Intelligent scheduling optimization strategy for double-stacker of intelligent stereoscopic warehouse with same track

Abstract

This invention discloses an intelligent scheduling optimization strategy for dual stacker cranes on the same track in an intelligent automated storage and retrieval system (AS / RS). First, an optimization model for the allocation and scheduling of storage locations for dual stacker cranes on the same track is established based on dynamic storage and composite operation modes. Simultaneously, a joint optimization model for storage location allocation and stacker crane scheduling is established. Then, a multi-objective evolutionary algorithm for intelligent scheduling optimization of dual stacker cranes on the same track is used to solve the problem. This algorithm employs an improved NSGA-III algorithm based on breadth-search, including improvements to encoding, population initialization, adaptive crossover and mutation operators, and a reference point-based selection operator. This optimization strategy ensures that the time and running distance for stacker cranes to perform inbound and outbound tasks are minimized, while also considering the overall stability of the racking system and the turnover rate of goods within the warehouse, thereby improving the operational efficiency of the AS / RS.

Description

Technical Field

[0001] This invention belongs to the field of scheduling, and more specifically relates to an intelligent scheduling optimization strategy for dual stacker cranes on the same track in an intelligent automated warehouse. Background Technology

[0002] Automated Storage / Retrieval Systems (AS / RS) are a crucial component of modern warehousing, playing a vital role in logistics and supply chain management. AS / RS systems aim to integrate storage, transportation, picking, and management processes into a highly efficient, unified system to meet increasing production demands and logistical complexity. With the continuous intelligent upgrading of global logistics systems, the efficiency and operational requirements of AS / RS systems are also constantly increasing. Traditional intelligent automated warehouse (AS / RS) inbound and outbound operation modes can no longer meet the needs of modern warehousing environments. To address these issues, this invention proposes an innovative solution aimed at improving the performance and efficiency of AS / RS systems. The core lies in the AS / RS system's dual-track stacker crane, which employs both Single Command (SC) and Dual Command (DC) operation modes when performing inbound and outbound tasks. Selecting the correct operation mode is crucial for system performance; however, this also increases the complexity of system management. Currently, effectively scheduling dual-track stacker cranes to maximize their efficiency remains a challenging problem. Summary of the Invention

[0003] This invention introduces an intelligent scheduling optimization strategy for dual-track stacker cranes in smart warehouses. This strategy not only combines SC and DC operating modes but also dynamically allocates storage locations based on task queue characteristics and stacker crane position status. This enables more efficient storage and retrieval, minimizing cargo movement distance and storage location conflicts. In addition to the storage location allocation strategy, this invention also includes a stacker crane scheduling optimization algorithm. This algorithm considers the motion characteristics and operating mode switching of dual-track stacker cranes to rationally arrange stacker crane movement and task execution sequence. By reducing unnecessary waiting time and energy consumption, the performance of the entire AS / RS system can be improved.

[0004] To achieve the above objectives, the present invention employs the following technical solution: the optimization strategy includes:

[0005] Establish an optimization model for the allocation and scheduling of storage locations for dual stacker cranes on the same track, based on dynamic storage and composite operation modes;

[0006] Establish a joint optimization model for warehouse location allocation and stacker crane scheduling;

[0007] The above model is solved using a multi-objective evolution-based intelligent scheduling optimization algorithm for dual stacker cranes on the same track.

[0008] Furthermore, the aforementioned optimization model for the allocation and scheduling of storage locations for dual stacker cranes on the same track makes the following assumptions:

[0009] 1) The left and right stacker cranes in the same lane have the same attribute parameters;

[0010] 2) All storage locations / shelves within the automated warehouse are of standard dimensions;

[0011] 3) The inbound and outbound order information of the I / O units on both sides of the same aisle is known, and the existing storage location information in the automated warehouse is known.

[0012] Furthermore, the aforementioned optimization model for the allocation and scheduling of storage locations for dual stacker cranes on the same track includes:

[0013] Calculation of stacker crane running time and distance;

[0014] When a stacker crane performs inbound and outbound tasks, its travel distance and time directly affect the operating efficiency of the AS / RS system. Therefore, it is necessary to accurately calculate the operating performance of the stacker crane performing inbound and outbound tasks. The calculation process is as follows:

[0015] The operation of the stacker crane in the intelligent automated warehouse consists of the horizontal movement of the right slide rail and the longitudinal movement of the hoist. Therefore, the movement of the stacker crane is calculated from two directions. At the same time, since the two stacker cranes of the same rail double stacker crane belong to the I / O units at both ends of the aisle, the running distance of the stacker cranes on the left and right sides is calculated separately.

[0016] Furthermore, in mixed tasks, the operation modes of SC and DC are not the same, and calculations need to be carried out separately for different operation models.

[0017] Furthermore, the aforementioned joint optimization model for cargo location allocation and stacker crane scheduling;

[0018] Based on the operational requirements of the AS / RS system, the optimization of storage location allocation and stacker cranes must be ensured to minimize the time and travel distance for stacker cranes to perform inbound and outbound tasks.

[0019] At the same time, the overall stability of the shelving and the turnover rate of goods in the warehouse also need to be considered on the impact of the automated warehouse operation efficiency.

[0020] Therefore, an objective function is established to describe the joint optimization problem of intelligent automated warehouse location allocation and stacker crane.

[0021] Furthermore, the intelligent scheduling optimization algorithm for dual stacker cranes on the same track based on multi-objective evolution adopts the NSGA-III algorithm improved based on the breadth-search idea, specifically improving the following four parts:

[0022] 1) Encoding;

[0023] 2) Initialize the population;

[0024] 3) Adaptive crossover and mutation operators;

[0025] 4) Selection operator based on reference point.

[0026] Furthermore, in the joint optimization problem of CSP and SLAP, the aforementioned encoding includes two parts for the decision variable: the location allocation of inbound orders and the stacker crane scheduling of outbound orders; the solution of the model is defined using integer encoding.

[0027] Furthermore, the initial population is designed based on the idea of ​​breadth-first search, which is a method for dimensionality reduction of the initial population.

[0028] When initializing the population, a CSP initialization scheme is first generated using a random initialization strategy; then, the previous order is used as the initial node for breadth-first search to obtain the initial solution for the SLAP task.

[0029] Furthermore, the adaptive crossover and mutation operators increase the crossover and mutation probabilities when the fitness function values ​​of individuals in the population tend to a local optimum.

[0030] When the fitness function values ​​of individuals in a population are relatively dispersed, the probabilities of crossover and mutation are relatively reduced.

[0031] Meanwhile, individuals with relatively poor fitness levels have relatively high crossover and mutation probabilities, while individuals with relatively low fitness levels have relatively low crossover and mutation probabilities.

[0032] Furthermore, the reference point-based selection operator is introduced to improve the algorithm.

[0033] Beneficial effects of this invention:

[0034] The multi-domain applications of this invention aim to improve the performance and efficiency of AS / RS systems, meet ever-changing logistics needs and environmental challenges, and provide efficient logistics and warehousing solutions for multiple industries.

[0035] This invention introduces an intelligent scheduling optimization strategy for dual-track stacker cranes in smart warehouses. This strategy not only combines SC and DC operating modes but also dynamically allocates storage locations based on task queue characteristics and stacker crane position status. This enables more efficient storage and retrieval, minimizing cargo movement distance and storage location conflicts. In addition to the storage location allocation strategy, this invention also includes a stacker crane scheduling optimization algorithm. This algorithm considers the motion characteristics and operating mode switching of dual-track stacker cranes to rationally arrange the movement of stacker cranes and the order of task execution. By reducing unnecessary waiting time and energy consumption, the system can improve the performance of the entire AS / RS system. Attached Figure Description

[0036] Figure 1 Diagram showing SC and DC operation tasks;

[0037] Figure 2 This is a schematic diagram of a breadth-first search mechanism.

[0038] Figure 3 This is a schematic diagram of chromosome coding.

[0039] Figure 4 This is a schematic diagram of chromosome reconstruction.

[0040] Figure 5 Flowchart for improving the NSGA-III algorithm. Detailed Implementation

[0041] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Typical embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0042] Unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this invention and in its specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0043] The modern warehousing and logistics industry faces numerous challenges, one of which is maximizing the efficiency of AS / RS (Automatic Storage and Retrieval System) inbound and outbound operations. Traditional stacker crane systems have limitations in meeting production demands and logistical complexities because they typically operate under a SC (Search and Load) mode. While this mode can handle basic inbound and outbound tasks, it struggles to meet the demands of high-frequency and high-efficiency operations. Furthermore, the allocation of storage locations and orders during inbound and outbound operations presents significant challenges. In inbound operations, rationally allocating storage locations based on factors such as the frequency of goods movement, rack stability, and stacker crane travel distance to reduce goods movement and improve storage efficiency has been a persistent industry challenge. Similarly, in outbound operations, intelligently selecting goods that match order requirements while minimizing outbound time is also complex. Moreover, traditional AS / RS systems often lack the flexibility to adapt to varying inbound and outbound demands and environmental changes. This leads to inefficient resource utilization and increased long-term operating costs. Therefore, seeking an innovative solution to improve the performance, efficiency, and flexibility of AS / RS systems has become an urgent need.

[0044] Based on the aforementioned problems and challenges, the objective of this invention is to provide an intelligent scheduling optimization method for dual stacker cranes in intelligent automated storage and retrieval systems (AS / RS). By combining SC and DC operation modes with intelligent location allocation and stacker crane scheduling optimization strategies, this method addresses the limitations of traditional AS / RS systems. Ultimately, it aims to improve inbound and outbound efficiency, reduce energy consumption, increase resource utilization, shorten outbound time, and enable AS / RS systems to better cope with constantly changing logistics demands and environments. This will bring a more efficient and intelligent solution to the modern warehousing and logistics industry.

[0045] Optimization Model for Location Allocation and Scheduling of Dual Stacker Cranes on the Same Track Based on Dynamic Storage and Composite Operation Mode

[0046] Model assumptions:

[0047] 1) The left and right stacker cranes in the same lane have the same attribute parameters;

[0048] 2) All storage locations / shelves within the automated warehouse are of standard dimensions;

[0049] 3) The inbound and outbound order information of the I / O units on both sides of the same aisle is known, and the existing storage location information in the automated warehouse is known;

[0050] Symbol explanation:

[0051] Q shelves : The set of shelves in AS / RS, i∈Q shelves Indicates the index of the shelf;

[0052] Q aisles : The set of shelves in AS / RS, j∈Qaisles Indicates the index of the lane;

[0053] Q cargoes : The set of goods types in AS / RS, a∈Q cargoes Indicates a goods index;

[0054] Stacker crane located on the left or right side of lane j;

[0055] The sliding rail located on the left or right side of the j-th lane;

[0056] The location of the i-th shelf at position (x, y);

[0057] I / O unit located on the left or right side of the i-th shelf;

[0058] Decision variables:

[0059] Indicates the cargo location at time T. Is the storage location available? If this storage location is available at time T, otherwise

[0060] Indicates stacker crane Is the receiving task for goods A being performed? If the stacker crane is... In performing the task of receiving goods A into the warehouse, otherwise

[0061] express Is the outbound task for goods A being executed? If the stacker crane is... In executing the outbound task of goods A, otherwise

[0062] Indicate whether goods A are in the storage location. If goods a are in the storage location Warehousing otherwise

[0063] Indicate whether goods A are in the storage location. If goods A are in storage location when goods are shipped out. Outbound otherwise

[0064] Stacker crane running time and distance calculation

[0065] When a stacker crane performs inbound and outbound tasks, its travel distance and time directly affect the operating efficiency of the AS / RS system. Therefore, it is necessary to accurately calculate the operating performance of the stacker crane performing inbound and outbound tasks. The calculation process is as follows:

[0066] The operation of a stacker crane in an automated storage and retrieval system (AS / RS) consists of the horizontal movement of the right slide rail and the longitudinal movement of the hoist. Therefore, the movement of the stacker crane needs to be calculated from two directions. Furthermore, since the two stacker cranes on the same rail belong to I / O units at opposite ends of the aisle, the travel distances of the left and right stacker cranes need to be calculated separately. Moreover, in mixed operations, the operating modes of SC and DC are not entirely the same; therefore, calculations need to be performed separately for different operating models.

[0067] For the stacker crane performing the SC task on the left:

[0068]

[0069]

[0070]

[0071] Equation (1) represents the distance calculation method for the SC task performed on the left. This distance consists of four parts, namely the running distance of the slide rail when the stacker crane enters or leaves the storage location. and and the distance the elevator travels and Equations (2) and (3) are detailed calculation methods for the travel distance of the slide rail and the travel distance of the elevator, respectively. and , , and , respectively, are the x and y coordinates of the storage location for goods a. and These are the width and height of the storage space, respectively.

[0072] For the stacker crane performing the SC task on the right:

[0073]

[0074]

[0075]

[0076] Equation (4) represents the distance calculation method for the SC task performed on the right. This distance consists of four parts, namely the running distance of the slide rail when the stacker crane enters or leaves the storage location. and and the distance the elevator travels and Equations (5) and (6) are detailed calculation methods for the travel distance of the slide rail and the travel distance of the elevator, respectively, where x max and y max The x and y coordinates represent the positions of the right-side I / O station, respectively.

[0077] When executing SC tasks, regardless of whether it's outbound or inbound, the stacker crane only performs a single task, therefore, the following limitations exist:

[0078]

[0079]

[0080] Similarly, stacker cranes performing DC tasks also need to be calculated based on the I / O unit distribution of the stacker crane:

[0081] For the stacker crane performing the DC task on the left:

[0082]

[0083]

[0084]

[0085]

[0086] Equation (9) is the formula for calculating the running distance of a stacker crane performing a DC task, which is divided into three parts for calculation. This refers to the travel distance of the stacker crane from picking up goods from the I / O unit to storing the goods in the storage location. The distance a stacker crane travels from storing one item to picking the next. Equations (10)-(11) are detailed calculation methods for the stacker crane's running distance from picking up goods to shipping goods from the I / O unit.

[0087] For the stacker crane performing the DC task on the right:

[0088]

[0089]

[0090]

[0091]

[0092] Equation (13) represents the calculation method of the running distance of the right stacker crane when performing DC tasks, which is similar to the calculation method of the left stacker crane. Equations (14)-(15) are the detailed calculation methods of the running distance of the three parts of the right stacker crane.

[0093] When performing a DC (Distribution Control) task, the stacker crane needs to first carry the goods into the automated warehouse (AS / RS). After storing the goods in the AS / RS, it also needs to retrieve the goods for outbound orders and remove them from the racks. Therefore, the DC task has the following constraints:

[0094]

[0095]

[0096] Since the stacker cranes performing SC and DC tasks use different methods for calculating the running distance, the running time for different task types still needs to be calculated separately. The following is the method for calculating the running time of a stacker crane performing an SC task:

[0097]

[0098]

[0099]

[0100] Equation (19) represents the formula for calculating the running time of the stacker crane performing the SC task, and equations (20) and (21) represent the calculation methods for the running time of the two stages of the SC task, where and These represent the travel time of the lateral slide rail and the longitudinal lift when the stacker crane performs the first phase of the SC task, respectively. and The running time of the transverse slide rail and longitudinal lift when the stacker crane performs the second phase of SC task respectively meets the standard.

[0101] The following is the method for calculating the running time of a stacker crane when performing a DC task:

[0102]

[0103]

[0104]

[0105]

[0106] Equation (22) represents the formula for calculating the running time of the stacker crane performing the DC task, and equations (23)-(25) represent the calculation method for the running time of the three stages of the DC task, where and These represent the travel time of the lateral slide rail and the longitudinal lift when the stacker crane performs the first phase of the DC task, respectively. and These represent the travel time of the lateral slide rail and the longitudinal lift when the stacker crane performs the second phase of the DC task, respectively. and These represent the running time of the lateral slide rail and the longitudinal elevator, respectively, when the stacker crane performs the third stage of the DC task.

[0107] Joint optimization model of storage location allocation and stacker crane scheduling

[0108] Based on the operational requirements of the AS / RS system, the joint optimization problem of storage location allocation and stacker crane needs to simultaneously ensure that the stacker crane performs inbound and outbound tasks with the shortest possible time and travel distance. It also needs to consider the impact of factors such as the overall racking stability and the turnover rate of goods within the warehouse on the operational efficiency of the automated storage and retrieval system (AS / RS). Therefore, this invention establishes the following objective function to describe the joint optimization problem of storage location allocation and stacker crane in an intelligent automated storage and retrieval system:

[0109]

[0110]

[0111]

[0112]

[0113]

[0114]

[0115] Constraints:

[0116]

[0117]

[0118]

[0119]

[0120]

[0121]

[0122]

[0123]

[0124]

[0125]

[0126]

[0127] 2j-1≤i≤2j+1 (43)

[0128] 2j-1≤i'≤2j+1 (44)

[0129] Equation (26) is the first objective function, representing the shortest time required for the stacker crane to complete the order task in the intelligent automated warehouse. Equation (27) is the second objective function, representing the shortest running distance required for the stacker crane to complete the order task in the intelligent automated warehouse. Equation (29) is the third objective function, representing the optimal order storage stability in the intelligent automated warehouse. Equation (28) is the formula for calculating the centroid of the rack. Equation (31) is the fourth objective, representing the optimal turnover rate of goods in and out of the intelligent automated warehouse. Equation (30) represents the turnover rate of goods. Calculation formula; Constraints (32)-(35) represent the anti-collision strategy characterization formula of the stacker crane operation process, constraints (36)-(37) represent that all inbound and outbound orders must be satisfied, constraint (38) represents that the goods stored in the automated warehouse meet the outbound order requirements, constraints (39) and (40) represent that the goods in each inbound order meet the volume and weight limit requirements of the storage location, and constraints (41) and (42) represent that the maximum operating speed of the stacker crane needs to meet the minimum operating speed limit of the stacker crane.

[0130] A Multi-Objective Evolutionary Intelligent Scheduling Optimization Algorithm for Dual Stacker Cranes on the Same Track

[0131] Since the Storage Location Assignment Problem (SLAP) and the Crane Scheduling Problem (CSP) have been proven to be NP-hard, the SLAP-CSP ensemble optimization problem is also NP-hard. Furthermore, due to the enormous scale and complex constraints of the proposed dual-track stacker crane storage location assignment and scheduling optimization problem, exact algorithms struggle to find the optimal task combination within a finite time, while traditional optimization algorithms are prone to getting trapped in local optima and failing to search for the globally optimal task combination. Therefore, considering the complexity of the problem and the diversity of the objectives, a multi-objective heuristic evolutionary algorithm is designed to solve this problem.

[0132] The idea of ​​breadth-first search algorithm

[0133] like Figure 2 As shown, Breadth-First Search (BFS) is a basic graph search algorithm. Its principle is to start from the initial node, traverse all the neighbor nodes of the initial node layer by layer, and then visit the neighbors of the neighbor nodes, and so on, until the target node is found or the entire search space has been traversed.

[0134] Breadth-first search (BFS) provides a new approach for stacker cranes to perform outbound tasks, effectively reducing the search space and improving the algorithm's efficiency.

[0135] The NSGA-III algorithm is improved based on the breadth-search concept.

[0136] 1) Encoding

[0137] like Figure 3 As shown, in the joint optimization problem of CSP and SLAP, the decision variable ... Figure 3 As shown, the initial solution is first defined based on the stacker crane's inbound and outbound order tasks. The initial solution consists of four parts, namely the inbound order of stacker crane No. 1 (the number of orders is D). in1 Outbound orders (number of orders: D) out1 ), the inbound order for stacker crane No. 2 (order quantity: D) in2 Outbound orders (number of orders: D) out2 Therefore, the dimension of the initial solution is: D in1 +D out1 +D in2 +D out1 The circular boxes represent order indexes, and the square boxes represent chromosomes, which are code-based to the location index numbers. For example, in the inbound order of stacker crane No. 1, order No. 1 is stored in location 15, order No. 2 is stored in location 78, and order No. 34 is stored in location 62. In the outbound order of stacker crane No. 2, order No. 88 is outbound from location 58, order No. 89 is outbound from location 25, and order No. 97 is outbound from location 2.

[0138] For the joint optimization of CSP and SLAP under the established hybrid operation mode, it is necessary to calculate the number of tasks performed by the two stacker cranes in SC and DC modes. The number of tasks can be obtained according to the following calculation formula:

[0139] N DC =min{D in D out} (45)

[0140] N SC =max{D in D out}-N DC (46)

[0141] Where D in Let D be the number of inbound orders for any stacker crane. out Let N be the quantity of any stacker crane outbound order. DC and N SCThese represent the number of DC and SC tasks executed, respectively. To indicate the number of chromosomes that need to be reconstructed for the stacker crane to execute DC and SC tasks, the following is an example: Figure 4 As shown, the company has 12 orders, including 4 outbound orders and 8 inbound orders. According to the task calculation formula above, this order has a total of 4 DC tasks and 4 SC tasks. The first DC task is: the stacker crane transports the No. 1 inbound order from the I / O station to the No. 15 storage location and stores it. Then it goes to the No. 74 storage location to transport the goods loaded in the No. 74 storage location to the I / O station to meet the outbound demand of the No. 35 outbound order. When executing the SC task, the stacker crane only performs a single outbound / inbound task. The No. 34 inbound order is an SC task. The stacker crane transports the No. 34 order from the I / O station to the No. 25 storage location and stores it, and then returns to the I / O station. During this period, no other tasks are executed.

[0142] To further describe the operation of the stacker crane, this invention introduces an S-shaped motion curve to describe the operation of the stacker crane, and establishes the following relationship to describe the relationship between the stacker crane's operating distance and operating time:

[0143]

[0144] Where T f J and a are the movement times of the stacker crane. max and V max Describe the range of values ​​for jerk, acceleration, and velocity during the stacker crane's movement process, S. ref The distance traveled by the stacker crane during its operation.

[0145] 2) Initialize the population

[0146] Population initialization has a significant impact on the solution performance of NSGA-III. A good initialization strategy can not only effectively locate the search region of the initial population, thereby improving the algorithm's solution efficiency, but also effectively increase population diversity, which is beneficial for later iterations and updates. This invention addresses the problem of the enormous solution space of the established stacker crane location allocation and scheduling optimization model. Based on the breadth-first search approach, it designs an initialization population dimensionality reduction method. During population initialization, a CSP initialization scheme is first generated using a random initialization strategy. Then, the previous order is used as the initial node for breadth-first search to obtain the initial solution for the SLAP task.

[0147] 3) Adaptive crossover and mutation operators

[0148] Crossover is essentially the process by which two parent chromosomes exchange information to produce offspring. To avoid infeasible solutions, we implement crossover operations between chromosomes by randomly selecting the crossover length. Mutation is the process of randomly selecting chromosome nodes on the same chromosome to mutate local information. However, both crossover and mutation operations can result in chromosome solution structures that do not meet the requirements of the established model. Therefore, after each crossover operation, we need to perform a breadth-first search to update the solution of the SLAP task based on the new chromosome in order to eliminate conflicts in the chromosome solution set.

[0149] The traditional NSGA-III crossover operator has relatively poor global search capabilities, and the polynomial mutation operator contains subjective parameters, resulting in insufficient objective solution capabilities. To overcome these shortcomings, this invention designs crossover and mutation operators with adaptive dynamic adjustment. When the fitness function values ​​of individuals in the population tend to a local optimum, the crossover and mutation probabilities increase. When the fitness function values ​​of individuals in the population are relatively dispersed, the crossover and mutation probabilities decrease. At the same time, individuals with relatively poor fitness levels have relatively high crossover and mutation probabilities, while individuals with relatively low fitness levels have relatively low crossover and mutation probabilities.

[0150]

[0151]

[0152] Where f 11 ,f 12 It is the fitness function value of two individuals in the first target random population, f 21 ,f 22 For the second objective, select the fitness function values ​​f of two individuals in the population. 31 ,f 32 For the third objective, select the fitness function values ​​f of two individuals in the population. 41 ,f 42 For the fourth objective, select the fitness function values ​​f of two individuals in the population. 1avg ,f 2avg ,f 3avg ,f 4avg f is the average fitness value of the current population. 1max ,f 2max ,f 3max ,f 4max f is the maximum fitness value of the current population. 1min ,f 2min ,f 3min ,f 4min It is the minimum fitness value of the current population, f1' = max(f 11 ,f 12 ), f2' = max(f21 ,f 22 ), f3' = max(f 31 ,f 32 ), f4' = max(f 41 ,f 42 ).

[0153] 4) Reference point-based selection operator

[0154] Traditional NAGS-III combines parent and offspring generations into a 2N-sized population using an elitist strategy, and selects the top N individuals with higher dominance levels to form a new population. This invention improves the algorithm by introducing a reference point-based selection operator:

[0155]

[0156]

[0157] In the NSGA-III algorithm, the ideal point is defined as the minimum value of each objective function. In the above formula, ω j Is target f j The weighting coefficient in the axial direction, ρ is the weighting coefficient, and its value is 10. -4 a i It calculates the distance between the line connecting two extreme points and the target axis. The method for constructing the reference point is the same as that of the traditional NSGA-III. The algorithm flow is as follows: Figure 5 As shown in Table 1 below, the effects of several algorithms are compared.

[0158] Table 1 Comparison of Technical Effects

[0159]

[0160] The invention has broad applications across various aspects of the AS / RS field, particularly in several key areas of the modern warehousing and logistics industry. Firstly, it can be used in modern warehousing and sorting centers, enabling them to more efficiently handle increasing inbound and outbound demands, accelerate storage and retrieval operations, and improve the accuracy of inventory management, thereby providing superior customer service. Secondly, e-commerce logistics centers can also benefit from this invention to increase order processing speed, reduce operating costs, and achieve efficient order fulfillment, thus meeting the rapid growth and competition in the e-commerce sector. Raw material inventory management and parts supply in manufacturing are also application areas; this invention helps manufacturers reduce production downtime, improve production efficiency, and ensure production line stability. Furthermore, the high demands of the food and pharmaceutical industries for logistics and inventory management make this another important area; this invention helps improve storage efficiency, reduce error rates, ensure the quality and safety of food and pharmaceuticals, and meet timely delivery and regulatory requirements. Finally, supply chain management can also benefit from this invention by optimizing logistics and warehousing operations, reducing inventory costs, and improving the sustainability and efficiency of the supply chain to meet the needs of modern business. In summary, the multi-domain applications of this invention aim to improve the performance and efficiency of AS / RS systems, meet ever-changing logistics needs and environmental challenges, and provide efficient logistics and warehousing solutions for multiple industries.

[0161] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0162] It should be understood that the above detailed description of the technical solutions of the present invention with reference to preferred embodiments is illustrative and not restrictive. Those skilled in the art can modify the technical solutions described in the embodiments or make equivalent substitutions for some of the technical features based on reading this specification; however, these modifications or substitutions do not cause the essence of the corresponding technical solutions to depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent scheduling optimization strategy for a same-track double stacker crane in an intelligent stereoscopic warehouse, characterized in that: The optimization strategies include: Establish an optimization model for the allocation and scheduling of storage locations for dual stacker cranes on the same track, based on dynamic storage and composite operation modes; Establish a joint optimization model for warehouse location allocation and stacker crane scheduling; The above model is solved using a multi-objective evolution-based intelligent scheduling optimization algorithm for dual stacker cranes on the same track. The aforementioned optimization model for the allocation and scheduling of storage locations for dual stacker cranes on the same track makes the following assumptions: 1) The left and right stacker cranes in the same lane have the same attribute parameters; 2) All storage locations / shelves within the automated warehouse are of standard dimensions; 3) The inbound and outbound order information of the I / O units on both sides of the same aisle is known, and the existing storage location information within the automated warehouse is also known; The aforementioned optimization model for the allocation and scheduling of storage locations for dual stacker cranes on the same track includes: Calculation of stacker crane running time and distance; When a stacker crane performs inbound and outbound tasks, its travel distance and time directly affect the operating efficiency of the AS / RS system. Therefore, it is necessary to accurately calculate the operating performance of the stacker crane performing inbound and outbound tasks. The calculation process is as follows: The operation of the stacker crane in the intelligent automated warehouse consists of the horizontal movement of the slide rail and the longitudinal movement of the hoist. Therefore, the movement of the stacker crane is calculated from two directions. At the same time, since the two stacker cranes of the same track double stacker crane belong to the I / O units at both ends of the aisle, the running distance of the stacker cranes on the left and right sides is calculated separately. In mixed-task operations, the operation modes of SC and DC are not the same, and calculations need to be performed separately for different operation models. The aforementioned joint optimization model for cargo location allocation and stacker crane scheduling; Based on the operational requirements of the AS / RS system, the optimization of storage location allocation and stacker cranes must be ensured to minimize the time and travel distance for stacker cranes to perform inbound and outbound tasks. At the same time, the overall stability of the shelving and the turnover rate of goods in the warehouse also need to be considered on the impact of the automated warehouse operation efficiency. Therefore, an objective function is established to describe the joint optimization problem of intelligent automated warehouse location allocation and stacker crane; The intelligent scheduling optimization algorithm for dual stacker cranes on the same track based on multi-objective evolution adopts the NSGA-III algorithm improved based on the breadth-search idea, and specifically improves the following four parts: coding; Initialize the population; Adaptive crossover and mutation operators; Reference point-based selection operator; In the joint optimization problem of stacker crane scheduling and storage location allocation, the decision variables consist of two parts: storage location allocation for inbound orders and stacker crane scheduling for outbound orders; the solution of the model is defined using integer encoding. The initialization population described herein is based on a breadth-first search approach, which designs a method for dimensionality reduction in the initialization population. When initializing the population, a stacker crane scheduling problem initialization scheme is first generated using a random initialization strategy; then, the previous order is used as the initial node for breadth-first search to obtain the initial solution for the cargo location allocation problem. The adaptive crossover and mutation operators increase the crossover and mutation probabilities when the fitness function values ​​of individuals in the population tend to a local optimum. When the fitness function values ​​of individuals in a population are relatively dispersed, the probabilities of crossover and mutation are relatively reduced. Meanwhile, individuals with relatively poor fitness levels have relatively high crossover and mutation probabilities, while individuals with relatively good fitness levels have relatively low crossover and mutation probabilities. The aforementioned reference point-based selection operator improves the algorithm by introducing a reference point-based selection operator.

Images