

[0102] When integrating and scheduling multiple operating equipment, attention must be paid to the connection between different equipment. Only when the equipment in the next stage extracts the task container can the operating equipment in the previous stage be released. Formula (48) constrains that the container truck can be released only after it connects to the container. Formula (49) constrains that the container truck can be released only after the quay crane grabs the container. Formula (50) stipulates that the quay crane can be released only after the container truck is connected to the container. Formula (51) stipulates that the quay crane can be released only after the terminal crane grabs the container.

[0103]

[0104] S8: According to the problem structure, a hybrid heuristic algorithm SA-ALNS based on simulated annealing algorithm (SA) and adaptive large neighborhood search algorithm (ALNS) is designed to solve the problem. The Metropolis criterion of the simulated annealing algorithm is used as the acceptance criterion of ALNS. The algorithm flow is as follows: Figure 5 As shown in the figure. The algorithm starts with the initial solution, and the ALNS algorithm searches in the neighborhood space of the solution. The ALNS algorithm constructs a new solution by adaptively selecting the destruction operator and the repair operator. When the new solution is better than the existing solution, the new solution is accepted and the search is performed in the neighborhood space of the new solution. If the new solution is not better than the current solution, the probability of retaining the inferior solution is calculated according to the Metropolis criterion to increase the search space.

[0105] The initial solution is constructed using heuristic rules. The algorithm assigns operating machinery to refrigerated container tasks and arranges the order of operations according to the order of the start time of the operation time window, and then inserts ordinary container tasks at feasible positions. The execution time of the operation task is based on the principle of starting as early as possible, and the execution time of each task is calculated in sequence.

[0106] Since the optimization problem involves the scheduling of three types of equipment, the algorithm coding adopts a three-layer coding method. Figure 5 A simple example is used to illustrate the algorithm coding method. Take the operation system with 3 quay cranes, 2 container trucks and 4 yard cranes as an example. Figure 6 A feasible solution for scheduling 7 tasks is shown. A three-layer real number encoding is used, and the real number represents the container task number. Each layer of encoding represents a different type of operating equipment. Tuples are used to distinguish different equipment in one layer of encoding. For example, the quay crane encoding layer contains three tuples, which represent the operating plans of the three quay cranes. The tasks in the tuple are arranged in the order of operation. For example, container truck No. 1 will transport task 1, task 4 and task 6 in turn. The three-layer real number encoding with tuples intuitively expresses the resource allocation and operation sequence decision of multi-equipment, multi-task and multi-stage scheduling problems.

[0107] In addition, in order to improve the solution efficiency and avoid deadlock problems during the solution process, the algorithm adds a decoding order table. Figure 6 As shown in the figure, the sequence table records the predecessor and successor tasks of each container task. When a task is deleted due to damage to the solution, the sequence table synchronously extracts and deletes the task. When repairing the damaged solution, the repair operator reads the sequence table, obtains the feasible insertion position of the current task to be inserted, and then selects an insertion position from the feasible insertion positions according to the repair rule of the repair operator to perform the repair. Each time a task is inserted after repair, the sequence table is updated synchronously, and then the next repair and insertion is performed. The dynamic sequence table effectively avoids invalid insertions during repair and improves the efficiency of the algorithm.

[0108] At the same time, two destruction operators are designed according to the characteristics of the optimization problem: random destruction operator and worst-case destruction operator. The random destruction operator randomly selects tasks from all job tasks for deletion. The worst-case destruction operator gives priority to refrigerated container tasks with too long power outages, and then selects job tasks with longer total operation time for destruction. Accordingly, three repair operators are designed: random insertion repair, greedy insertion repair, and earliest insertion repair. Random insertion repair randomly selects one from the feasible positions to perform insertion repair for the current task to be inserted. Greedy insertion repair selects the position with the shortest total operation time after insertion from the feasible positions for insertion. The earliest insertion repair inserts the current task to be inserted into the position where the operation can start the earliest.

[0109] The constructed scheduling optimization model involves the allocation, operation sequence and execution time of three key operating equipment. Based on a small-scale example, the SA-ALNS algorithm is applied to solve examples of different scales under the limitation of 1800s solution time. In the example, refrigerated container tasks account for 10% of the total tasks, and only one refrigerated container area is set in the yard. The specific example parameters and solution results are shown in Table 1.

[0110] Table 1 SA-ALNS algorithm verification examples and results

[0111]

[0112] In the example parameters, “|·|” represents the cardinality of the set; Obj.Gap = (Obj.SA-ALNS-Obj.CPLEX) / Obj.CPLEX; Obj.SA-ALNS is the objective function value solved by SA-ALNS, and Obj.CPLEX is the optimal function value solved by CPLEX.

[0113] The above-described embodiments merely express the implementation methods of the present invention, but they cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A method for optimizing port production operation scheduling considering refrigerated container operation requirements, characterized in that: The following steps are involved: S1: Constructing mathematical model: minZ MAX (1) Among them, Z MAX The total time to complete the assignment; S2: Establish constraint conditions: In the formula, Assemble for the refrigerated container mission; is the start time of the task u's operation time window, which is determined by the completion time of the auxiliary operation preceding task u; is the end time of the operation time window of task u; the duration of the operation time window The allowed power outage duration for task u; The time when the yard crane starts operation; The moment to release the quay crane for in-port tasks; The time when the quay crane is occupied for task u; The moment when the field bridge is released for task u; S3: Establish constraints based on the characteristics of quay crane operations; Let u DS and u DE They are virtual start and end tasks respectively. Gathering for the quay cranes in the port, is the set of all refrigerated container and ordinary container tasks that the quay crane n should perform; the constraint condition formula is established: Where u,u' represents a refrigerated container task, u DS and u DE For virtual start and end tasks; Decision variables for the operation sequence of port production scheduling: For quay crane n, if task u is executed before task u', it is 1, otherwise it is 0; is 1, which means task u' is executed before task u. For task u DS If it is executed before task u', it is 1; For task u in task u DE If it is executed before, it is 1; Formula (15) means that there can only be one sequential relationship between tasks u and u'; S4: Establish constraints based on the characteristics of truck operations: Where: u is the task set; v is the container truck set; If task u is assigned to set v, it is 1, otherwise it is 0; If the card v executes task u first and then task u', it is 1, otherwise it is 0; If the card v executes task u' first and then task u, it is 1, otherwise it is 0; To collect card v, first perform task u DS , and then execute task u', it is 1; For card v, first execute task u', then execute task u DE Then it is 1; If the card v first executes task u' and then executes task u'', then it is 1; If the same task u has a sequential relationship on the set card v, it is 1, otherwise it is 0; S5: Establish constraints based on the characteristics of the field crane operation; For each unloading task According to its task attributes, the alternative container storage sub-box area and the corresponding yard crane are known Create constraint conditions: In the formula, For field bridge assembly; If task u is stored in the block to which bridge d belongs, it is 1, otherwise it is 0; If the field bridge d executes task u first and then task u', it is 1, otherwise it is 0; Execute task u for field bridge d first DS , and then execute task u', it is 1; For field bridge d, first execute task u, then execute task u DE Then it is 1; If the same task u has a sequential relationship on the field bridge d, it is 1, otherwise it is 0; If the field bridge d first executes task u' and then executes task u, it is 1, otherwise it is 0; If the field bridge d first executes task u' and then executes task u'', it is 1, otherwise it is 0; cap d The storage capacity of the block to which the yard crane belongs; S6: Setting constraints related to the truck travel time; For each unloading task According to its task attributes, the alternative container storage sub-box area and the corresponding yard crane are known There are four main types of container truck driving paths; make is the time required for the container truck to travel empty to task u' after completing task u, then equations (36-39) respectively represent the travel time of the container truck to travel empty to the subsequent task u' after completing the previous task u; equation (36) shows that if the previous task is loading and the subsequent task is unloading, the empty travel distance is To subsequent tasks Formula (37) shows that the empty distance is the quay crane from the previous task To the yard sub-block where the subsequent tasks are located Equations (38) and (39) show that the allocation decision of idle time and its yard resources The idle time is the time from the yard sub-block of the previous task to the yard sub-block of the subsequent task. The idle time is from the yard sub-block of the previous task to the quay crane of the subsequent task. In the formula, are the quay crane positions for tasks u and u' respectively, is the field bridge position of task u'; It is expressed as the time required for the vehicle to travel from the end point of task u to the starting point of task u'; For the quay crane arrive the time required; For the quay crane Arrival Bridge the time required; If task u is stored in the block to which bridge d belongs, it is 1, otherwise it is 0; From field bridge d to Time required; From the yard crane d to the shore crane Time required; Formula (40) is the starting time for the truck to execute the next task Greater than the time it takes to complete the previous task Plus idle time M is a positive integer; make is the travel time required for the container truck to complete task u, then equations (41) and (42) respectively represent the constraints of the heavy-load travel time when the task is loading and unloading tasks; equation (41) indicates that the horizontal transportation time of the loading task is the block where the yard bridge is located To the location of its quay crane Formula (42) represents the horizontal transportation time of the unloading task and its storage allocation decision Related; Formula (43) gives the starting and ending times of horizontal transportation of task u and The relationship between the transport completion time and the transport start time plus the transport time; In the formula, Indicates the bridge from the field To the shore crane the time required; If task u is stored in the block to which field bridge d belongs, it is 1, otherwise it is 0; From the shore crane The time required to reach the field bridge d; S7: Set equipment operation related constraints; Formula (44) and Formula (45) indicate that when the quay crane and yard crane are operating, on the same operating equipment, the start time of the subsequent task is later than the completion time of its immediate predecessor task plus the buffer time of the equipment between tasks; t QCI and t YI are the buffer time required for the quay crane and yard crane to continuously perform the operation task; Formula (46) represents the completion time of the task on the quay crane Greater than the time it starts working Add the quay crane operation time t QCO ; Formula (47) constrains the completion time of the task on the field bridge Greater than the time it starts working Add the bridge operation time t YO ; In the formula, Indicates the time when the port operation task u' occupies the quay crane; Indicates the time when the port operation task u' occupies the yard bridge; If task u on quay crane n is the predecessor task of task u', it is 1, otherwise it is 0; It is 1 if task u on field bridge d is the predecessor task of task u'; For container loading tasks Each constraint in equation (48) to equation (51) represents a release condition; S8: According to the problem structure, a hybrid heuristic algorithm SA-ALNS based on the simulated annealing algorithm SA and the adaptive large neighborhood search algorithm ALNS is designed to solve the problem.

2. A method for optimizing port production operation scheduling considering refrigerated container operation demand according to claim 1, characterized in that: In S1, if the start time of the work shift is set to 0, the container loading task is scheduled at the same time. and unloading and stacking tasks The total time to complete the task Z MAX It can be calculated by formula (2) and (3): In the formula, Assemble for the unloading task; Assemble for the task of loading boxes on board; The moment of releasing the yard bridge for the port operation task u; The moment to release the quay crane for port operations.

3. A method for optimizing port production operation scheduling considering refrigerated container operation demand according to claim 1, characterized in that: In the above-mentioned S6, the four types of container truck driving paths are specifically as follows: the first is from the quay crane to the storage yard, completing the task of unloading and stacking containers; the second is from the storage yard to the quay crane, completing the task of picking up containers and loading them onto the ship; the third is from the storage yard to the storage yard, that is, after stacking the containers, go to another container area to pick up the containers; the fourth is from the quay crane to the quay crane, that is, after completing the task of loading and delivering containers on a quay crane, go to another quay crane to start the task of unloading and picking up containers.

4. A method for optimizing port production operation scheduling considering refrigerated container operation demand according to claim 1, characterized in that: The value of M in formula (40) is 9 99 .

5. The method for optimizing port production operation scheduling considering refrigerated container operation demand according to claim 1, characterized in that: The S8 is specifically: Starting from the initial solution, the ALNS algorithm searches in the neighborhood space of the solution. The ALNS algorithm constructs a new solution by adaptively selecting the destruction operator and the repair operator. When the new solution is better than the existing solution, the new solution is accepted and searched in the neighborhood space of the new solution. If the new solution is not better than the current solution, the probability of retaining the inferior solution is calculated according to the Metropolis criterion to increase the search space. The initial solution is constructed using heuristic rules. The algorithm assigns operating machines to refrigerated container tasks and arranges the order of operations according to the order of the start time of the operation time window, and then inserts ordinary container tasks at feasible locations. The execution time of the work task is to start as early as possible, and the execution time of each task is calculated sequentially.

6. A method for optimizing port production operation scheduling considering refrigerated container operation demand according to claim 1, characterized in that: In the above-mentioned S8, in order to improve the solution efficiency and avoid deadlock problems during the solution process, the algorithm adds a decoding sequence table; the sequence table records the predecessor and successor tasks of each container task; when a task is deleted due to damage to the solution, the sequence table synchronously extracts and deletes the task; when the damaged solution is repaired, the repair operator reads the sequence table, obtains the feasible insertion position of the current task to be inserted, and then selects an insertion position from the feasible insertion positions according to the repair rules of the repair operator to perform the repair; each time a task is repaired and inserted, the sequence table is synchronously updated, and then the next repair and insertion is performed.

7. A method for optimizing port production operation scheduling considering refrigerated container operation requirements according to any one of claims 1 to 6, characterized in that: The hypothetical scenarios of the port production operation scheduling optimization method include: Scenario 1: The cross-berth scheduling of the quay crane is not considered, that is, the quay crane operation area is known; Scenario 2: During the work shift, the remaining capacity of the yard is known; Scenario 3: The shipping containers have been assembled at the port in advance, and the assembly method and arrival time are determined by the route optimization model; Scenario 4: Container trucks in the port follow the established driving rules without considering the path conflicts between container trucks; Scenario 5: Only the internal container truck scheduling is considered, and the impact of external container trucks is not considered; Scenario 6: The yard crane moves within a fixed container area, and cross-regional scheduling is not considered.

Images