A method and system for ship segment logistics scheduling that integrates cross-system business data

Abstract

This application provides a method and system for ship segment logistics scheduling based on cross-system business data fusion, belonging to the technical field of ship segment logistics scheduling. The invention acquires yard status information; obtains path selection information based on the yard status information; obtains an objective function for a yard status prediction sequence based on the path selection information; obtains ship storage location prediction information based on the objective function of the yard status prediction sequence; and completes ship segment logistics scheduling based on the ship storage location prediction information. This invention enables effective training of large-scale yard layout optimization models, achieving yard layout optimization while protecting shipyard data.

Description

Technical Field

[0001] This application relates to the technical field of ship segment logistics scheduling, and in particular to a ship segment logistics scheduling method and system that integrates cross-system business data. Background Technology

[0002] In the field of smart manufacturing, warehouses play a crucial role in the supply chain process. Warehouses balance costs and customer service quality by improving space utilization and optimizing spatial layout. Warehouses comprise four main functions: receiving, storage, order picking, and shipping. Storage is the most common function in a warehouse, and the process may involve issues such as regional item allocation, cross-regional item allocation, and storage location allocation. Disadvantages of related technologies:

[0003] 1. A large amount of data is needed to train the model, but currently available datasets make modeling difficult.

[0004] 2. Existing deep reinforcement learning models can only be used for small and medium-sized storage yards and cannot solve the state space explosion problem caused by large-scale storage yards. Summary of the Invention

[0005] The main objective of this application is to provide a method and system for ship segment logistics scheduling that integrates cross-system business data.

[0006] The technical solution adopted in this invention is:

[0007] On one hand, embodiments of the present invention provide a method for ship segment logistics scheduling based on cross-system business data fusion, the method comprising the following steps:

[0008] Obtain the status information of the stockpile;

[0009] Based on the stockpile status information, obtain path selection information;

[0010] Based on the path selection information, the objective function for the stockpile state prediction sequence is obtained;

[0011] Based on the objective function of the yard status prediction sequence, the predicted information of the ship storage location is obtained;

[0012] Based on the predicted ship storage location information, the logistics scheduling of ship segments is completed.

[0013] Furthermore, obtaining the stockpile status information includes the following steps:

[0014] Obtain vessel location information;

[0015] Obtain obstacle location information;

[0016] Obtain the timetable information of ships within the storage yard;

[0017] The ship location information, the obstacle location information, and the ship timetable information within the storage yard are used as the storage yard status information.

[0018] Furthermore, obtaining path selection information based on the yard status information includes the following steps:

[0019] Obtain the starting position of the new vessel;

[0020] Obtain the current location of the new vessel;

[0021] Obtain new candidate vessel locations;

[0022] Based on the new vessel's starting position, current position, and candidate target position, path loss information is calculated.

[0023] Based on the path loss information, the storage location sequence is calculated using the WNGO algorithm;

[0024] Based on the storage location sequence and the stacking yard status information, path selection information is obtained.

[0025] Furthermore, the path loss information is calculated based on the new vessel's starting position, current position, and candidate target position using the following formulas:

[0026]

[0027] in, Let be the total path loss in the stockpile state at time t. The actual path loss is the distance from the new vessel's starting position to its current position, including the loss from moving obstacle blocks. The path loss is the distance from the current position of the new vessel to the candidate target position of the new vessel.

[0028] Further, the step of calculating the storage location sequence using the WNGO algorithm based on the path loss information includes the following steps:

[0029] The WNGO algorithm is obtained by fusing the Whale Optimization Algorithm and the Neighborhood Gradient Optimization Algorithm.

[0030] Based on the path loss information and the WNGO algorithm, optimization information on the state transition lag and predictability of the stockpile is obtained;

[0031] Based on the optimization information of the state transition lag and predictability of the storage yard, a storage location sequence is obtained.

[0032] Furthermore, the objective function used to obtain the stockpile state prediction sequence based on the path selection information includes the following formula:

[0033]

[0034] Among them, R t Select a reward for the storage location of the stockpile state at time t, W = [ω1, ω2, ..., ω N [ ] represents the weights for the predicted state of the storage yard; MaxF(R t ,W) is the objective function for predicting the state sequence of the stockyard;

[0035] For path loss of new ship blocks entering the stockyard, The average path loss for all rearranged ship blocks transported out of the yard in the next rearrangement task. Average path loss for clearing the stockpile;

[0036] The number of times the new ship block is moved into the storage yard, This represents the average number of relocations of all rearranged ship blocks within the storage yard. The average number of moves required to clear the storage yard.

[0037] Further, obtaining the ship storage location prediction information based on the objective function of the yard state prediction sequence includes the following steps:

[0038] Based on the objective function of the stockpile state prediction sequence, state fitness evaluation information is obtained;

[0039] Based on the state fitness assessment information, duplicate states are eliminated to obtain the predicted ship storage location information.

[0040] On the other hand, embodiments of the present invention also provide a ship segment logistics scheduling system that integrates cross-system business data, the system comprising:

[0041] The first module is used to obtain the status information of the stockpile.

[0042] The second module is used to obtain path selection information based on the stockpile status information;

[0043] The third module is used to obtain the objective function of the stockpile state prediction sequence based on the path selection information;

[0044] The fourth module is used to obtain ship storage location prediction information based on the objective function of the yard state prediction sequence;

[0045] The fifth module is used to complete the segmented logistics scheduling of the ship based on the predicted information of the ship's storage location.

[0046] On the other hand, embodiments of the present invention also provide a ship segment logistics scheduling device for cross-system business data fusion, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the ship segment logistics scheduling method for cross-system business data fusion as described above.

[0047] On the other hand, embodiments of the present invention also provide a computer-readable storage medium storing computer-executable instructions for causing a computer to perform the methods described above.

[0048] The embodiments of this application include at least the following beneficial effects: This application provides a method and system for ship segment logistics scheduling based on cross-system business data fusion. The invention acquires yard status information; acquires path selection information based on the yard status information; obtains an objective function for a yard status prediction sequence based on the path selection information; obtains ship storage location prediction information based on the objective function of the yard status prediction sequence; and completes ship segment logistics scheduling based on the ship storage location prediction information. This invention enables effective training of large-scale yard layout optimization models, achieving yard layout optimization while protecting shipyard data. Attached Figure Description

[0049] Figure 1 This is a flowchart of a ship segment logistics scheduling method for cross-system business data fusion provided in an embodiment of the present invention;

[0050] Figure 2 This is a schematic diagram of the shipyard yard layout and rearrangement task flow provided in an embodiment of the present invention;

[0051] Figure 3 This is a schematic diagram of the decision-making process for new ship blocks entering the stockyard, provided by an embodiment of the present invention;

[0052] Figure 4 This is a schematic diagram of the optimization of large-scale stockpile layout based on deep reinforcement learning provided in an embodiment of the present invention;

[0053] Figure 5 This is a schematic diagram of the prediction of the storage status of the stockpile provided in an embodiment of the present invention;

[0054] Figure 6 This is a schematic diagram illustrating an example of using SOA for new ship block storage and rearrangement provided in an embodiment of the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0056] It is understood that the terms “first,” “second,” etc., used in this application may be used herein to describe various concepts, but unless otherwise stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the words “if,” “when,” or “in response to a determination” as used herein may be interpreted as “when…” or “when…” or “in response to a determination.”

[0057] As used in this application, the terms "at least one", "multiple", "each", "any", etc., "at least one" includes one, two or more, "multiple" includes two or more, "each" refers to each of the corresponding multiples, and "any" refers to any one of the multiples.

[0058] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0059] Before providing a detailed description of the embodiments of this application, some of the nouns and terms involved in the embodiments of this application will be explained first. The nouns and terms involved in the embodiments of this application are subject to the following interpretations.

[0060] 1) WOA (Whale Optimization Algorithm) is a heuristic optimization algorithm based on the hunting behavior of humpback whales in nature.

[0061] 2) NGO (Neighborhood Gradient Optimization) is an optimization algorithm that combines gradient information and neighborhood search.

[0062] 3) PSLAP (Partial-State Linear Approximation Problem) is a problem of partial-state linear approximation. This problem involves how to effectively approximate states when dealing with large-scale state spaces in order to reduce computational complexity and storage requirements.

[0063] 4) PPO (Proximal Policy Optimization) is an optimization algorithm used to solve reinforcement learning problems.

[0064] 5) DRL (Deep Reinforcement Learning) is a technique that combines deep learning and reinforcement learning to solve complex decision-making and control problems.

[0065] 6) SOA (Stack Optimization Algorithm) is an algorithm used to optimize stack storage.

[0066] 7) WNGO is the similar coding whale-eagle optimization algorithm of this invention. It combines the features of the Whale Optimization Algorithm (WOA) and the Eagle Optimization Algorithm (EOA) to solve the storage location optimization problem in the stockyard.

[0067] This invention addresses the challenges of state space explosion and the lack of public datasets when using deep reinforcement learning to solve large-scale PSLAP problems. By standardizing storage rules, this invention models large-scale PSLAP problems. Then, based on the lag in the state transition of the stockpile, a heuristic algorithm is used to obtain the top-k optimal storage coordinates of the current stockpile state. These top-k coordinates are then used to generate the corresponding stockpile state for pre-training of the deep reinforcement learning model.

[0068] The main objectives of this invention are as follows:

[0069] ① By using heuristic algorithms to simulate human experience in selecting the storage location of ship blocks in the yard, the problem of insufficient training data can be alleviated. The dataset constructed by heuristic algorithms achieves better results than human experience and is more conducive to model training.

[0070] ② Based on the above heuristic method for generating training data, and combined with the lag in the state transition of the storage yard, a deep reinforcement learning training method for large storage yards is proposed to solve the state space explosion problem in the model training process.

[0071] ③ Combining the PPO algorithm, design a deep reinforcement learning optimization model for yard layout optimization.

[0072] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0073] On the one hand, embodiments of the present invention provide 1. a method for ship segment logistics scheduling based on cross-system business data fusion, referring to... Figure 1 The method includes the following steps:

[0074] S100, Obtain the stockpile status information;

[0075] S200. Obtain route selection information based on the yard status information;

[0076] S300. Based on the path selection information, obtain the objective function of the stockpile state prediction sequence;

[0077] S400. Based on the objective function of the yard status prediction sequence, the predicted information of the ship storage location is obtained;

[0078] S500: Based on the predicted location of the ship, complete the logistics scheduling of ship segments.

[0079] This invention discloses step S100 for obtaining stockpile status information, including the following steps:

[0080] S110. Obtain vessel location information;

[0081] S120, Obtain obstacle location information;

[0082] S130. Obtain the timetable information of ships in the storage yard;

[0083] S140. The storage yard status information is based on the ship position information, obstacle position information, and the timetable information of ships in the storage yard.

[0084] This invention discloses step S200, which involves obtaining path selection information based on yard status information, including the following steps:

[0085] S210, Obtain the new vessel's starting position;

[0086] S220, Obtain the current location of the new vessel;

[0087] S230, Obtain the location of new candidate targets for ships;

[0088] S240. Based on the new vessel's starting position, current position, and candidate target position, the path loss information is calculated.

[0089] S250. Based on the path loss information, the storage location sequence is calculated using the WNGO algorithm;

[0090] S260. Obtain path selection information based on the storage location sequence and yard status information.

[0091] As an optional implementation, the present invention obtains the starting position of the new vessel in S210, that is, the position of the vessel before entering the yard. This position is usually the starting point of the vessel outside the yard, which can be the initial position of the dock or transport vehicle.

[0092] In step S220, the current position of the new vessel in the storage yard is obtained. This position is the coordinate of the vessel after it entered the storage yard and moved to the current point.

[0093] In step S230, candidate target locations for new vessels within the yard are obtained. These locations are places where vessels can be stored, determined based on the yard layout and the status of other vessels. Implementation methods include:

[0094] Candidate locations are generated: Based on the yard layout and storage rules, all possible target storage locations are calculated. This typically requires taking into account the locations of existing vessels and available units.

[0095] Evaluation of candidate locations: The candidate locations are initially evaluated using the status information of the storage yard (such as storage density and available space) to ensure their suitability.

[0096] Candidate position selection: Provide a list of qualified candidate positions to subsequent algorithms for further optimization.

[0097] In step S240, the path loss from the new vessel's current position to each candidate target position is calculated. Path loss refers to the total cost or time from the current position to the target position, including actual travel time and additional losses due to obstacles. Implementation methods include:

[0098] Define the path loss function: it includes the distance loss from the current position to the target position and the additional loss due to obstacles.

[0099] Path search: Calculate the loss value of different paths using the A* algorithm or other path search algorithms.

[0100] Calculate loss information: Based on the calculation results of the path loss function, obtain the loss information for each path.

[0101] In step S250, the WNGO algorithm (combining the Whale Optimization algorithm and the Northern Eagle Optimization algorithm) is used to calculate the optimal storage location sequence. This sequence represents the optimal storage location configuration, minimizing overall path loss and other constraints. Implementation methods include:

[0102] Initialize the WNGO algorithm: Set the algorithm parameters, including the weights of the Whale Optimization Algorithm and the Northern Eagle Optimization Algorithm.

[0103] Application of WNGO: The WNGO algorithm is used to optimize path loss information and generate a sequence of storage locations.

[0104] Update the model: Based on the results of the WNGO algorithm, update the storage location sequence and optimize the yard layout.

[0105] In step S260, based on the optimized storage location sequence and the current status information of the storage yard, the optimal route selection for the new vessel is determined. This information guides the vessel's specific storage operations within the storage yard. The implementation method includes:

[0106] Apply storage location sequence: Apply the optimized storage location sequence to the current yard state to calculate the final path selection.

[0107] Generate path information: Based on the sequence of storage locations, generate the final path selection information, including the movement route and storage points.

[0108] Update yard status: Apply route selection information to the yard layout to update the yard status in real time.

[0109] This invention discloses step S240, which calculates path loss information based on the new ship's starting position, current position, and candidate target position. The formula used includes:

[0110]

[0111] in, Let be the total path loss in the stockpile state at time t. This represents the actual path loss from the new vessel's starting position to its current position. The actual path loss includes the loss incurred from moving obstacle blocks. The path loss is the distance from the current position of the new vessel to the candidate target position of the new vessel.

[0112] This invention discloses step S250, which calculates the storage location sequence using the WNGO algorithm based on path loss information, including the following steps:

[0113] S251. The WNGO algorithm is obtained by fusing the whale optimization algorithm and the neighborhood gradient optimization algorithm.

[0114] S252. Based on the path loss information and the WNGO algorithm, obtain the optimization information on the state transition lag and predictability of the stockpile.

[0115] S253. Based on the state transition lag and predictability optimization information of the storage yard, obtain the storage location sequence.

[0116] This invention discloses step S300, which obtains the objective function of the stockpile state prediction sequence based on path selection information. The formula used includes:

[0117]

[0118]

[0119] Among them, R t Select a reward for the storage location of the stockpile state at time t, W = [ω1, ω2, ..., ω N [ ] represents the weights for the predicted state of the storage yard; MaxF(R t ,W) is the objective function for predicting the state sequence of the stockyard;

[0120] For path loss of new ship blocks entering the stockyard, The average path loss for all rearranged ship blocks transported out of the yard in the next rearrangement task. Average path loss for clearing the stockpile;

[0121] The number of times the new ship block is moved into the storage yard, This represents the average number of relocations of all rearranged ship blocks within the storage yard. The average number of moves required to clear the storage yard.

[0122] This invention discloses step S400, which obtains ship storage location prediction information based on the objective function of the yard status prediction sequence, including the following steps:

[0123] S410. Based on the objective function of the stockpile state prediction sequence, obtain the state fitness evaluation information;

[0124] S420. Based on the state fitness assessment information, duplicate states are eliminated to obtain the ship storage location prediction information.

[0125] This invention summarizes the problems related to the relevant technologies:

[0126] 1. The problem of allocating planar storage locations during ship section processing.

[0127] Large shipyards typically have stockyards divided into normal stockyards and common stockyards. Normal stockyards are for specific workshops, while common stockyards store timber blocks from multiple workshops. If a ship block has completed one process but the next process cannot begin immediately, it will be stored in a normal stockyard near the next workshop. If there is no space remaining in the normal stockyard, the ship block will be stored in a common stockyard. Normal stockyards are relatively small and can hold a limited number of ship blocks. Therefore, there is little temporary relocation work in normal stockyards when ship blocks are moved in and out of the yard. However, common stockyards are larger and can store a large number of ship blocks, leading to numerous obstacles when moving them in and out. In this situation, minimizing the use of transport vehicles is also important, as the resource consumption of transport vehicles during temporary relocations is significant.

[0128] The task of moving ship blocks from common stockyards to different locations is defined as a rearrangement task. This task is typically performed to remove obstructing ship blocks that hinder the movement of the target block. If transport vehicles had to perform rearrangement work every time they transported a ship block, it would not only affect the efficiency of the stockyard but also significantly increase transportation costs. Therefore, in large shipyards, rearrangement work is usually carried out during breaks after normal working hours. The rearrangement task moves ship blocks that need to be transported to the workshop for processing in the next working period to a specific location in the common stockyards near the main road (see...). Figure 2 The green area in Common Stockyards indicates the exit location. Rearranging is one of the tasks that must be minimized in the daily operation of the stockyard. Optimizing the stockyard layout by fully considering the storage locations of other stockyards and their exit times when they arrive is an effective way to reduce rearrangement tasks. (Reference: Shipyard Stockyard Layout and Rearrangement Task Process) Figure 2 .

[0129] 2. Decision-making issues in the optimization of shipyard yard layout

[0130] Optimizing the layout of a shipyard yard requires considering factors such as the storage location of new ship blocks upon arrival, the paths for ship blocks to enter and exit the yard, and the selection of storage locations for obstructed ship blocks. Taking the process of new ship blocks entering the yard as an example... Figure 3As shown, when a new ship block arrives, a target storage location needs to be selected. Then, based on the actual conditions of the storage yard, an entrance and a path to the target location are selected. The fitness of the target location is calculated along the selected path, and the above steps are repeated until the optimal storage location is confirmed. After determining the optimal storage location and path, if there are obstructing ship blocks in the path, the obstructing ship blocks need to be moved to another storage location.

[0131] As an optional implementation, the technology of the present invention includes:

[0132] 1. A large-scale storage yard layout optimization model based on deep reinforcement learning

[0133] Optimization modeling of yard layout is a complex process, especially when new ship blocks encounter obstacles when entering the target storage location. As the obstructing ship blocks are moved, the yard state changes accordingly, making it difficult for the model to obtain a realistic reward after executing actions. To overcome this challenge, this work fully considers the stacking positions and schedules of other ship blocks within the yard when designing the reward function, further mitigating the impact of yard state changes on storage location selection. Based on the lag and predictability of yard state transitions, this paper proposes a similar coding whale-eagle optimization algorithm (WNGO) for predicting the optimal storage location in the yard.

[0134] like Figure 4 As shown in the left figure, "Stockyard State Prediction and Layout Model Pre-training," the top-k sequence of storage locations with the highest fitness are selected and transformed into corresponding stockyard states. These states are then used for the pre-training of a large-scale stockyard layout model. Since the reward for selecting a storage location is affected by the path leading to the stockyard, different paths result in different rewards. This paper uses the A* algorithm for path search to obtain the optimal path. The selection of storage locations includes choosing the storage location after a new block arrives in the common stockyards and choosing the location when an obstacle block re-enters the common stockyards.

[0135] 2. Yard Condition Prediction Based on WNGO

[0136] (1) Prediction of the status of the storage yard

[0137] like Figure 5As shown, the system acquires the actual state of the storage yard at time t0 as the initial state. Based on this, the WNGO algorithm is used to calculate the storage location with the highest fitness, and this location is transformed into the predicted state of the storage yard at time t1. Similarly, the predicted state of the storage yard at time t2 is predicted using the predicted state of the storage yard at time t1 as the initial state, and so on, ultimately obtaining a state sequence consisting of n predicted states. Due to the high uncertainty in the prediction of the storage yard state, the above operation needs to be repeated to obtain more predicted state sequences. Then, the top-k predicted state sequences with the highest fitness are obtained, and duplicate states in the sequences are removed to form the pre-training state set for DRL.

[0138] To reduce computational complexity, changes in the stockpile state caused by storing ship blocks at the current time are not considered, except for those caused by other forms of changes in the stockpile state. That is, the fitness calculation of the storage location and the selection of the transportation route at time t1 only refer to the initial state of the stockpile at time t1, while the calculation at time t2 uses the stockpile state after the operation at time t1. The objective function of the stockpile state prediction sequence is defined as shown in Equation (1).

[0139]

[0140]

[0141] Among them, R t Select a reward for the storage location of the stockpile state at time t, W = [ω1, ω2, ..., ω N [ ] represents the weight of the predicted state of the storage yard. For path loss of new ship blocks entering the stockyard, The average path loss for all rearranged ship blocks transported out of the yard in the next rearrangement task. The average path loss for clearing the storage yard. Similarly, The number of times the new ship block is moved into the storage yard, This represents the average number of relocations of all rearranged ship blocks within the storage yard. The average number of moves required to clear the storage yard.

[0142] (2) WNGO optimization algorithm

[0143] WOA updates the positions of whales randomly based on their respective locations rather than the current best whale position. While this improves the global search capability of WOA in the early stages, its local search capability becomes insufficient in the later stages. NGO is designed to simulate the hunting behavior of northern eagles, including two stages: prey identification and attack, and pursuit and escape. Although it also updates the positions randomly based on the locations of their prey rather than the current best eagle position, its global search efficiency is lower than that of the WOA algorithm. The second stage of development enhances the local search capability of the NGO algorithm, enabling it to obtain better combinations than WOA within a feasible timeframe. This invention integrates the WOA and NGO algorithms to design the WNGO algorithm. It utilizes the WOA algorithm to optimize the positions of some eagle prey, guiding the NGO algorithm to search towards a more optimal space while avoiding getting trapped in local optima.

[0144] Phase 1: Prey Identification (Global Search)

[0145] In the first stage of hunting, the Northern Goshawk randomly selects prey and quickly attacks it. This random selection of prey in the search space enhances the global exploration capability of the WNGO algorithm, aiming to determine the optimal region. However, the efficiency of this random search is very low. The WNGO algorithm divides the prey's position X into two parts: one is the prey position generated using the WOA algorithm; the other is a prey position that prevents the algorithm from getting trapped in local optima. The guiding factor is used to quickly determine a high-quality search region, while the random factor prevents the algorithm from getting trapped in local optima.

[0146] P i ~X i , i=1,2,…,N, j=1,2,…,Z, (3)

[0147]

[0148] Among them, P i It is the location of the i-th prey of the northern goshawk. It is its objective function value. X i For the location of the Northern Goshawk, F i It is its objective function value. The new recommended position is generated at time t. This represents the position of the prey at time t-1. In formula (5), t is the iteration counter, and T is the maximum number of iterations. Formula (6) expresses the update process of the prey's position. This represents the j-th dimension of the prey's position at time t. This is the objective function value of the positional variable at time t-1. for The j-th dimension position variable, It is its objective function value. In formula (7) yes The j-th dimension position variable, Rd() is the rounding function, r is a random number in the interval [0,1], and I is a random number with a value of 1 or 2. Parameters r and I are random numbers used to generate randomness in the search and update behavior. In formula (8) This is the first phase of the new service package. It is its objective function value.

[0149] Phase Two: Chase and Escape (Local Search)

[0150] After a Northern Wilderness Eagle attacks its prey, the prey will attempt to escape. Therefore, the eagle continues to pursue and chase the prey. Due to its high speed, the eagle can catch its prey in almost any situation and ultimately capture it. Simulating this behavior enhances the algorithm's ability to perform local searches of the search space.

[0151]

[0152]

[0153] For the new solution, The objective function value, yes The j-th dimension position variable.

[0154] 3. Path selection optimization based on A* algorithm

[0155] The A* algorithm is an efficient heuristic search algorithm with good path optimization effect on two-dimensional grid maps. Therefore, the A* algorithm is used as the path search algorithm for storage locations in the storage yard. The definition of the A* algorithm is shown in formula (16).

[0156]

[0157] in, Let be the total path loss in the stockpile state at time t. This represents the actual path loss from the starting position to the current search position (including the loss from moving obstacle blocks). This represents the path loss from the current search location to the destination location.

[0158] The A* algorithm estimates the cost of each path node using a function and assigns it as the node's overall priority. When selecting the next node to traverse, it chooses the node with the highest overall priority, gradually finding the optimal path. Therefore, the A* algorithm can quickly find the path with the minimum cost and shortest distance. The specific process is as follows:

[0159] (1) Obtain the state of the stockpile at time t, and obtain the list of starting node s and target nodes D = [d1, d2, ..., d2]. L Initialize the open and close lists;

[0160] (2) s is the starting node, d i As the target node, s is treated as a square to be inspected and saved to the open list;

[0161] (3) Find the grid nodes that can be reached around the starting node s, put them into the open list, and set s as the parent node;

[0162] (4) Remove the starting node s from the open list and save it to the close list;

[0163] (5) Calculate the value of each neighboring square. value;

[0164] (6) Select from the open list The cell with the smallest value, M1, is removed from the open list and saved to the close list;

[0165] (7) Check all adjacent squares of M1:

[0166] ① The cells in the close list are no longer considered;

[0167] ② If the square is not in the open list, add it to the open list and calculate. Set the value and set M1 as the parent node;

[0168] ③ If an adjacent cell M2 is already in the open list, calculate the new path from s to cell M2, and determine whether the parent node and the open list need to be updated. Value: If the new node M2 If the value is lower, set M1 as the parent node and recalculate. The value of H remains unchanged if the new node M2's A higher value indicates that the new path has a higher cost. The value remains unchanged.

[0169] (8) Repeat (6) to (7) until the target node d appears in the open list. i Obtain the distance from the starting node s to the target node d. i The optimal path;

[0170] (9) Repeat (2) to (8) to obtain the optimal path from the starting node s to all target nodes in list D;

[0171] (10) Choose The path with the smallest value is taken as the globally optimal path.

[0172] 4. Optimization of Ship Block Storage in Stockyards Based on Deep Reinforcement Learning Algorithms

[0173] (1) Analysis of the optimization problem of ship block storage in the yard

[0174] To reduce the overall resource consumption of new ship block storage and rearrangement tasks in large common stockyards, this study trained a stockyard storage optimization agent (SOA). The SOA's state, actions, and rewards are defined in Table 1. The SOA's reward is proportional to the number of relocations and the length of the transport path at each stage. It is assumed that the stockyard is a rectangular grid with entrances on any side, containing n storage cells horizontally and m storage cells vertically. Due to uncertainties causing errors in the ship block schedule, this invention divides the storage cell state into four levels, such as... Figure 6 As shown. Level 0 indicates the storage unit is empty. Level 1 indicates the block within the storage unit will be used in the next working period. Level 2 indicates the block will be used in the near future. Level 3 indicates the block will not be used in the near future or its timetable is unclear. Level transitions can only strictly follow the constraints and decrease progressively. Therefore, the stockpile state at time t is... l = [0, 1, 2, 3]. Since the reward for selecting storage needs to consider the impact on the rearrangement task, and the ship blocks that need to be rearranged in the next working period within the yard are known, SOA still uses formula (2) as the reward function.

[0175] Table 1

[0176]

[0177] SOA can be used in both the new shipment arrival phase and the rearrangement phase. When a new shipment arrives at the yard perimeter, SOA first determines the storage location, then uses the A* algorithm to calculate the entrance coordinates and the route with the least loss. The new shipment and obstructing shipments are temporarily placed at the yard exit, which cannot be on the route with the least loss. After the obstructing shipments are moved, a transport vehicle moves the new shipment to the designated storage location. Then, the shipment schedule is observed; if the shipment at the exit is grade 1, it will remain at the exit; otherwise, it will be treated as a new shipment and a new storage location will be selected within the yard. If there are multiple obstructing shipments that need to be rearranged at the exit, they are rearranged in reverse order of their relocation. This rule is to prevent a shipment that entered the yard first from becoming a new obstructing shipment. The rearrangement phase differs in that the order of rearranging shipments within the yard needs to be determined first, and then the exit and route are determined sequentially using the A* algorithm. Finally, a transport vehicle is used to realize the shipment rearrangement and the placement of obstructing shipments.

[0178] (2) Reinforcement learning algorithm for solving the ship block layout problem

[0179] In real-world applications, state transitions in a stockpile are predictable, and the probability of large-scale state transitions is very low. Therefore, to maintain model stability, this invention uses the PPO algorithm to train the SOA. PPO is a policy gradient algorithm for reinforcement learning, designed to limit the magnitude of each policy update by introducing a regularization term, thereby ensuring the stability of the training process.

[0180] The Proof-of-Policy (PPO) algorithm has two main methods for updating policy gradients: KL penalty and Clip. Their primary function is to limit the magnitude of policy gradient updates, thereby deriving different neural network parameter update methods. Compared to KL penalty, Clip has advantages such as easier-to-understand update rules, no need to set additional hyperparameters, and more explicit constraints on policy distribution updates. Therefore, this invention uses Clip as the method for updating policy gradients. PPO optimizes the policy by calculating the surrogate loss to constrain the change magnitude between the old and new policies through the Clip policy constraint. At a given time step t, for action a... t The ratio of action probabilities r under the new and old strategies t The calculation of (θ) is shown in formula (15). Where π θ (a t |s t ) indicates that in state s t Take action a t New strategy π θ The probability, Indicates that in state s t Take action a t old strategy The probability of.

[0181]

[0182] The Clip loss function is used as part of the total loss function, and its definition is shown in formula (18). This represents the expected estimate for time step t. This is the estimated advantage value at time step t. `clip()` is a clamping function that takes the input r... t (θ) is restricted to the range (1-∈, 1+∈), where ∈ is a hyperparameter.

[0183]

[0184] use Let be the total loss function. For the value function loss, c2E r (π θ) represents the regularization term of the policy distribution entropy, and c1 and c2 are hyperparameters.

[0185]

[0186] The beneficial effects of this invention are as follows:

[0187] (1) Even when actual yard layout data is unavailable and there is a lack of corresponding public datasets, it is still possible to effectively train a large-scale yard layout optimization model, thereby optimizing the yard layout while protecting shipyard data.

[0188] (2) Since the generated data distribution is more uniform than that of actual human experience operation, our model will have higher robustness and accuracy when dealing with extreme situations and real-time storage decisions.

[0189] On the other hand, embodiments of the present invention also provide a ship segment logistics scheduling system that integrates cross-system business data, the system comprising:

[0190] The first module is used to obtain the status information of the stockpile.

[0191] The second module is used to obtain route selection information based on the yard status information;

[0192] The third module is used to obtain the objective function of the stockpile state prediction sequence based on the path selection information;

[0193] The fourth module is used to obtain ship storage location prediction information based on the objective function of the yard status prediction sequence.

[0194] The fifth module is used to complete the logistics scheduling of ship segments based on the predicted information of ship storage location.

[0195] On the other hand, embodiments of the present invention also provide a ship segment logistics scheduling device for cross-system business data fusion, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the ship segment logistics scheduling method for cross-system business data fusion as described above.

[0196] The ship segment logistics scheduling device for cross-system business data fusion according to an embodiment of the present invention includes a memory and a processor.

[0197] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0198] Memory can include various types of storage units, such as system memory, read-only memory (ROM), and permanent storage devices. ROM can store static data or instructions required by the processor or other modules of the computer. Permanent storage devices can be read-write storage devices. Permanent storage devices can be non-volatile storage devices that retain stored instructions and data even when the computer is powered off. In some embodiments, permanent storage devices use high-capacity storage devices (e.g., magnetic or optical disks, flash memory) as permanent storage devices. In other embodiments, permanent storage devices can be removable storage devices (e.g., floppy disks, optical drives). System memory can be a read-write storage device or a volatile read-write storage device, such as dynamic random access memory. System memory can store some or all of the instructions and data required by the processor during operation. Furthermore, memory can include any combination of computer-readable storage media, including various types of semiconductor memory chips (DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), and disks and / or optical disks can also be used. In some implementations, the memory may include removable storage devices that are readable and / or writable, such as laser discs (CDs), read-only digital versatile optical discs (e.g., DVD-ROMs, dual-layer DVD-ROMs), read-only Blu-ray discs, ultra-high density optical discs, flash memory cards (e.g., SD cards, mini SD cards, Micro-SD cards, etc.), magnetic floppy disks, etc. Computer-readable storage media do not contain carrier waves or transient electronic signals transmitted wirelessly or via wired connections.

[0199] The memory stores executable code, which, when processed by the processor, can cause the processor to execute some or all of the methods described above.

[0200] On the other hand, embodiments of the present invention also provide a computer-readable storage medium storing computer-executable instructions for causing a computer to perform the methods described above.

[0201] It will be understood by those skilled in the art that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0202] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A method for ship segment logistics scheduling that integrates cross-system business data, characterized in that, The method comprises the following steps: Obtain the status information of the stockpile; Based on the yard status information, path selection information is obtained, including: Obtain the starting position of the new vessel; Obtain the current location of the new vessel; Obtain new candidate vessel locations; Based on the new vessel's starting position, current position, and candidate target position, path loss information is calculated. Based on the path loss information, the storage location sequence is calculated using the WNGO algorithm, including: The WNGO algorithm is obtained by fusing the Whale Optimization Algorithm and the Neighborhood Gradient Optimization Algorithm. Based on the path loss information and the WNGO algorithm, optimization information on the state transition lag and predictability of the stockpile is obtained; Based on the optimization information regarding the state transition lag and predictability of the storage yard, a storage location sequence is obtained; Based on the storage location sequence and the stockpile status information, path selection information is obtained; Based on the path selection information, the objective function for predicting the stockpile state sequence is obtained, including: Among them, R t Select a reward for the storage location of the stockpile state at time t, W = [ω1, ω2, ..., ω N [ ] represents the weights for the predicted state of the storage yard; MaxF(R t ,W) is the objective function for predicting the state sequence of the stockyard; For the path loss of new ship blocks entering the stockyard, The average path loss for all rearranged ship blocks transported out of the yard in the next rearrangement task. Average path loss for clearing the stockpile; The number of times the new ship block is moved into the storage yard, This represents the average number of relocations of all rearranged ship blocks within the storage yard. The average number of moves required to clear the storage yard; Based on the objective function of the yard status prediction sequence, the predicted information of the ship storage location is obtained; Based on the predicted ship storage location information, the logistics scheduling of ship segments is completed.

2. The method according to claim 1, characterized in that, The process of obtaining the stockpile status information includes the following steps: Obtain vessel location information; Obtain obstacle location information; Obtain the timetable information of ships within the storage yard; The ship location information, the obstacle location information, and the ship timetable information within the storage yard are used as the storage yard status information.

3. The method according to claim 1, characterized in that, The path loss information is calculated based on the new vessel's starting position, current position, and candidate target position. The formulas used include: in, Let be the total path loss in the stockpile state at time t. The actual path loss is the distance from the new vessel's starting position to its current position, including the loss from moving obstacle blocks. The path loss is the distance from the current position of the new vessel to the candidate target position of the new vessel.

4. The method according to claim 1, characterized in that, The process of obtaining ship storage location prediction information based on the objective function of the yard status prediction sequence includes the following steps: Based on the objective function of the stockpile state prediction sequence, state fitness evaluation information is obtained; Based on the state fitness assessment information, duplicate states are eliminated to obtain the predicted ship storage location information.

5. A ship segment logistics scheduling system that integrates cross-system business data, used to implement the method as described in any one of claims 1 to 4, characterized in that, The system includes: The first module is used to obtain the status information of the stockpile. The second module is used to obtain path selection information based on the stockpile status information; The third module is used to obtain the objective function of the stockpile state prediction sequence based on the path selection information; The fourth module is used to obtain ship storage location prediction information based on the objective function of the yard state prediction sequence; The fifth module is used to complete the segmented logistics scheduling of the ship based on the predicted information of the ship's storage location.

6. A ship segment logistics scheduling device for cross-system business data fusion, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the ship segment logistics scheduling method for cross-system business data fusion as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions for causing a computer to perform the method as described in any one of claims 1 to 4.

Images