Multi-target scheduling optimization method for intelligent warehouse logistics system

Abstract

The invention provides a multi-target scheduling optimization method for a smart warehouse logistics system, and belongs to the technical field of smart warehouse logistics distribution. An intelligent decision-making strategy is introduced, firstly, according to an attribute relationship between intelligent warehouse logistics entities, when there is an order, statistics is performed on the order, and related information is counted into a database; a multi-objective optimization model for storage environment shelf stability, operation cargo distribution uniformity, in-out warehouse carrying efficiency and the shortest transportation operation path is established, a good point set is adopted to uniformly initialize a population, a position disturbance factor is introduced to update and adjust the particle position of a particle swarm algorithm, and multi-objective optimization solution is performed. Determining an optimal cargo position storage decision under the current condition; an enhanced A * algorithm is put forward to plan a general path of transportation vehicle-mounted order operation from a storage place to a position below a goods shelf, an output path scheme is taken as an initial path planning solution record of an improved particle swarm algorithm, and accurate optimization is further realized by the improved particle swarm algorithm, so that the lowest consumption is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of smart warehousing logistics and distribution, and in particular relates to a multi-objective scheduling optimization method for a smart warehousing logistics system. Background Art

[0002] In recent years, China's warehousing industry has faced multiple challenges due to increasing cost pressure, slowing growth in warehousing demand, and uncertain social consumption factors. The traditional low-value-added service model that relies on land and labor costs is no longer sustainable. Therefore, relying on digitalization and intelligence can effectively improve resource allocation efficiency and operational quality. In order to achieve visual monitoring of warehousing physical resources and intelligent allocation decisions for cargo space, intelligent scheduling has become an important way to achieve smart warehousing logistics.

[0003] To meet these challenges, the smart warehousing and logistics system has improved the degree of automation and reduced labor costs by virtue of functions such as automatic handling, intelligent monitoring, operation process management and multi-equipment coordinated scheduling; its flexible warehouse layout, high space utilization and high flexibility of four-way vehicles also reduce land costs, making it an effective solution for the construction of smart warehousing and logistics systems.

[0004] In the smart warehousing and logistics system, the optimization of cargo location decisions is achieved by comprehensively considering factors such as material weight, equipment scheduling frequency, scheduling path, operation balance and material type. This can dynamically allocate cargo locations and improve the efficiency of warehousing operations to cope with the growing order demand. Therefore, the optimization of cargo location decisions has become the core issue in improving the efficiency of warehousing and logistics systems.

[0005] Traditional research mainly focuses on optimizing the stability of storage shelves and finding the shortest scheduling path, which has improved the efficiency of logistics distribution to a certain extent. However, there are still some limitations:

[0006] 1. Traditional research often only considers the Manhattan path between the goods and the target location, ignoring the actual path of obstacles, inventory or other equipment, resulting in the real-time and quality of cargo allocation failing to meet actual needs.

[0007] 2. Traditional research does not take energy consumption and job distribution constraints into account, and it is necessary to redesign the path planning algorithm to consider dynamic constraints such as equipment and inventory. Therefore, if the actual cargo location decision needs are to be better met, the goals of low energy consumption and balanced distribution of jobs on each shelf layer must also be considered, which increases the complexity of constraints and scheduling resources, making intelligent cargo location allocation more difficult.

[0008] 3. Functions such as cargo location decision, equipment monitoring, and equipment scheduling can be independently turned into services, and can be combined with the Web to form a visual page. Current research still has some difficulties in the decoupling of various functional modules in the warehousing system, the development of visual applications, and their use in actual application environments. Summary of the invention

[0009] In view of the above problems, the present invention provides a multi-objective scheduling optimization method for a smart warehousing logistics system, which includes the following processes:

[0010] S1, build an entity information model based on the actual warehousing environment; determine the relationship model of each factor, generate the ER diagram in the database table, that is, an entity association diagram including at least six entities: warehouse, order, operation, cargo location, inventory, and pallet. When there is an order, count the order and store the relevant information in the database table;

[0011] Set three-dimensional coordinate information (x, y, z) to mark the location of goods and shelf channel information, and provide data information for multi-objective cargo scheduling optimization;

[0012] S2, establish a model for multi-objective optimization problems such as shelf stability in the storage environment, uniform distribution of cargo, efficiency of inbound and outbound handling, and shortest transportation operation path, which conforms to the operating procedures of the actual storage logistics system;

[0013] S3, a meta-heuristic algorithm is designed to solve the multi-objective optimization problem of cargo location. The population is initialized uniformly using the good point set, the parameters in the speed update formula are adjusted using nonlinear transformation, and the position disturbance factor is introduced to adjust the particle position update of the particle swarm algorithm. Multi-objective optimization is performed to determine the optimal cargo location storage decision under the current situation.

[0014] Preferably, the road surface of the smart warehousing and logistics environment is modeled, and an enhanced A* algorithm is designed to plan the path for goods leaving the warehouse to obtain a rough path; the rough path result is used as the initial path plan, and an improved particle swarm optimization algorithm is used to optimize the path planning problem to determine the final route trajectory.

[0015] Preferably, the complete warehouse management process is first determined in S1. When there are orders for ordering and replenishment, the information is uniformly entered through order management, and different operations are generated according to the time and type of the order. According to the operation type, the storage and access location of the goods is determined, and the location information is transmitted to the warehouse transport vehicle for operation scheduling to realize the storage and access operations of the warehouse goods. After determining the operation process of warehousing and logistics, an entity relationship model of the smart warehousing and logistics system is established, and a relational database is established. It is mapped with various types in the warehousing and logistics system to complete the storage and access of information in the warehousing and logistics system. Among them, the ER diagram contains six entities: warehouse, order, operation, cargo location, inventory, and pallet; their respective related connection attributes are set as follows:

[0016] Warehouse: warehouse number, warehouse name;

[0017] Order: order number, order date, logistics information, customer number;

[0018] Job: job number, job details, job item, job item number;

[0019] Cargo location: cargo location number, cargo location, number of layers, layer name, layer number, layer specification;

[0020] Inventory: inventory number, shelf number, pallet number;

[0021] Pallet: Pallet number, maximum load, pallet type, pallet specifications.

[0022] Preferably, the shelf stability in S2 is to minimize the product of the weight of the goods and the number of layers where the goods are located, and the optimization objective function F1 model is:

[0023]

[0024] Where M xyz represents the weight of the stored and retrieved goods, h represents the height of the goods, and k, p, and q correspond to the total number of shelf columns, rows, and layers respectively;

[0025] The uniform distribution of the operation cargo considers the uniform distribution of the operation cargo, and evenly distributes the existing cargo and the operations to be completed to form the operation uniform distribution objective function F2 model:

[0026]

[0027] Among them, the smaller the value of F2, the more uniform and reasonable the operation distribution; q>1 means that when performing the operation, only the case where the number of shelves is more than two layers is considered; z Indicates the number of cargo spaces occupied by goods on the shelf at layer z. Indicates the average number of cargo spaces occupied by cargo on each floor;

[0028] The in-and-out handling efficiency takes into account the movement time of operators in the rows, rows, and layers of goods entering the warehouse and the corresponding in-and-out frequency, forming the in-and-out handling efficiency objective function F3 model:

[0029]

[0030] Among them, M, L, and H correspond to the length, width, and height information of the cargo location respectively; v x ,v y ,v z Indicates the movement speed of the operator, S xyz Indicates the frequency of goods entering and leaving the warehouse;

[0031] The transportation operation path is the shortest, and the design takes the minimum value of the matching degree between the actual cargo position and the scheduling decision as the shortest transportation operation path; specifically, determine the location of the cargo on the platform (x1, y1, z1), the cargo placement coordinates (x2, y2, z2), the cargo elevator coordinates (x m ,y m ) constitutes the shortest objective function F4 model of transportation operation path:

[0032]

[0033] When the position of the storage platform and the cargo placement position are on the same floor, the value is dis1. If they are on different floors, the shortest path distance is dis1+dis2+dis3. dis2 represents the floor difference between the storage platform and the cargo placement position, and dis3 represents the row difference between the elevator and the cargo position.

[0034] The above-mentioned storage environment shelf stability, uniform distribution of operating goods, in-and-out warehouse handling efficiency, and shortest transportation operation path optimization model together constitute a multi-objective optimization design problem.

[0035] Preferably, the specific process of S3 is:

[0036] S31, use the best point set to evenly distribute the population in the search space, set the particle swarm population, the maximum number of iterations and related parameters:

[0037]

[0038] Among them, X i represents the position of the i-th particle in the initial population; ub and lb correspond to the upper and lower limits of the search space respectively; represents the best point taken; u satisfies 1≤u≤D;

[0039] S32, through the data transmitted by the smart warehouse logistics system database, the feasible domain of shelf locations is updated in real time, and the number of decision-making operations is determined as the spatial dimension;

[0040] S33, calculate the position of each particle and determine gb est , gb est represents the global optimal position;

[0041] S34, updating the velocity and position information of the particle respectively according to the following formula (6) and the formula (7) with the position disturbance operator added, and the formulas (8) to (11);

[0042] The update formulas for velocity v and position x are as follows:

[0043]

[0044] Where t represents the number of iterations of the calculation; i represents the particle number i=1,...,n, n represents the particle population size; ω represents the inertia weight; r1, r2 correspond to random numbers between (0,1) respectively; c1, c2 represent acceleration constants, p best represents the individual optimal position of the particle, that is, the optimal cargo position value obtained by the particle after one iteration; g best represents the global optimal position, that is, the record with the best optimal cargo position value among all particles after one iteration;

[0045] The three parameters of inertia weight ω and acceleration constants c1 and c2 are transformed nonlinearly. The nonlinear transformation is as follows:

[0046]

[0047] Among them, the value range of ω is [ω min ,ω max ] decreases according to the law; the acceleration constant c1 is [c 11 ,c 12 ]; the acceleration constant c2 changes between [c 21 ,c 22 ]; i represents the current iteration number, i max Indicates the maximum number of iterations of the algorithm;

[0048] The perturbation operator added in combination with formula (7) is as follows:

[0049]

[0050] Where u is the u-th dimension component in the D-dimensional search space; x iu The value range is [SU u ,SL u ]; r is a random number randomly generated between [0,1], b is the disturbance degree factor, t represents the current iteration number, T represents the total iteration number, and the random value of l is 0 or 1;

[0051] S35, calculating the function value according to the optimization objectives of shelf stability in the storage environment, uniform distribution of the cargo, efficiency of inbound and outbound handling, and shortest transportation operation path;

[0052] S36, update the particle's p best and g best ;

[0053] S37, the number of iterations is increased by 1, and it is determined whether the process has reached the set maximum number of iterations. If not, it returns to S34; if it has, it outputs the optimal solution for the multi-objective cargo location decision optimization.

[0054] Preferably, the road surface modeling of the smart warehousing logistics environment is specifically as follows:

[0055] The actual environment is converted into a grid map to provide an environmental basis for path planning; the grid map is to decompose the actual environment into multiple grid units of the same size L×W, each unit M lw Represented by state variables 0-1; forming a two-dimensional map information:

[0056]

[0057] Among them, l and w represent the row and column where the current grid is located, and their respective maximum values ​​correspond to L and W; when M lw When the value is 1, it means there is an obstacle, and when it is 0, it means it is passable;

[0058] At the same time, the path planning objective function is designed; the path planning goal of the transport vehicle is to find a shortest path from the starting point (a1, b1) of the cargo to the end point (a2, b2) under the shelf without touching obstacles; therefore, the objective function of the path planning problem is designed as follows:

[0059]

[0060] in, represents the penalty operator, which is used to exclude paths that pass through the obstacle range, f o It represents the sum of the Euclidean distances of all points in the path trajectory, which is expressed as follows:

[0061]

[0062] Set f obs Represents the variable that determines the obstacle level of the current planned path, and its initial value is set to 0:

[0063]

[0064] Among them, (a s ,bs ) represents the coordinate value of the sth path point on the planned path; (xobs k ,yobs k ) corresponds to the coordinate value of the kth obstacle; n represents the total number of points on the planned path, nobs represents the total number of obstacles, i.e. other shelves and obstacles placed on the path in the warehouse logistics environment; r k Represents the obstacle radius of an abstract obstacle.

[0065] Preferably, in the A* algorithm in smart warehousing logistics, the location of the cargo is the starting point of the transport vehicle, and the cargo unloading location is the end point of the transport vehicle for path planning. The evaluation function of the A* algorithm for path planning from the starting point to the target end point is as follows:

[0066] f(n)=h(n)+g(n) (17)

[0067] Where n represents the current node position, h(n) represents the estimated distance from the current position to the end point, and g(n) represents the actual driving distance from the starting point to the current position; h(n) is expressed as follows:

[0068]

[0069] Among them, (a1, b1) and (a2, b2) correspond to the coordinates of the current node and the target end point respectively;

[0070]

[0071] Among them, g(n) is the previous node (a1, b1) according to the current node (a1, b1). s ,b s ) is obtained by accumulating the shortest path distance;

[0072] By calculating the location cost between the current node and the target end point, the shortest path requirement can be gradually met;

[0073] To improve the calculation effect of path planning, add a nonlinear factor P:

[0074]

[0075] Among them, length represents the length of the grid, and width represents the width of the grid; the nonlinear factor is added to the evaluation function:

[0076] f(n)=h(n)+P*g(n) (21)

[0077] If the current node is far from the target end point, the P value is larger to enhance the search efficiency. On the contrary, if the current node is close to the target end point, the P value is smaller to enhance the search accuracy of the algorithm.

[0078] Preferably, the improved particle swarm optimization algorithm is used to optimize the path planning problem to determine the final route trajectory, and the approximate path record Gbest(n) obtained by the enhanced A* algorithm is used as the initial optimal solution record of the improved particle swarm algorithm;

[0079] The shortest distance from the starting point to the end point is determined according to the evaluation function expressed in formula (13), and the node at the next position is selected according to the estimated value as the fitness function of the improved particle swarm algorithm.

[0080] Compared with the prior art, the present invention has the following beneficial effects:

[0081] The method of the present invention introduces an intelligent decision-making strategy. First, according to the attribute relationship between smart warehousing and logistics entities, an entity association diagram of six entities, namely, warehouse, order, operation, cargo location, inventory, and pallet, is established. When there is an order, the order is counted and the relevant information is counted in a database. A multi-objective optimization model for the stability of the shelf in the storage environment, the uniform distribution of the operation cargo, the efficiency of in-and-out handling, and the shortest transportation operation path is established. The population is uniformly initialized using a good point set, and a position disturbance factor is introduced to update and adjust the particle position of the particle swarm algorithm, and a multi-objective optimization solution is performed to determine the optimal cargo location storage decision under the current situation. An enhanced A* algorithm is proposed to plan the approximate path of the transport vehicle order operation from the storage place to the bottom of the shelf, and the output path plan is recorded as the initial path planning solution of the improved particle swarm algorithm, and the improved particle swarm algorithm is used for further precise optimization to form a double-layer optimization, thereby achieving the lowest cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 It is a schematic diagram of the overall process of the multi-objective scheduling optimization method of the intelligent warehousing logistics system of the present invention.

[0083] Figure 2 This is the operation flow chart of the multi-intelligent warehousing logistics system of the present invention.

[0084] Figure 3 This is the order and job ER diagram of the present invention.

[0085] Figure 4 This is a multi-objective cargo location optimization flow chart for the intelligent warehousing logistics system of the present invention.

[0086] Figure 5 The optimal path flow chart for optimizing the intelligent warehousing logistics system of the present invention. DETAILED DESCRIPTION

[0087] The present invention is further described below in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0088] The present invention provides a multi-objective scheduling optimization method for a smart warehousing logistics system. The overall process is as follows: Figure 1 As shown:

[0089] S1, smart warehousing and logistics system cargo location decision entity information modeling. Combined with the actual warehousing environment, the entity information model is constructed, and the relationship model of each factor is determined to generate the ER diagram in the database table, including warehouse, cargo location, inventory, order and other related factors. The designed database table is built with the object class to establish a corresponding relationship, and the warehousing and logistics related database is obtained. The three-dimensional coordinate information (x, y, z) is set to mark the cargo location and shelf channel information, providing data information for multi-objective cargo scheduling optimization.

[0090] S2, multi-objective cargo location decision optimization design modeling for smart warehousing and logistics systems. Consider the relevant mathematical modeling of warehousing operations in warehousing and logistics systems. Including the establishment of multi-objective optimization models for storage environment shelf stability, uniform distribution of operating goods, in-and-out handling efficiency, and shortest transportation operation path. Fit the operating procedures of the actual warehousing and logistics system.

[0091] S3, design a meta-heuristic algorithm to solve the multi-objective optimization problem of cargo location. Use the improved particle swarm algorithm for optimization and solution. The particle swarm population represents the solution to the multi-objective optimization problem. The population generation of the original particle swarm algorithm is optimized by the good point set, and the position disturbance factor is introduced to adjust the particle position update. Apply it to solve the multi-objective cargo location optimization problem and obtain the optimal cargo location information.

[0092] S4, the path planning algorithm design for the transport vehicle to transport goods to the bottom of the shelf. After determining the storage decision of the goods location for the current order, the transport vehicle is used to transport the goods to the cargo location. An enhanced A* algorithm is proposed to plan the path when the goods are out of the warehouse, and a rough path is obtained; the rough path result is used as the initial path plan, and the improved particle swarm optimization algorithm is used to perform double-layer optimization on the path planning problem to determine the final route trajectory.

[0093] 1. Entity information modeling for cargo location decision making in smart warehousing and logistics system

[0094] (1) Analysis of warehouse logistics operation process design. This part consists of three parts: order management, cargo location decision and operation execution.

[0095] (2) First, determine the complete warehouse management process. When there are orders for ordering and replenishment, the information is uniformly entered through the order management, and different operations are generated according to the time and type of the order. According to the operation type, the storage and access location of the goods is determined, and the location information is transmitted to the warehouse transport vehicle for operation scheduling to realize the storage and access operations of the warehoused goods. The flowchart of the operation process of the intelligent warehouse logistics system is as follows Figure 2 shown.

[0096] (3) Multi-attribute relational database mapping modeling. After determining the operation process of warehousing and logistics, the entity relationship model of the intelligent warehousing and logistics system is established, and a relational database is established to map the relationship between it and various types of warehousing and logistics systems to complete the storage and access of information in the warehousing and logistics system.

[0097] The present invention constructs relational database modeling, that is, designs a warehousing logistics entity relationship model diagram (ER diagram) to describe the information type stored in the database.

[0098] The ER diagram contains six entities: warehouse, order, operation, cargo location, inventory, and pallet. Their related connection attributes are set as follows:

[0099] (1) Warehouse: warehouse number and warehouse name;

[0100] (2) Order: order number, order date, logistics information, customer number;

[0101] (3) Job: job number, job details, job item, job item number;

[0102] (4) Cargo location: cargo location number, cargo location, number of layers, layer name, layer number, layer specifications;

[0103] (5) Inventory: inventory number, shelf number, pallet number;

[0104] (6) Pallet: pallet number, maximum load, pallet type, pallet specifications.

[0105] There are related connections between different entities. Figure 3 Give the associative entity relationship diagram of orders and jobs.

[0106] According to the attributes of each entity in the designed warehousing logistics system, an ER diagram is constructed to complete the design and modeling of the database.

[0107] (4) Set the three-dimensional coordinate information (x, y, z) to mark the cargo location and shelf channel information. Simulate the actual warehouse shelf cargo placement, where x represents the number of shelves, y represents the number of shelves, and z represents the number of operating floors of the cargo elevator. According to the storage conditions of the shelves corresponding to the set three-dimensional coordinate information, a three-dimensional environment model is obtained.

[0108] Smart warehousing and logistics system information mapping. When a new order is generated in the smart warehousing and logistics system, the cargo information is set based on the current cargo situation and the attributes determined by the ER diagram, and added to the warehousing and logistics system database to complete order management.

[0109] 2. Multi-objective cargo location decision optimization design modeling

[0110] After the order information is managed through the database, it is necessary to make multi-objective optimization storage and retrieval decisions based on the current warehouse logistics storage situation, generate operation instructions, and send them to the transport vehicle for operation and transportation to complete the storage and retrieval of goods.

[0111] The multi-objective cargo location decision optimization design meets the optimization goal and converts the actual situation into a mathematical expression, including the establishment of a multi-objective optimization model design for the stability of the storage environment shelves, the uniform distribution of the operating goods, the efficiency of in-and-out handling, and the shortest transportation operation path.

[0112] (1) Shelf stability. The above goal is to consider the safety of smart warehousing logistics scheduling. Reasonable cargo placement is to place the overall center of gravity of the cargo close to the bottom area of ​​the shelf to reduce the impact of the high center of gravity caused by placing the cargo at a high position. That is, the product of the cargo weight and the number of layers where the cargo is located is minimized. Therefore, the present invention designs a shelf stability optimization objective function F1 model from a safety perspective:

[0113]

[0114] Where M xyz represents the weight of the stored and retrieved goods, h represents the height of the goods location, and k, p, q correspond to the total number of shelf columns, rows, and layers respectively.

[0115] (2) Even distribution of cargo. The above objective is based on the current storage situation of cargo in the warehouse. If the cargo is too densely stored, it will cause congestion and reduce the efficiency of the operation. Therefore, the present invention considers the even distribution of cargo, evenly distributes the existing cargo and the operations to be completed, and forms the objective function F2 model of even distribution of operations:

[0116]

[0117] Among them, the smaller the value of F2, the more uniform and reasonable the operation distribution is; q>1 means that when performing the operation, only the case where the number of shelves is more than two layers is considered. z Indicates the number of cargo spaces occupied by goods on the shelf at layer z. Indicates the average number of cargo spaces occupied by goods on each floor.

[0118] (3) In-and-out handling efficiency. The above target is for the economic requirements of warehousing and logistics. When performing in-and-out operations, the principle of storing goods nearby is the primary consideration. The handling efficiency is determined based on the frequency of goods in and out of the warehouse and the location of goods access. The present invention considers the movement time of operators in the rows, rows, and layers of goods entering the warehouse and the corresponding frequency of in-and-out, and forms the in-and-out handling efficiency target function F3 model:

[0119]

[0120] Among them, M, L, and H correspond to the length, width, and height information of the cargo location respectively; v x ,v y ,v z Indicates the movement speed of the operator, S xyz Indicates the frequency of goods entering and leaving the warehouse.

[0121] (4) The transportation operation path is the shortest. The above goal is to consider the distance from the cargo to the target location. The storage location of the cargo during the operation should be kept close to the target location as much as possible. Therefore, the design of the present invention takes the minimum value of the matching degree between the actual cargo location and the scheduling decision as the shortest transportation operation path. Specifically, the location of the cargo on the warehouse (x1, y1, z1), the cargo placement coordinates (x2, y2, z2), and the cargo elevator coordinates (x m ,y m ) constitutes the shortest objective function F4 model of transportation operation path:

[0122]

[0123] When the storage platform and the cargo placement are on the same floor, the value is dis1. If they are on different floors, the shortest path distance is dis1+dis2+dis3. dis2 represents the floor difference between the storage platform and the cargo placement, and dis3 represents the row difference between the elevator and the cargo placement.

[0124] The above-mentioned storage environment shelf stability, uniform distribution of operating goods, in-and-out warehouse handling efficiency, and shortest transportation operation path optimization model together constitute a multi-objective optimization design problem.

[0125] 3. Improved particle swarm algorithm to solve multi-objective decision optimization of cargo location in smart warehousing logistics system

[0126] The smart warehousing logistics system is very complex in the actual environment. According to the designed smart warehousing cargo location optimization goal, the design obtains multiple optimization goals. The present invention uses an improved particle swarm algorithm to solve the multi-objective optimization problem, such as Figure 4 As shown, the improved particle swarm algorithm is used to solve the path planning algorithm design for transporting goods from the transport vehicle to the position under the shelf, and double-layer optimization is performed to determine the route trajectory to save computing resources.

[0127] The multi-objective particle swarm algorithm is developed based on the original particle swarm algorithm for solving single-objective optimization problems. The algorithm generates a set of Pareto optimal solutions within the constraints in each iteration and optimizes through mutual learning among particles.

[0128] (1) Original particle swarm algorithm. The particle swarm algorithm is a set of particle populations randomly generated in a D-dimensional space. Each particle corresponds to a set of solutions to the multi-objective optimization problem, that is, a feasible solution to the intelligent warehousing logistics cargo location decision. With the iterative update of the calculation, the particles adjust their own position and speed and continuously update towards the optimal solution position, that is, the optimal cargo position. The speed v and position x update formulas of the original particle swarm algorithm are as follows:

[0129]

[0130] Where t represents the number of iterations of the calculation; i represents the particle number i=1,...,n, n represents the particle population size; ω represents the inertia weight; r1, r2 correspond to random numbers between (0,1) respectively; c1, c2 represent acceleration constants, p best represents the individual optimal position of the particle, that is, the optimal cargo position value obtained by the particle after one iteration; g best Represents the global optimal position, that is, the record with the best optimal cargo position value among all particles after one iteration.

[0131] (2) Improved particle swarm algorithm.

[0132] The good point set evenly distributes the population. The population initialization of the original particle swarm algorithm is randomly generated in the D-dimensional search space. This makes the population distribution uneven. Therefore, in order to improve the global search ability of the particle swarm algorithm, the good point set is used to evenly distribute the population in the search space:

[0133]

[0134] Among them, X i represents the position of the i-th particle in the initial population; ub and lb correspond to the upper and lower limits of the search space respectively; represents the optimal point taken; u satisfies 1≤u≤D.

[0135] (3) Particle velocity parameter transformation. According to formula (5), it can be seen that in the particle velocity update formula, the inertia weight ω and the acceleration constants c1 and c2 have an important influence on the algorithm performance.

[0136] A larger value of ω increases the global search performance of the algorithm, and vice versa, it enhances the local search capability of the algorithm.

[0137] For c1 and c2, when c1 takes a small value and c2 takes a large value, the algorithm has good local search capability. Otherwise, the algorithm has better global search capability.

[0138] In order to improve the influence of inertia weight and acceleration constant on algorithm performance, the present invention proposes to perform nonlinear transformation on the three parameters. The nonlinear transformation is as follows:

[0139]

[0140] Among them, the value range of ω is [ω min ,ω max ] decreases according to the law; the acceleration constant c1 is [c 11 ,c 12 ]; the acceleration constant c2 changes between [c 21 ,c 22 ]; i represents the current iteration number, i max Indicates the maximum number of iterations of the algorithm.

[0141] (4) Perturbation operator of particle position. It can be understood that the particle position is updated in a fixed direction in each iteration, which makes the algorithm easy to fall into the local optimal solution. Adding a perturbation operator can enhance the local search ability of the algorithm. The perturbation operator added in combination with formula (6) is as follows:

[0142]

[0143] Where u is the u-th dimension component in the D-dimensional search space; x iu The value range is [SU u ,SL u ]. r is a random number randomly generated between [0,1], b is the disturbance degree factor, t represents the current iteration number, T represents the total iteration number, and l is a random value of 0 or 1.

[0144] By adding a position perturbation operator into formula (6), the local search capability of the algorithm is enhanced.

[0145] (5) Improved particle swarm algorithm to solve the multi-objective cargo location decision of smart warehousing logistics system. The process of optimizing the above improved particle swarm algorithm to solve the smart warehousing cargo location decision is as follows:

[0146] S1, according to the best point set of formula (7), set the particle swarm population, maximum number of iterations and related parameters;

[0147] S2, through the data transmitted by the smart warehousing logistics system database, updates the feasible domain of shelf locations in real time and determines the number of decision-making operations as the spatial dimension;

[0148] S3, calculate the position of each particle and determine g best ;

[0149] S4, updating the velocity and position information of the particle according to formula (5), formula (6) with the position disturbance operator added, and formulas (8) to (11);

[0150] S5, calculates the function value according to the optimization objectives of shelf stability in the storage environment, uniform distribution of cargo, efficiency of inbound and outbound handling, and shortest transportation operation path;

[0151] S6, update the particle’s p best and g best ;

[0152] S7, the number of iterations is increased by 1, and it is determined whether the process has reached the set maximum number of iterations. If not, it returns to S4; if it has, it outputs the optimal solution for the multi-objective cargo location decision optimization.

[0153] Through the above steps, the optimal cargo storage and access location for real-time order generation operations is obtained.

[0154] 4. Double-layer optimization determines the route trajectory of the transport vehicle transporting goods to the bottom of the shelf

[0155] The overall process is as follows Figure 5 As shown in the figure. First, complete the road modeling of the smart warehousing logistics environment. Convert the actual environment into a grid map to provide an environmental basis for path planning. The grid map is to decompose the actual environment into multiple grid units of the same size L×W, each unit M lw Indicated by state variable 0-1.

[0156] Constitute a two-dimensional map information.

[0157]

[0158] Among them, l and w represent the row and column where the current grid is located, and their respective maximum values ​​correspond to L and W. lw When the value is 1, it means there is an obstacle, and when it is 0, it means it is passable.

[0159] Secondly, design the path planning objective function. The path planning goal of the transport vehicle is to find the shortest path from the starting point (a1, b1) of the cargo to the end point (a2, b2) under the shelf without touching any obstacles. Therefore, the objective function of the path planning problem is designed as follows:

[0160]

[0161] in, represents the penalty operator, which is used to exclude paths that pass through the obstacle range.o Represents the sum of the Euclidean distances of all points in the path trajectory. It is expressed as follows:

[0162]

[0163] Furthermore, f obs Represents the variable used to determine the obstacle level of the currently planned path, and its initial value is set to 0.

[0164]

[0165] Among them, (a s ,b s ) represents the coordinate value of the sth path point on the planned path; (xobs k ,yobs k ) corresponds to the coordinate value of the kth obstacle; n represents the total number of points on the planned path, n obs represents the total number of obstacles, i.e. other shelves and obstacles placed on the path in the warehouse logistics environment; r k Represents the obstacle radius of an abstract obstacle.

[0166] Finally, the A* algorithm is designed. It can be understood that the A* algorithm determines the shortest distance from the starting point to the end point based on the evaluation function, and selects the node at the next location according to the valuation. In smart warehousing logistics, the location of the operating goods is the starting point of the transport vehicle, and the cargo unloading location is the end point of the transport vehicle's path planning. The evaluation function of the A* algorithm for path planning from the starting point to the target end point is as follows:

[0167] f(n)=h(n)+g(n) (17)

[0168] Where n represents the current node location, h(n) represents the estimated distance from the current location to the end point, and g(n) represents the actual driving distance from the starting point to the current location. h(n) is expressed as follows:

[0169]

[0170] Among them, (a1, b1) and (a2, b2) correspond to the coordinates of the current node and the target end point respectively.

[0171]

[0172] Among them, g(n) is the previous node (a1, b1) according to the current node (a1, b1). s ,b s ) is obtained by summing up the shortest path distance.

[0173] By calculating the location cost between the current node and the target end point, the shortest path requirement can be gradually met.

[0174] Furthermore, an enhanced A* algorithm is designed. According to the above analysis, the evaluation function of the A* algorithm is obtained by linearly combining the actual travel distance g(n) from the starting point of the goods to the current node and the estimated distance h(n) from the current location to the end point. The two parts have a similar overall impact on the evaluation function. Therefore, in order to improve the calculation effect of path planning, a nonlinear factor P is added.

[0175]

[0176] Where length is the length of the grid and width is the width of the grid. Add nonlinear factors to the evaluation function:

[0177] f(n)=h(n)+P*g(n) (21)

[0178] Therefore, if the current node is far from the target end point, the P value is larger to enhance the search efficiency. Conversely, if the current node is close to the target end point, the P value is smaller to enhance the search accuracy of the algorithm.

[0179] Through the above design, the enhanced A* algorithm is applied to solve the intelligent warehousing logistics path planning problem. The initial position of the operating goods is used as the starting point of the path planning. The transport vehicle performs optimal path planning and unloads at the target destination after reaching the optimal path trajectory. A rough path planning solution Gbest(n) is obtained.

[0180] The approximate path result scheme obtained by the enhanced A* algorithm is used as the initial path record, and the initial optimal solution of the particle population of the improved particle swarm algorithm is used. The improved particle swarm algorithm is used to further optimize the calculation to obtain a more accurate optimal trajectory path, realizing double-layer optimization.

[0181] The enhanced A* algorithm is used to obtain a rough path planning scheme for the transport vehicle to transport the goods. Since the A* algorithm needs to judge the Euclidean distance of the surrounding grids, as the number of grid points increases, the search results become inaccurate and the optimal path length increases.

[0182] To this end, the present invention uses the improved particle swarm algorithm to further optimize the solution and realize double-layer optimization calculation. The approximate path record Gb obtained by the enhanced A* algorithm est (n) is used as the initial optimal solution record of the improved particle swarm algorithm.

[0183] Furthermore, the shortest distance from the starting point to the end point is determined according to the evaluation function expressed in formula (13), and the node at the next position is selected according to the estimated value as the fitness function of the improved particle swarm algorithm.

[0184] Through the above design, the roughly planned path obtained by the enhanced A* algorithm is used as the initial solution record of the improved particle swarm algorithm. The improved particle swarm further refines the calculation to obtain a more accurate path planning solution.

[0185] Taking the initial position of the cargo as the starting point of the path planning, the transport vehicle performs the optimal path planning. After reaching the optimal path trajectory, the cargo is unloaded at the target destination and placed under the shelf. After the cargo is unloaded, it is stored and retrieved to the corresponding cargo position through the elevator. The intelligent warehousing logistics system is realized with double-layer optimization, and the cargo is placed under the shelf accurately and timely.

[0186] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0187] Although the above describes the specific implementation methods of the present invention, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A multi-objective scheduling optimization method for a smart warehousing logistics system, characterized in that: The process includes: S1, build an entity information model based on the actual warehousing environment; determine the relationship model of each factor, generate the ER diagram in the database table, that is, an entity association diagram including at least six entities: warehouse, order, operation, cargo location, inventory, and pallet. When there is an order, count the order and store the relevant information in the database table; Set three-dimensional coordinate information (x, y, z) to mark the location of goods and shelf channel information, and provide data information for multi-objective cargo scheduling optimization; S2, establish a model for multi-objective optimization problems such as shelf stability in the storage environment, uniform distribution of cargo, efficiency of inbound and outbound handling, and shortest transportation operation path, which conforms to the operating procedures of the actual storage logistics system; S3, a meta-heuristic algorithm is designed to solve the multi-objective optimization problem of cargo location. The population is initialized uniformly using the good point set, the parameters in the speed update formula are adjusted using nonlinear transformation, and the position disturbance factor is introduced to adjust the particle position update of the particle swarm algorithm. Multi-objective optimization is performed to determine the optimal cargo location storage decision under the current situation.

2. A multi-objective scheduling optimization method for a smart warehousing logistics system as claimed in claim 1, characterized in that: The road surface of the smart warehousing and logistics environment is modeled, and an enhanced A* algorithm is designed to plan the path for goods out of the warehouse to obtain a rough path; the rough path result is used as the initial path plan, and the improved particle swarm optimization algorithm is used to optimize the path planning problem to determine the final route trajectory.

3. The multi-objective scheduling optimization method for a smart warehousing logistics system according to claim 1, characterized in that: In S1, the complete warehouse management process is first determined. When there are orders for ordering and replenishment, the information is uniformly entered through the order management, and different operations are generated according to the time and type of the order. According to the operation type, the storage and access location of the goods is determined, and the location information is transmitted to the warehouse transport vehicle for operation scheduling to realize the storage and access operations of the warehouse goods. After determining the operation process of warehousing and logistics, an entity relationship model of the smart warehousing and logistics system is established, and a relational database is established. It is mapped with various types in the warehousing and logistics system to complete the storage and access of information in the warehousing and logistics system. Among them, the ER diagram contains six entities: warehouse, order, operation, cargo location, inventory, and pallet; their respective related connection attributes are set as follows: Warehouse: warehouse number, warehouse name; Order: order number, order date, logistics information, customer number; Job: job number, job details, job item, job item number; Cargo location: cargo location number, cargo location, number of layers, layer name, layer number, layer specification; Inventory: inventory number, shelf number, pallet number; Pallet: Pallet number, maximum load, pallet type, pallet specifications.

4. The multi-objective scheduling optimization method for a smart warehousing logistics system according to claim 1, characterized in that: The shelf stability in S2 is to minimize the product of the weight of the goods and the number of layers where the goods are located. The optimization objective function F1 model is: Where M xyz represents the weight of the stored and retrieved goods, h represents the height of the goods, and k, p, and q correspond to the total number of shelf columns, rows, and layers respectively; The uniform distribution of the operation cargo considers the uniform distribution of the operation cargo, and evenly distributes the existing cargo and the operations to be completed to form the operation uniform distribution objective function F2 model: Among them, the smaller the value of F2, the more uniform and reasonable the operation distribution; q>1 means that when performing the operation, only the case where the number of shelves is more than two layers is considered; z Indicates the number of cargo spaces occupied by goods on the shelf at layer z. Indicates the average number of cargo spaces occupied by cargo on each floor; The in-and-out handling efficiency takes into account the movement time of operators in the rows, rows, and layers of goods entering the warehouse and the corresponding in-and-out frequency, forming the in-and-out handling efficiency objective function F3 model: Among them, M, L, and H correspond to the length, width, and height information of the cargo location respectively; v x ,v y ,v z Indicates the movement speed of the operator, S xyz Indicates the frequency of goods entering and leaving the warehouse; The transportation operation path is the shortest, and the design takes the minimum value of the matching degree between the actual cargo position and the scheduling decision as the shortest transportation operation path; specifically, determine the location of the cargo on the platform (x1, y1, z1), the cargo placement coordinates (x2, y2, z2), the cargo elevator coordinates (x m ,y m ) constitutes the shortest objective function F4 model of transportation operation path: When the position of the storage platform and the cargo placement position are on the same floor, the value is dis1. If they are on different floors, the shortest path distance is dis1+dis2+dis3. dis2 represents the floor difference between the storage platform and the cargo placement position, and dis3 represents the row difference between the elevator and the cargo position. The above-mentioned storage environment shelf stability, uniform distribution of operating goods, in-and-out warehouse handling efficiency, and shortest transportation operation path optimization model together constitute a multi-objective optimization design problem.

5. The multi-objective scheduling optimization method for a smart warehousing logistics system according to claim 1, characterized in that: The specific process of S3 is as follows: S31, use the best point set to evenly distribute the population in the search space, set the particle swarm population, the maximum number of iterations and related parameters: Among them, X i represents the position of the i-th particle in the initial population; ub and lb correspond to the upper and lower limits of the search space respectively; represents the best point taken; u satisfies 1≤u≤D; S32, through the data transmitted by the smart warehouse logistics system database, the feasible domain of shelf locations is updated in real time, and the number of decision-making operations is determined as the spatial dimension; S33, calculate the position of each particle and determine g best , g best represents the global optimal position; S34, updating the velocity and position information of the particle respectively according to the following formula (6) and the formula (7) with the position disturbance operator added, and the formulas (8) to (11); The update formulas for velocity v and position x are as follows: Where t represents the number of iterations of the calculation; i represents the particle number i=1,...,n, n represents the particle population size; ω represents the inertia weight; r1, r2 correspond to random numbers between (0,1) respectively; c1, c2 represent acceleration constants, pb est represents the individual optimal position of the particle, that is, the optimal cargo position value obtained by the particle after one iteration; g best represents the global optimal position, that is, the record with the best optimal cargo position value among all particles after one iteration; The three parameters of inertia weight ω and acceleration constants c1 and c2 are transformed nonlinearly. The nonlinear transformation is as follows: Among them, the value range of ω is [ω min ,ω max ] decreases according to the law; the acceleration constant c1 is [c 11 ,c 12 ]; the acceleration constant c2 changes between [c 21 ,c 22 ]; i represents the current iteration number, i max Indicates the maximum number of iterations of the algorithm; The perturbation operator added in combination with formula (7) is as follows: Where u is the u-th dimension component in the D-dimensional search space; x iu The value range is [SU u ,SL u ]; r is a random number randomly generated between [0,1], b is the disturbance degree factor, t represents the current iteration number, T represents the total iteration number, and the random value of l is 0 or 1; S35, calculating the function value according to the optimization objectives of shelf stability in the storage environment, uniform distribution of the cargo, efficiency of inbound and outbound handling, and shortest transportation operation path; S36, update the particle's p best and g best ; S37, the number of iterations is increased by 1, and it is determined whether the process has reached the set maximum number of iterations. If not, it returns to S34; if it has, it outputs the optimal solution for the multi-objective cargo location decision optimization.

6. A multi-objective scheduling optimization method for a smart warehousing logistics system as claimed in claim 2, characterized in that: The road modeling of the smart warehousing logistics environment is specifically as follows: The actual environment is converted into a grid map to provide an environmental basis for path planning; the grid map is to decompose the actual environment into multiple grid units of the same size L×W, each unit M lw Represented by state variables 0-1; forming a two-dimensional map information: Among them, l and w represent the row and column where the current grid is located, and their respective maximum values ​​correspond to L and W; when M lw When the value is 1, it means there is an obstacle, and when it is 0, it means it is passable; At the same time, the path planning objective function is designed; the path planning goal of the transport vehicle is to find a shortest path from the starting point (a1, b1) of the cargo to the end point (a2, b2) under the shelf without touching obstacles; therefore, the objective function of the path planning problem is designed as follows: in, represents the penalty operator, which is used to exclude paths that pass through the obstacle range, f o It represents the sum of the Euclidean distances of all points in the path trajectory, which is expressed as follows: Set f obs Represents the variable that determines the obstacle level of the current planned path, and its initial value is set to 0: Among them, (a s ,b s ) represents the coordinate value of the sth path point on the planned path; (xobs k ,yobs k ) corresponds to the coordinate value of the kth obstacle; n represents the total number of points on the planned path, nobs represents the total number of obstacles, i.e. other shelves and obstacles placed on the path in the warehouse logistics environment; r k Represents the obstacle radius of an abstract obstacle.

7. A multi-objective scheduling optimization method for a smart warehousing logistics system as claimed in claim 6, characterized in that: In the A* algorithm in smart warehousing logistics, the location of the cargo is the starting point of the transport vehicle, and the cargo unloading location is the end point of the transport vehicle's path planning. The evaluation function of the A* algorithm for path planning from the starting point to the target end point is as follows: f(n)=h(n)+g(n) (17) Where n represents the current node position, h(n) represents the estimated distance from the current position to the end point, and g(n) represents the actual driving distance from the starting point to the current position; h(n) is expressed as follows: Among them, (a1, b1) and (a2, b2) correspond to the coordinates of the current node and the target end point respectively; Among them, g(n) is the previous node (a1, b1) according to the current node (a1, b1). s ,b s ) is obtained by accumulating the shortest path distance; By calculating the location cost between the current node and the target end point, the shortest path requirement can be gradually met; To improve the calculation effect of path planning, add a nonlinear factor P: Among them, length represents the length of the grid, and width represents the width of the grid; the nonlinear factor is added to the evaluation function: f(n)=h(n)+P*g(n) (21) If the current node is far from the target end point, the P value is larger to enhance the search efficiency. On the contrary, if the current node is close to the target end point, the P value is smaller to enhance the search accuracy of the algorithm.

8. A multi-objective scheduling optimization method for a smart warehousing logistics system as claimed in claim 7, characterized in that: The improved particle swarm optimization algorithm is used to optimize the path planning problem to determine the final route trajectory, and the approximate path record Gbest(n) obtained by the enhanced A* algorithm is used as the initial optimal solution record of the improved particle swarm algorithm; The shortest distance from the starting point to the end point is determined according to the evaluation function expressed in formula (13), and the node at the next position is selected according to the estimated value as the fitness function of the improved particle swarm algorithm.

Images