Electric power material warehouse intelligent scheduling strategy based on improved A-star and whale optimization algorithm

Abstract

The invention discloses an electric power material warehouse intelligent scheduling strategy based on an improved A-star and whale optimization algorithm, and belongs to the technical field of artificial intelligence, and the strategy comprises the steps: S1, obtaining various data information contained in an electric power material warehouse; establishing a database; s2, updating the material information in the database according to the map information; s3, obtaining the name of a required cargo, searching a database to query the position information of the cargo, and providing position selection for the cargo with a plurality of positions by adopting a mixed integer linear programming algorithm; s4, performing path planning by using an improved A-star algorithm according to the starting point of the unmanned vehicle and the cargo position information, and constructing an initial path of each section of pickup of the unmanned vehicle; s5, optimizing the path; according to the electric power material warehouse intelligent scheduling strategy based on the improved A-star and whale optimization algorithm, the occupied system memory during path planning can be effectively reduced, the situation that a local optimal solution is trapped in a complex environment is avoided, and a new scheme and thought are provided for the path planning method.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an intelligent dispatching strategy for power material warehouses based on improved A-star and Beluga optimization algorithms. Background Art

[0002] Intelligent dispatching of power material warehouses is an important research field in modern supply chain management. With the diversification of market demand, the rapid development of e-commerce, and the advancement of information technology and artificial intelligence, enterprises are facing increasingly complex material management challenges. Intelligent dispatching relies on real-time data analysis, optimization algorithms, and automation technology to improve response speed and customer satisfaction. Commonly used path planning methods in intelligent dispatching include Dijkstra algorithm, A-star algorithm, etc., but these traditional path planning algorithms may have many problems in dynamic environments and complex scenarios.

[0003] Dijkstra's algorithm is a classic shortest path algorithm, which is mainly used to find the shortest path from the source node to other nodes in a weighted graph. It uses a greedy strategy to gradually expand the nodes of the known shortest path, ensuring that the nodes of the current shortest path are selected for expansion each time. This algorithm is applicable to all graphs with non-negative edge weights. However, its main disadvantage is that its time complexity is high, especially when processing large graphs, even if a priority queue is used, there is still a problem of low efficiency. In addition, the Dijkstra algorithm cannot handle negative weight edges.

[0004] The A-star algorithm is a heuristic search algorithm that aims to find the shortest path from a starting point to a target node. It combines the advantages of the Dijkstra algorithm for heuristic search. The A-star algorithm generally performs well in path planning, especially in complex environments, but its performance depends on the choice of heuristic function. If the heuristic function is not accurate enough, it may lead to inefficient search. In addition, the A-star algorithm may need to frequently recalculate the path when dealing with dynamic environments or graphs that are updated in real time, which increases the computational burden.

[0005] The White Whale Optimization Algorithm (WOA) is a nature-inspired swarm intelligence optimization algorithm based on the hunting behavior of white whales, with many significant features. It can effectively explore and utilize the search space by simulating bubble net hunting and spiral swimming strategies, has strong global search capabilities, strong adaptability, and is suitable for a variety of optimization problems. WOA has a simple structure, few parameters, and is easy to implement and debug, but its performance is sensitive to parameter settings and may fall into local optimal solutions in complex problems. Therefore, in practical applications, combining other optimization methods or improved strategies may help improve its performance. Summary of the invention

[0006] The purpose of the present invention is to provide an intelligent scheduling strategy for power material warehouses based on improved A-star and Beluga optimization algorithms, which can perform path planning for unmanned vehicles and pickup locations on a large map with complex situations, so as to solve the intelligent scheduling problem of power material warehouses in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides an intelligent dispatching strategy for power material warehouse based on improved A-star and Beluga optimization algorithms, comprising the following steps:

[0008] S1. Obtain various data information contained in the power material warehouse, including material name, material location information, power material warehouse map information, unmanned vehicle starting point for picking up goods, and goods placement point; and establish a database containing material name and material location information;

[0009] S2. Update the material information in the database according to the map information;

[0010] S3, obtaining the name of the required goods, searching the database for the goods location information, and using the mixed integer linear programming (MILP) algorithm to provide location selection for goods with multiple locations;

[0011] S4. Based on the starting point of the unmanned vehicle and the location of the goods, the improved A-star algorithm is used for path planning to construct the initial path for each section of the unmanned vehicle to pick up the goods;

[0012] S5, path optimization, based on the fusion of improved A-star and Beluga optimization algorithms, optimizes each initial path of the unmanned vehicle to pick up goods and returns the optimized total path.

[0013] Preferably, the specific steps of establishing a database in S1 are:

[0014] a. Obtain the map information of the power material warehouse, the starting point of the unmanned vehicle for picking up goods, and the cargo placement point, and extract the required material information from it and upload it to the database, including the name of the material and the warehouse code of the material location information;

[0015] b. Obtain the location information of the materials in the warehouse according to the warehouse location code and update the database, including the two-dimensional plane coordinates of the materials and the number of shelves on which they are located.

[0016] Preferably, the mixed integer linear programming algorithm used in S3 takes the cost into consideration when providing location selection.

[0017] Preferably, the improved A-star algorithm in S4 introduces a jump point search algorithm, a bidirectional search mechanism in a bidirectional RRT algorithm, and Floyd's shortest path idea to optimize the path. By introducing the jump point search algorithm, the traversal points can be reduced; by introducing the bidirectional search mechanism in the bidirectional RRT algorithm, the problem of long algorithm running time is solved, the number of search nodes is reduced, the algorithm running time is shortened, and the algorithm search efficiency is improved; by introducing the Floyd's shortest path idea to optimize the path, the problem of a large number of turning points is solved, the number of turns in the path planning process is reduced, and the smoothness of the path is improved.

[0018] Preferably, the improved A-star algorithm in S4 also introduces the environmental obstacle rate as the coefficient of the adaptive heuristic function, and the cost function f(n) is adaptively adjusted according to the number of obstacles within the rectangular area from the node to the target;

[0019] The calculation formula of obstacle rate δ is:

[0020]

[0021] Among them, x s is the horizontal coordinate of the starting point s, y s is the ordinate of the starting point s, x g is the horizontal coordinate of the end point g, y s is the ordinate of the end point g, and N is the number of obstacles within the rectangular range formed by the start point and the end point;

[0022] The obstacle rate is determined by the number of obstacles and the positions of the starting point and the end point. The environmental obstacle rate is used as the weight coefficient of the heuristic function h(n). By adjusting the weight coefficient, the adaptive change of the cost function f(n) is achieved. The improved cost function f(n) is:

[0023] f(n)=g(n)+exp(1-Obs)×h(n);

[0024] Among them, g(n) represents the actual cost, Obs represents the obstacle rate, and h(n) represents the estimated cost.

[0025] When there are fewer obstacles, the obstacle rate exp(1-δ) is larger, the coefficient of the heuristic function h(n) is larger, the more heuristic information is obtained, and the algorithm searches faster. On the contrary, when there are more obstacles in the environment, the obstacle rate exp(1-δ) is smaller, the less heuristic information is obtained, the algorithm reduces the search speed, but expands the search range to avoid falling into the local optimum.

[0026] Preferably, S5 specifically includes:

[0027] S51, generating a valid passable point list;

[0028] S52. Initialize the whale position. Since the white whale optimization algorithm is based on the population mechanism, the white whale is regarded as a search agent; and each white whale is a candidate solution and is updated during the optimization process. The matrix to the search agent position is modeled as:

[0029]

[0030] Among them, n is the population size of beluga whales, representing each solution, d is the dimension of the design variable, and each dimension represents the middle point randomly sampled from the passable list for each path;

[0031] The initial path obtained based on the improved A-star algorithm in S4 is used as the optimal path and the target fitness value of the optimization algorithm, and the path length at this time is recorded as the optimal path length. At the same time, the White Whale optimization algorithm can switch from exploration to development, which depends on the balance factor B f , its mathematical model is:

[0032]

[0033] S53, exploration phase;

[0034] The exploration phase of the Beluga optimization algorithm is built by considering the swimming behavior of beluga whales. Beluga whales can perform social behaviors in different postures, such as two pairs of beluga whales swimming closely together in a synchronized or mirrored manner. Therefore, the position of the search agent is determined by the swimming of a pair of beluga whales;

[0035] S54, development stage;

[0036] The development phase of the White Whale Optimization Algorithm was inspired by the hunting behavior of white whales. White whales can cooperate in foraging and moving based on the locations of neighboring white whales. Therefore, white whales prey by sharing each other's location information, considering the best candidate solution and other solutions. The strategy of Levy flight was introduced in the development phase of the White Whale Optimization Algorithm to improve convergence.

[0037] S55, whale fall stage;

[0038] To simulate the behavior of whale falls in each iteration, the probability of a whale falling from the population is chosen as a subjective assumption to simulate small changes in the population. It is assumed that these belugas either move to other places or are shot down and fall into the deep sea. To ensure that the population size does not change, the updated position is established using the position of the beluga and the step size of the whale's descent.

[0039] Preferably, the position update formula of the white whale in S53 is:

[0040]

[0041] Where t is the current iteration number, is the new position of the i-th beluga whale in the j-th dimension, p j (j=1,2,…,d) is randomly selected from the d-dimensional space. is the position of the i-th beluga whale in the j dimension, and is the current position of the i-th and r-th beluga, and r represents a randomly selected beluga, r1 and r2 are random numbers in the range (0,1) used to enhance the random operator in the exploration phase, and sin(2πr2) and cos(2πr2) are used to average the random numbers between fins. Depending on the dimensions selected by odd and even numbers, the updated positions reflect the synchronization or mirroring behavior of the beluga when swimming or diving.

[0042] Preferably, the position update formula of the white whale in S54 is:

[0043]

[0044] Where t is the current iteration number, and is the current position of the i-th beluga whale and a random beluga whale, is the new position of the i-th beluga whale, is the best position in the whale, r3 and r4 are random numbers in the range (0,1), C1 = 2r4 (1-t / T) is used to measure the random jump intensity of Levy flight intensity, L F is the Levy flight function, and T represents the maximum number of iterations.

[0045] Preferably, the position update formula of the white whale in S55 is:

[0046]

[0047] Among them, r5, r6 and r7 are random numbers between (0,1), X step is the step length of the whale fall, determined as: C2 is the step size factor related to the probability of whale decline and population size (C2 = 2W f ×n), ub and lb are the upper and lower bounds of the variable, respectively. It can be seen that the step size is affected by the bounds of the design variable, the number of iterations, and the maximum number of iterations.

[0048] Therefore, the present invention adopts the above-mentioned intelligent scheduling strategy for power material warehouse based on the improved A-star and Beluga optimization algorithm, and applies path planning based on the integration of the improved A-star and Beluga optimization algorithms to the field of intelligent scheduling. Compared with the traditional A* algorithm, it can effectively reduce the system memory occupied during path planning and avoid falling into local optimal solutions in complex environments, providing a new solution and idea for path planning methods.

[0049] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 This is a flow chart of an intelligent dispatching strategy for power material warehouses based on improved A-star and Beluga optimization algorithms of the present invention;

[0051] Figure 2 A flow chart of updating material location information provided by an embodiment of the present invention;

[0052] Figure 3 The flowchart of the improved A-star algorithm provided by the embodiment of the present invention. DETAILED DESCRIPTION

[0053] The following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. 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.

[0054] like Figure 1 As shown in the figure, an intelligent dispatching strategy for power material warehouse based on improved A-star and Beluga optimization algorithm specifically includes the following steps:

[0055] S1. Obtain various data information contained in the power material warehouse, including material name, material location information, power material warehouse map information, unmanned vehicle starting point for picking up goods, and goods placement point. At the same time, establish a database containing material name and location information, specifically:

[0056] The data information used in this embodiment is the material information in the actual power material warehouse, which mainly includes the name and number of the stored materials and the warehouse code containing the location information.

[0057] Step 1.1, obtain the map information of the power material warehouse, the starting point of the unmanned vehicle for picking up goods, and the cargo placement point, extract the required material information from the data set and upload it to the database, including the name of the material and the warehouse code containing the material location information.

[0058] Step 1.2: Update the database by obtaining the location information of the materials in the warehouse according to the warehouse location code, including the two-dimensional coordinates of the materials and the number of shelves on which they are located.

[0059] S2. Update the material information in the database according to the map information.

[0060] Since the two-dimensional location information obtained through the bin coding is only the literal row and column of the shelf and cannot be used as the real coordinates to provide to the unmanned vehicle for picking up goods, it is necessary to convert the two-dimensional location information of the shelf where the existing materials are located into the two-dimensional location information that the unmanned vehicle can reach in the actual map based on the plane map information of the power material warehouse. Figure 2 The figure shows the material location information update flow chart.

[0061] Step 2.1: Extract the location information of the shelf where the materials are located

[0062] Retrieve each material name from the database, and extract the corresponding two-dimensional location information of the shelf for each material.

[0063] Step 2.2: Conversion of shelf position coordinates

[0064] Assuming that the actual arrangement order of the shelves obtained according to the map information is the order of the shelf column numbers from small to large from the right side of the map to the left, and the order of the shelf row numbers from small to large from the bottom of the map to the top, then a coordinate conversion method is generated.

[0065] Assume that the total length of the map is L, the total width of the map is W, the number of rows and columns of the shelves of the acquired materials is r, and the number of columns is c. Define that as long as a shelf appears in each row or column of the map, it represents a row or column of the shelf. The number of times the shelf appears from right to left is col, and the number of times the shelf appears from bottom to top is row. The row of the shelf on the real map is real_row, and the column on the real map is real_col. Therefore, the coordinates of the shelf on the real map are:

[0066] real_row=L-row (1)

[0067] real_col=W-col (2)

[0068] Step 2.3: Conversion of reachable locations of unmanned vehicles

[0069] The coordinates obtained from step 2.2 are the real coordinates of the materials on the map, but they are not yet the coordinates of the reachable location of the unmanned vehicle. Therefore, it is necessary to continue to convert the coordinates to the map coordinates that the unmanned vehicle can reach.

[0070] Here, we consider that the shelves for storing materials are double-opened, so the unmanned vehicle can obtain materials from either side of the shelf. However, there are also cases where the shelves are arranged continuously. In this case, the unmanned vehicle can only take goods from the side without shelves. Therefore, the coordinate conversion here is only to determine whether the left and right sides of the shelf are empty, and then update the coordinates of the materials in the column direction:

[0071]

[0072] Among them, real_col represents the column of the material on the real map, r represents the right side of the shelf, l represents the left side of the shelf, and "0" and "1" represent whether the space is empty.

[0073] S3. Obtain the name of the required goods, search the database for the goods location information, and use the mixed integer linear programming (MILP) algorithm to provide cost-effective goods location selection for goods in multiple locations, which specifically includes the following steps:

[0074] Mixed Integer Linear Programming (MILP) is a powerful mathematical optimization tool widely used in various fields. The core of MILP is its ability to handle both continuous and integer variables, which enables it to effectively model and solve many complex decision problems. In MILP, the decision maker's goal is to maximize or minimize a linear objective function while satisfying a series of linear constraints. These constraints can represent resource limitations, demand requirements, or other business rules. The flexibility of MILP is reflected in its ability to handle situations where some variables must take integer values. By using MILP, decision makers can find the optimal resource allocation plan while considering all relevant constraints, thereby improving efficiency, reducing costs, and achieving better decision support.

[0075] Step 3.1: Definition of the model objective function.

[0076] Suppose there are n kinds of materials needed and each material i may have multiple pickup locations j. For each pickup location j, there is a different pickup cost c ij , including the cost of horizontal and vertical movement. The ultimate goal is to select a pickup location for each material so that the total cost of the unmanned vehicle after obtaining all materials is minimized.

[0077] Define the decision variables:

[0078] Set the binary decision variable x ij , when selecting the pickup location j for item i, x ij =1, otherwise x ij =0.

[0079] Objective function: Minimize the total cost, including the pickup cost and the layer cost:

[0080]

[0081] Among them, c ij is the cost of picking up location j including the cost of horizontal and vertical movement.

[0082] Step 3.2: Define constraints.

[0083] You can only select one pickup location for each shipment:

[0084]

[0085] Add the layer cost to the objective function:

[0086] c ij =m ij +k·l ij (6)

[0087] Among them, l ij is the number of floors of pickup location j, k is the weight coefficient of the number of floors, m ij is the corresponding horizontal movement cost.

[0088] The decision variables are binary:

[0089]

[0090] Step 3.3: Implementation of the model.

[0091] The above model is solved by calling the appropriate MILP solver. After the solution is completed, the results can be analyzed by accessing the values ​​of the decision variables. By checking the pickup location for each cargo, the optimal pickup strategy can be obtained.

[0092] S4. Path planning is performed using the improved A-star algorithm based on the starting point of the unmanned vehicle and the location of the cargo, and the initial path for each section of the unmanned vehicle to pick up the cargo is constructed, including:

[0093] Introduce a jump point search algorithm into the A-star algorithm to reduce the number of traversal points. In order to solve the problem of long algorithm running time, introduce the bidirectional search mechanism in the bidirectional RRT algorithm to reduce the number of search nodes, shorten the algorithm running time, and improve the algorithm search efficiency. In order to solve the problem of many turning points, introduce the Floyd shortest path idea to optimize the path, reduce the number of turns in the path planning process, and improve the smoothness of the path. In order to solve the problem of fixed search efficiency, introduce the environmental obstacle rate as the coefficient of the adaptive heuristic function, so that the cost function f(n) can be adaptively adjusted according to the number of obstacles within the rectangular area from the node to the target point.

[0094] Bidirectional search:

[0095] The bidirectional search mechanism of alternating forward and reverse searches can effectively prevent path search failures. First, perform a forward search with S as the starting point and G as the target point to obtain the extended node S1. Then start a reverse search with G as the starting point and S1 as the target point to obtain the extended child node G1. The search is repeated until the same extended node appears in the forward and reverse directions, then the search is stopped. The path obtained at this time is the optimal path.

[0096] Fusion Floyd algorithm optimization path:

[0097] Floyd's algorithm is called the insertion point method and is used to solve the shortest distance between two points. The algorithm gradually traverses all potential paths between adjacent nodes and updates the path matrix to determine the shortest path between any two points. In this process, the algorithm removes redundant nodes and reduces turning points in the path.

[0098] Define D(A,D) to represent the distance between two points, and P(A,B,D) to represent that point A to point D needs to pass through node B. If there is an obstacle between the two points, the distance between the two points is set to infinity, that is, D(A,D) = ∞. If there is an intermediate point C between B and D, and D(A,C)+D(C,D)<D(A,B)+D(B,D), then let D(A,D) = D(A,C)+D(C,D) discard point B as the passing point, and use point C instead, and P(A,B,D) is changed to P(A,C,D).

[0099] Introducing obstacle rate to improve the cost function:

[0100] The environmental obstacle rate is used to describe the density of obstacles in the map. Introducing the environmental obstacle rate can help the algorithm better adapt to different environments and fully consider the actual characteristics of the map when evaluating potential paths, thereby improving the efficiency and accuracy of path planning. The obstacle δ calculation formula is:

[0101]

[0102] Among them, x s is the horizontal coordinate of the starting point s, y s is the ordinate of the starting point s. g is the horizontal coordinate of the end point g, y s is the ordinate of the end point g. N is the number of obstacles within the rectangular range formed by the start point and the end point.

[0103] The environmental obstacle rate δ is determined by the number of obstacles and the positions of the starting point and the end point. The environmental obstacle rate δ is used as the weight coefficient of the heuristic function h(n). By adjusting the weight coefficient, the adaptive change of the cost function f(n) can be achieved. The improved cost function f(n) is as follows:

[0104] f(n)=g(n)+exp(1-Obs)×h(n) (9)

[0105] When there are fewer obstacles, the obstacle rate exp(1-δ) is larger, the coefficient of the estimated cost h(n) is larger, the more heuristic information is obtained, and the faster the algorithm searches. On the contrary, when there are more obstacles in the environment, the obstacle rate exp(1-δ) is smaller, the less heuristic information is obtained, the algorithm slows down the search speed, but expands the search range and avoids falling into the local optimum.

[0106] Improve the A-star algorithm process as follows Figure 3 As shown, the implementation steps of the algorithm are as follows:

[0107] Step 4.1: Algorithm initialization.

[0108] Create Open List01, Open List02, Close List01, and Close List02 to store the forward search starting point, reverse search starting point, forward search extension node, and reverse search extension node respectively. Add the starting node S and target node G to

[0109] In OpenList01 and OpenList02, it serves as the starting node for forward and reverse searches.

[0110] Step 4.2: Forward jump point search.

[0111] Take S as the starting node and G as the target node, perform a forward search first with a straight line and then with a diagonal line, and add S to CloseList01.

[0112] Step 4.3, add expansion nodes.

[0113] The expanded nodes of the forward search are added to Open List01, and their estimated costs are calculated by formula (9). The node with the smallest estimated cost is used as the starting point of the next forward search.

[0114] Step 4.4: Reverse jump point search.

[0115] Take G as the starting node and S as the target node, perform a forward search first with a straight line and then with a diagonal line, and add G to CloseList02.

[0116] Step 4.5, add expansion nodes.

[0117] The expanded nodes of the reverse search are added to Open List02, and their estimated costs are calculated by formula (9). The node with the smallest estimated cost is used as the starting point of the next reverse search.

[0118] Step 4.6: Judge and update.

[0119] Determine whether there are common points in Close List01 and Close List02. If there are common points, it means that the algorithm has found the optimal path. According to the common point nodes, reversely traverse the respective parent nodes and connect them into paths, and connect the forward and reverse paths to obtain the optimal path of the algorithm. If not, put the expansion node with the smallest forward and reverse estimated cost into Open List01 and Open List02 as the starting point of a new round of search, and return to step 4.2 to continue running the algorithm.

[0120] Step 4.7: Use the Floyd algorithm to optimize the merge path and reduce redundant points.

[0121] S5, path optimization, based on the fusion of improved A-star and White Whale optimization algorithm, optimizes each initial path of the unmanned vehicle to pick up goods and returns the optimized total path. The specific operations are as follows:

[0122] Step 5.1: Generate a list of valid traversable points.

[0123] Step 5.2, initialize the whale position.

[0124] Due to the population-based mechanism of the White Whale Optimization Algorithm, the White Whales are regarded as search agents, and each White Whale is a candidate solution and is updated during the optimization process. The matrix from the matrix to the search agent position is modeled as:

[0125]

[0126] Among them, n is the population size of beluga whales, representing each plan, d is the dimension of the design variable, and each dimension represents the middle point randomly sampled from the passable list for each path.

[0127] The initial path obtained based on the improved A-star algorithm in S4 is used as the optimal path and the target fitness value of the optimization algorithm, and the path length at this time is recorded as the optimal path length.

[0128] At the same time, the White Whale optimization algorithm can switch from exploration to development, depending on the balance factor B f , its mathematical model is:

[0129]

[0130] Step 5.3: Exploration phase.

[0131] The exploration phase of the beluga optimization algorithm is built by considering the swimming behavior of beluga whales. Beluga whales can perform social behaviors in different postures, such as two pairs of beluga whales swimming closely together in a synchronized or mirrored manner. Therefore, the position of the search agent is determined by the swimming of a pair of beluga whales, and the position of the beluga whales is updated as follows:

[0132]

[0133] Where t is the current iteration number, is the new position of the i-th beluga whale in the j-th dimension, p j (j=1,2,…,d) is randomly selected from the d-dimensional space. is the position of the i-th beluga whale in the j dimension, and are the current positions of the ith and rth beluga whales (r represents a randomly selected beluga whale), r1 and r2 are random numbers in the range (0,1) used to enhance the random operator in the exploration phase. sin(2πr2) and cos(2πr2) are used to average the random numbers between fins. Depending on the dimensions chosen by odd and even numbers, the updated positions reflect the synchronized or mirrored behavior of the beluga whales when swimming or diving.

[0134] Step 5.4, development phase.

[0135] The development phase of the White Whale Optimization Algorithm was inspired by the hunting behavior of white whales. White whales can cooperate to forage and move based on the positions of neighboring white whales. Therefore, white whales prey by sharing each other's position information, considering the best candidate solution and other solutions. The Levy flight strategy was introduced in the development phase of the White Whale Optimization Algorithm to improve convergence. We assume that they can catch prey using the Levy flight strategy, and its mathematical model is expressed as:

[0136]

[0137] Where t is the current iteration number, and is the current position of the i-th beluga whale and a random beluga whale, is the new position of the ith beluga whale, is the optimal position in the whale, r3 and r4 are random numbers in the range (0,1), and C1=2r4(1-t / T) is used to measure the random jump intensity of the Levy flight intensity.

[0138] L F is the Levy flight function, which is calculated as follows:

[0139]

[0140] Among them, u and v are random numbers from the normal distribution, and β is the default constant, which defaults to 1.5 here.

[0141] Step 5.5, whale fall stage.

[0142] To simulate the behavior of whale falls in each iteration, the probability of a whale falling from the population is chosen as a subjective assumption to simulate small changes in the population. It is assumed that these belugas either move to other places or are shot down and fall into the deep sea. In order to ensure that the population size remains unchanged, the updated position is established using the position of the beluga and the step size of the whale's descent. The mathematical model is expressed as:

[0143]

[0144] Among them, r5, r6 and r7 are random numbers between (0,1), X step is the step length of the whale fall, determined as:

[0145]

[0146] Where C2 is the step size factor related to the probability of whale decline and population size (C2 = 2W f ×n),u b and l b are the upper and lower bounds of the variables, respectively. It can be seen that the step size is affected by the bounds of the design variable, the number of iterations, and the maximum number of iterations.

[0147] In this model, the probability of a whale falling is W f is calculated as a linear function:

[0148]

[0149] The probability of a whale falling drops from 0.1 in the initial iteration to 0.05 in the final iteration, indicating that the beluga whale is less dangerous as it gets closer to the food source during the optimization process.

[0150] Step 5.6: Update the results.

[0151] The improved A-star algorithm is integrated into the exploration, development and whale fall stages of each algorithm iteration to calculate the fitness values ​​corresponding to different whales. The smallest fitness represents the best effect, and the optimal path is updated to the path with the smallest fitness to complete the optimization.

[0152] Therefore, the present invention adopts the above-mentioned intelligent scheduling strategy for power material warehouse based on the improved A-star and Beluga optimization algorithm. Compared with the traditional A* algorithm, it can effectively reduce the system memory occupied during path planning and avoid falling into the local optimal solution in a complex environment, providing a new solution and idea for the path planning method.

[0153] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. An intelligent dispatching strategy for power material warehouse based on improved A-star and Beluga optimization algorithm, characterized in that: The following steps are involved: S1. Obtain various data information contained in the power material warehouse, including material name, material location information, power material warehouse map information, unmanned vehicle starting point for picking up goods, and goods placement point; and establish a database containing material name and material location information; S2. Update the material information in the database according to the map information; S3. Obtain the name of the required goods, search the database for the goods location information, and use a mixed integer linear programming algorithm to provide location selection for goods with multiple locations; S4. Based on the starting point of the unmanned vehicle and the location of the goods, the improved A-star algorithm is used for path planning to construct the initial path for each section of the unmanned vehicle to pick up the goods; S5, path optimization, based on the fusion of improved A-star and Beluga optimization algorithms, optimizes each initial path of the unmanned vehicle to pick up goods and returns the optimized total path.

2. According to claim 1, an intelligent dispatching strategy for power material warehouse based on improved A-star and Beluga optimization algorithm is characterized in that: The specific steps for establishing a database in S1 are: a. Obtain the map information of the power material warehouse, the starting point of the unmanned vehicle for picking up goods, and the cargo placement point, and extract the required material information from it and upload it to the database, including the name of the material and the warehouse code of the material location information; b. Obtain the location information of the materials in the warehouse according to the warehouse location code and update the database, including the two-dimensional plane coordinates of the materials and the number of shelves on which they are located.

3. According to claim 1, a smart dispatching strategy for power material warehouse based on improved A-star and Beluga optimization algorithm is characterized by: S3 uses a mixed integer linear programming algorithm to provide location selection with reference to cost.

4. According to claim 1, a smart dispatching strategy for power material warehouse based on improved A-star and Beluga optimization algorithm is characterized by: The improved A-star algorithm in S4 introduces a jump point search algorithm, a bidirectional search mechanism in the bidirectional RRT algorithm, and Floyd's shortest path idea to optimize the path.

5. According to claim 1, the intelligent dispatching strategy for power material warehouse based on improved A-star and Beluga optimization algorithm is characterized by: The improved A-star algorithm in S4 also introduces the environmental obstacle rate as the coefficient of the adaptive heuristic function, and the cost function f(n) is adaptively adjusted according to the number of obstacles within the rectangular area from the node to the target; The calculation formula of obstacle rate δ is: Among them, x s is the horizontal coordinate of the starting point s, y s is the ordinate of the starting point s, x g is the horizontal coordinate of the end point g, y s is the ordinate of the end point g, and N is the number of obstacles within the rectangular range formed by the start point and the end point; The obstacle rate is determined by the number of obstacles and the positions of the starting point and the end point. The environmental obstacle rate is used as the weight coefficient of the heuristic function h(n). By adjusting the weight coefficient, the adaptive change of the cost function f(n) is achieved. The improved cost function f(n) is: f(n)=g(n)+exp(1-Obs)×h(n); Among them, g(n) represents the actual cost, Obs represents the obstacle rate, and h(n) represents the estimated cost.

6. According to claim 1, the intelligent dispatching strategy for power material warehouse based on improved A-star and Beluga optimization algorithm is characterized in that: S5 specifically includes: S51, generating a list of passable points; S52, initialize the whale position. Since the white whale optimization algorithm is based on the population mechanism, the white whale is regarded as a search agent; S53, exploration phase; S54, development stage; S55, whale fall stage.

7. The intelligent dispatching strategy for power material warehouse based on improved A-star and Beluga optimization algorithm according to claim 6 is characterized in that: The formula for updating the position of the white whale in S53 is: Where t is the current iteration number, is the new position of the i-th beluga whale in the j-th dimension, p j (j=1,2,…,d) is randomly selected from the d-dimensional space. is the position of the i-th beluga whale in the j dimension, and is the current position of the i-th and r-th beluga whales, and r represents a randomly selected beluga whale. r1 and r2 are random numbers in the range (0,1) used to enhance the random operator in the exploration phase. sin(2πr2) and cos(2πr2) are used to average the random numbers between fins.

8. The intelligent dispatching strategy for power material warehouse based on improved A-star and Beluga optimization algorithm according to claim 7 is characterized in that: The position update formula of the white whale in S54 is: Where t is the current iteration number, and is the current position of the i-th beluga whale and a random beluga whale, is the new position of the ith beluga whale, is the best position in the whale, r3 and r4 are random numbers in the range (0,1), C1 = 2r4 (1-t / T) is used to measure the random jump intensity of Levy flight intensity, L F is the Levy flight function, and T represents the maximum iteration.

9. The intelligent dispatching strategy for power material warehouse based on improved A-star and Beluga optimization algorithm according to claim 8 is characterized in that: The formula for updating the position of the white whale in S55 is: Among them, r5, r6 and r7 are random numbers between (0,1), X step is the step length of the whale fall, determined as: C2 is the step size factor related to the probability of whale decline and population size (C2 = 2W f ×n), ub and lb are the upper and lower bounds of the variable respectively.

Images