Dynamic logistics backpack group intelligent trajectory planning method

Abstract

The invention relates to the technical field of logistics distribution, in particular to a dynamic logistics backpack group intelligent trajectory planning method, which comprises the steps of constructing a logistics terminal distribution model; according to the logistics terminal distribution model, constructing a dynamic logistics backpack optimization model by taking minimization of total distribution distance cost, maximization of total customer satisfaction and maximization of total vehicle backpack value rate as a joint optimization target; solving the dynamic logistics backpack optimization model by adopting an improved grey wolf optimization algorithm to obtain an optimal distribution scheme; according to the invention, the transportation cost can be reduced, the distribution time is shortened, and the resource utilization rate is optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics distribution, and in particular to an intelligent trajectory planning method for a dynamic logistics backpack group. Background Art

[0002] Logistics distribution is an important part of the logistics system. Specifically, it refers to the logistics activities of selecting, processing, packaging, combining and distributing goods according to customer requirements within an economically reasonable area, and finally delivering them to the user's location. Logistics distribution is affected by many factors. The main issues currently concerned include the location of the logistics distribution center (Regional Distribution Center, RDC), the vehicle routing problem (VRP), the vehicle filling problem (VFP) and the layout planning of the logistics park. Among them, VRP is a typical combinatorial optimization problem and an important part of logistics distribution optimization, that is, in the distribution process, a reasonable and scientific method is used to arrange suitable distribution vehicles, plan distribution routes, etc., to effectively optimize the vehicle distribution distance, vehicle load utilization, etc. VFP is also an important part of logistics distribution optimization, that is, to reasonably load the goods on the premise of meeting the number of vehicles and the size of the order quantity, so as to effectively improve the vehicle load utilization. The cargo loading problem is also called the packing problem or the knapsack problem (KP for short). The knapsack problem is a combinatorial optimization problem, which is characterized by the refinement of VFP, allowing it to have a wider range of application scenarios. Logistics distribution VRP and VFP often complement each other and are jointly optimized to improve logistics distribution efficiency, reduce distribution costs, improve corporate service levels, and ultimately reduce corporate logistics costs and enhance corporate competitiveness.

[0003] Scholars at home and abroad have conducted corresponding research on the optimization of distribution vehicle routes, vehicle loading, vehicle scheduling, and the selection of distribution centers. When studying these issues, the weight and volume restrictions of the distributed goods and customer satisfaction during distribution were considered respectively; the latest comprehensive logistics problems have also begun to study the dynamic factors in the distribution process, such as: abnormal detection of logistics distribution, dynamic time window requirements of customers, etc. However, there are still a large number of distribution dynamic factors that have not been considered, such as the vehicle full load rate problem in the distribution process, traffic conditions in logistics distribution, actual distribution changes, and logistics loading and unloading synchronization, etc., mainly lacking a comprehensive consideration of these factors and actual distribution optimization. In addition, scholars at home and abroad have begun to combine traditional knapsack problems with logistics problems, but the research content is relatively small, and the optimization of various dynamic factors in the distribution process has not been deeply considered. Summary of the invention

[0004] To solve the above problems, the present invention provides a method for intelligent trajectory planning of a dynamic logistics backpack group, comprising the following steps:

[0005] S1. Construct a logistics terminal distribution model, which includes N+1 distribution points and K vehicles. N+1 distribution points consist of 1 distribution center and N customer terminals.

[0006] S2. According to the logistics terminal distribution model, a dynamic logistics backpack optimization model is constructed with the joint optimization objectives of minimizing the total distribution distance cost, maximizing the total customer satisfaction and maximizing the total vehicle backpack value rate;

[0007] S3. Use the improved grey wolf optimization algorithm to solve the dynamic logistics backpack optimization model and obtain the optimal distribution plan.

[0008] Beneficial effects of the present invention:

[0009] In order to solve the problem that the existing algorithms are prone to fall into local optimum and the late iteration speed is slow when solving the logistics backpack problem, the present invention provides a dynamic logistics backpack group intelligent trajectory planning method, takes the vehicle load rate, vehicle volume and total postage of the logistics distribution center as sub-objective functions, and obtains a comprehensive objective function through normalization, and proposes an improved gray wolf optimization algorithm for optimization. The present invention can reduce transportation costs, shorten delivery time, and optimize resource utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 is a flow chart of the method of the present invention;

[0011] Figure 2 Improved gray wolf optimization algorithm flow chart for the present invention;

[0012] Figure 3 This is a simulation diagram of the experiment of the present invention. DETAILED DESCRIPTION

[0013] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not 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.

[0014] The present invention provides a method for intelligent trajectory planning of a dynamic logistics backpack group. Figure 1 As shown, the following steps are included:

[0015] S1. Construct a logistics terminal distribution model, which includes N+1 distribution points and K vehicles. The N+1 distribution points consist of 1 distribution center and N customer terminals.

[0016] Specifically, logistics distribution transportation refers to that the enterprise (distribution center) arranges reasonable vehicles according to the requirements of customers (distribution points) and distributes to all customers in turn. Among them, the transportation route selection schemes for logistics distribution are mainly divided into two types, two-level distribution center distribution route planning and one-level distribution center distribution route planning. The distribution process in the two-level distribution center distribution route planning includes two layers of distribution and transportation operations. The first layer is the distribution center and distribution point of distribution transportation, which is generally distributed from the factory or general supplier to each wholesale enterprise or distributor; the second layer is from the distribution point to the customer point, which is generally distributed from the distributor to the customer; the distribution process in the first-level distribution center distribution route planning is only the distribution center to each customer point, omitting the distribution transit. In the actual logistics distribution process, the distribution of the general supplier and the distributor / retailer can be carried out separately, so the present invention adopts the distribution route planning distribution process of the first-level distribution center, which can not only reflect the basic characteristics of logistics distribution, but also adapt to different logistics distribution scenarios.

[0017] S2. According to the logistics terminal distribution model, a dynamic logistics backpack optimization model is constructed with the joint optimization objectives of minimizing the total distribution distance cost, maximizing the total customer satisfaction and maximizing the total vehicle backpack value rate.

[0018] Specifically, in the actual logistics distribution process, the dynamic logistics path optimization problem is inseparable from the logistics backpack problem. When planning the vehicle distribution plan, the requirements of vehicle loading must be met, and the loading order of goods is obtained by the distribution plan, so the two influence and restrict each other. In the embodiment of the present invention, the dynamic logistics path optimization mainly reflects the relationship between the distribution vehicle and the transportation path, how to plan the path of the vehicle to distribute goods so that the cost of logistics distribution activities is as small as possible; the logistics backpack problem mainly reflects the relationship between the vehicle backpack loading and the ordered goods, how to make full use of the vehicle space, load more goods, and make the vehicle full load rate as large as possible during the distribution process.

[0019] The dynamic logistics backpack optimization model constructed by the present invention is expressed as follows:

[0020] L=MinZ1+MaxZ2+MaxZ3

[0021]

[0022] stC1:

[0023] C2:

[0024] C3:

[0025] C4:

[0026] C5:

[0027] Among them, Z1 represents the total delivery distance cost, Z2 represents the total customer satisfaction, and Z3 represents the total vehicle backpack value rate; c ij represents the delivery distance cost from delivery point i to delivery point j; f(t i ) represents the customer satisfaction of distribution point i, X i represents the customer demand for goods at distribution point i; M represents the quantity of goods, p m Represents the value of goods m; ω m Indicates the weight of cargo m, W k represents the load limit of vehicle k; v m Represents the volume of cargo m, V k represents the volume limit of vehicle k. ijk is a binary variable, represented by

[0028]

[0029] If vehicle k is transported from delivery point i to delivery point j, then x ijk =1, otherwise x ijk =0;

[0030] y ik is a binary variable, represented by

[0031]

[0032] If vehicle k performs loading and unloading activities at distribution point i, then y ik =1, otherwise y ik =0.

[0033] In the above expression, C1 is the load constraint of vehicle k, C2 is the volume constraint of vehicle k; C3 indicates that each delivery point is delivered by only one vehicle, that is, the uniqueness of the delivery vehicle; C4-C5 indicates that the goods are delivered from delivery point i to delivery point j by the vehicle, and the sending and receiving flags are 1 at the same time.

[0034] Specifically, in the original dynamic path optimization problem, the dynamic constraint condition is the dynamic time window requirement of the customer to accept the goods, and the present invention converts the dynamic time window requirement of the customer into a customer satisfaction objective function. Customer satisfaction is the customer's overall evaluation of the service quality of the logistics company, which may include commodity price, service attitude, expectation, etc. The present invention mainly evaluates the customer's satisfaction with the delivery time and converts the evaluation result into one of the objective functions of the optimization problem. That is, convert the customer's dynamic time window requirement into the satisfaction objective function of the terminal customer point. Assume that the delivery time required by the delivery point i is t i, that is, the time from the start of delivery to the delivery point i is t i , then customer satisfaction f(t i ) is expressed as

[0035]

[0036] Among them, E.T. i represents the earliest delivery time that delivery point i can accept, LT i It represents the latest delivery time that delivery point i can accept.

[0037] S3. Use the improved grey wolf optimization algorithm to solve the dynamic logistics backpack optimization model and obtain the optimal distribution plan.

[0038] Specifically, the Grey Wolf Optimizer (GWO) is a swarm intelligence algorithm that simulates a series of processes such as attack, pursuit, encirclement, and predation of prey by a grey wolf group. According to the predation mechanism of the grey wolf group, the algorithm is mainly divided into three processes: encirclement, hunting, and attack, as follows:

[0039] 1) Social Hierarchy: When designing the gray wolf optimization algorithm, we first need to build a gray wolf social hierarchy model. Calculate the fitness of each individual in the population, and mark the three gray wolves with the best fitness in the wolf pack as α, β, and δ, and the remaining gray wolves as ω. In other words, the social hierarchy in the gray wolf group is arranged from high to low as α, β, δ, and ω. The optimization process of GWO is mainly guided by the best three solutions (i.e., α, β, and δ) in each generation of the population.

[0040] 2) Encircling Prey: When searching for prey, the gray wolf will gradually approach the prey and surround it. The mathematical model of this behavior is as follows

[0041] D=|CXp(t)-X(t)

[0042] X(t+1)=X p (t)-AD

[0043] A=2ar1-a

[0044] C=2r2

[0045] Among them, t represents the current iteration number, A, C, a are position operators, and X p represents the location of the prey (i.e. the optimal solution), X represents the current location of the gray wolf; during the entire iteration process, a decreases linearly from 2 to 0; r1 and r2 are random numbers in [0, 1].

[0046] 3) Hunting: Gray wolves have the ability to identify the location of potential prey (optimal solution), and the search process is mainly completed under the guidance of gray wolves α, β, and δ. However, the characteristics of the solution space of many problems are unknown, and gray wolves cannot determine the exact location of prey (optimal solution). In order to simulate the search behavior of gray wolves (candidate solutions), it is assumed that α, β, and δ have a strong ability to identify the location of potential prey. Therefore, in each iteration process, the best three gray wolves (α, β, δ) in the current population are retained, and then the positions of other search agents (including ω) are updated based on their location information. The mathematical model of this behavior can be expressed as follows

[0047] D α =|C1X α -X|,D β =|C2X β -X|,D δ =|C3X γ -X

[0048] X1=X α -A α D α , X2=X β -A β D β , X3=X δ -A δ D δ

[0049]

[0050] Among them, X α , X β , X δ They represent the positions of α, β, and δ in the current population, X represents the position of the current gray wolf individual, and D α , D β , D δ They represent the distances between the current gray wolf individual and the gray wolves α, β, and δ respectively, and X(t+1) represents the updated position of the current gray wolf individual.

[0051] When |A|<1, the gray wolf will focus on searching for prey in one or certain areas.

[0052] 4) Attacking Prey: According to the formula in 2), A is a random vector in the interval [-a,a], where a decreases linearly during the iteration process. The decrease in the value of a will cause the value of A to fluctuate. When A is in the interval [-1,1], the next position of the search agent can be anywhere between the current gray wolf and the prey. When |A|>1, the gray wolves try to disperse in various areas and search for prey. This search method enables GWO to perform global search. Another search coefficient (position operator) in the GWO algorithm is C. From the formula in 2), we can see that C is a vector of random values ​​in the interval [0,2]. This coefficient provides random weights for prey to increase (|C|>1) or decrease (|C|<1). This helps GWO show random search behavior during the optimization process to avoid the algorithm falling into local optimality. It is worth noting that C does not decrease linearly. C is a random value during the iteration process. This coefficient is helpful for the algorithm to jump out of the local state, which is particularly important in the later stage of the iteration.

[0053] The present invention proposes an improved Grey Wolf Optimal algorithm based on the GWO algorithm, namely the Fast Non-Dominated Sorting Grey Wolf Optimal (FNSGWO for short); Figure 2 As shown in the figure, the improved grey wolf optimization algorithm is used to solve the dynamic logistics backpack optimization model, including:

[0054] S31. Initialize the gray wolf population P and the position of the gray wolf population P Represents the position of individual i in the gray wolf population P, and initializes the position operator; calculates the fitness of each individual in the gray wolf population P according to the objective function L, and initializes the iteration number gen=1.

[0055] S32. Execute a non-dominated sorting algorithm based on the gray wolf population P to generate the first generation gray wolf population Q; then set gen=gen+1.

[0056] Specifically, step S32 executes a non-dominated sorting algorithm according to the gray wolf population P to generate a first-generation gray wolf population Q, including:

[0057] S321. Initialize the number of layers q = 1; for each individual in the population P, based on fitness, use the Pareto dominance definition to compare its dominance relationship with other individuals in the population P;

[0058] S322. Traverse each individual in the population P. If the individual is not dominated by other individuals, store the individual in the offspring population Q. q ;

[0059] S323. Calculate and update the offspring population Q q The fitness of each individual in , and then execute step S324;

[0060] S324. Eliminate the offspring population Q from the population P q , and then determine whether the population P is empty. If so, merge all the offspring populations into the population Q and execute step S325. If not, set q = q + 1 and re-execute step S322.

[0061] S324. Update position operator; arrange all individuals in population Q in descending order of fitness, select the first individual as the alpha wolf and update its position, select the second individual as the alpha wolf and update its position, select the third individual as the alpha wolf and update its position; calculate and update the positions of the remaining individuals according to the position of the alpha wolf, and obtain the position Represents the position of individual i in population Q.

[0062] S33. Merge the gray wolf population P with the first generation gray wolf population Q to form a new population R, where the new population R is located at Represents the position of individual i in population R.

[0063] S34. Perform fast non-dominated sorting according to the new population R to obtain a non-dominated population cluster F = {F1, F2, ..., F N’}, F n represents the non-dominated individual population of level n=1, 2,…, N'; N' represents the level of the non-dominated individual population; calculate the crowding degree of each individual in each level of the non-dominated individual population.

[0064] Specifically, step S34 performs a fast non-dominated sort according to the new population R, including:

[0065] S341. Determine the dominating set and the number of dominated for each individual, including:

[0066] For individual i∈R, initialize the dominant individual set S of individual i i is an empty set, the number of dominated i = 0; compare individual i with the rest of the individuals in population R one by one. If individual i dominates individual j, j∈R and i≠j, then store individual j in the dominated individual set S i , otherwise let n i =n i +1;

[0067] S342. Store all individuals in population R whose number of dominated individuals is 0 into the non-dominated individual population F of level n=1, 2, ..., N' n ;

[0068] S343. Let the nth level non-dominated individual population F n The non-dominated value of each individual in is n;

[0069] S344. For the nth level non-dominated individual population F n For each individual in the set, reduce the number of individuals dominated by each individual in the set of individuals it dominates by 1, and store all individuals whose number of dominated individuals is 0 after reducing it by 1 into the set H;

[0070] S345. Let n = n + 1, and denote the set H as the nth level non-dominated individual population F n , and returns to step S343; the loop ends when all individuals in population R are graded.

[0071] S35. The nth level non-dominated individual population F n Add a new population P' and determine whether the number of individuals in the new population P' is greater than p. If so, execute step S36; if not, set n=n+1 and repeat step S35.

[0072] S36. Sort all individuals in the non-dominated population F according to hierarchy and crowding, extract the first p-|P'| individuals and put them into the new parent gray wolf population P+1, where |P'| represents the number of individuals in the new population P'; determine whether the number of individuals in the new parent gray wolf population P+1 is greater than p, if so, execute step S37, otherwise repeat step S36.

[0073] Specifically, the congestion calculation formula is:

[0074] id=(f m (i+1)-f m (i-1))

[0075] Among them, i d represents the crowding degree of individual i, f m (i+1) represents the objective function value of individual i+1, f m (i-1) represents the objective function value of individual i-1.

[0076] Specifically, step S36 sorts all individuals in the non-dominated species cluster F according to the level and crowding degree, including:

[0077] The more excellent individuals in the non-dominated individual population are in the front, that is, the individuals with smaller non-dominated values ​​are in the front;

[0078] If two individuals have the same non-dominated population, the one with the higher crowding degree comes first.

[0079] S37. Take the three best individuals after sorting as the alpha, beta and delta, and update the population according to the gray wolf algorithm to obtain a new offspring population Q+1; determine whether the maximum number of iterations has been reached. If so, output the optimal position in the new parent population P+1; if not, set gen=gen+1, merge the new parent population P+1 and the new offspring population Q+1 to form a new population R, and return to step S34.

[0080] In one embodiment, according to the operating principle of the simplified model of the logistics system, the following experiments are conducted:

[0081] First, the system generates the city point location coordinates and order quantity: MATLAB software generates them randomly; the corresponding parameters are configured as follows:

[0082] Delivery vehicle load limit: 800kg;

[0083] Delivery vehicle transportation distance constraint: 90km;

[0084] Delivery vehicle speed: 70km / h;

[0085] Delivery vehicle service time: 40 minutes.

[0086] This experiment was conducted on a Windows 10 Education Edition platform with a 2.40 GHz Intel i3 processor and 24 GB of memory, and the algorithm was implemented using MATLAB software and GUI tools.

[0087] The dynamic logistics backpack optimization model based on multiple objectives proposed in this invention is compared with the basic terminal distribution model. The basic terminal distribution model only optimizes the distribution distance as the objective function. The comparative experimental steps are as follows: different time and different distribution centers are selected in the dynamic logistics backpack optimization model system, and the improved non-dominated sorting gray wolf optimization algorithm proposed in this paper is used to solve the model and compare it with the solution of the distribution distance optimization model containing only the distance objective function. The effectiveness of the fast non-dominated sorting gray wolf optimization algorithm (FNSGWO) is verified by comparing the results before and after the optimization of the number of distribution vehicles, distribution distance, distribution vehicle backpack value rate, and customer satisfaction.

[0088] The results of the comparison test between the dynamic logistics backpack optimization model and the basic terminal distribution model are shown in Table 1.

[0089] Table 1 Comparison of optimization results of different distribution centers on different dates

[0090]

[0091] According to Table 1, the basic terminal distribution model only considers the optimization of the distance distribution target and only processes the distance information between the distribution point and the distribution center, which leads to a low full load rate of the distribution vehicle and low customer satisfaction. The dynamic logistics backpack optimization model integrates the three objective functions to form a multi-objective optimization problem. At the same time, it loads the customer's requirements for the delivery time window and obtains the real-time load rate attribute of the vehicle. After solving the model, the vehicle backpack value rate obtained has increased significantly by about 14.4%, and customer satisfaction has increased by 12.3%; relatively speaking, the distribution distance has not changed significantly, and has decreased by about 1.62% on average. This shows that arranging the same distribution vehicle and keeping the distribution distance cost constant can improve customer satisfaction and increase the number of goods delivered by the vehicle, and ultimately effectively improve the distribution efficiency.

[0092] In the simplified model of logistics simulation system built by the present invention, based on the dynamic logistics backpack multi-objective optimization model, the solution effect of FNSGWO algorithm is compared with that of non-dominated sorting grey wolf optimization algorithm (NSGWO), grey wolf optimization algorithm (GWO), and particle swarm optimization algorithm (PSO). The specific experimental method is as follows.

[0093] The same date (January 10) and the same distribution center were set, and the FNSGWO algorithm, NSGWO algorithm, GWO algorithm and PSO algorithm were used to solve the dynamic logistics backpack multi-objective optimization model. At the same time, different dates and distribution centers were set, and multiple groups of experiments were conducted to obtain the convergence curves of the four algorithms as shown in the figure below. Figure 3 shown.

[0094] Figure 3 In the figure, green represents the fast non-dominated sorting grey wolf optimization algorithm (FNSGWO). The four algorithms solve the dynamic logistics multi-objective optimization model under different iteration times and obtain the corresponding logistics distribution plan. The three objective function values ​​corresponding to the distribution plan are compared. The comparison results are shown in Table 2 below:

[0095] Table 2 Distribution solutions solved by the algorithm under different running times

[0096]

[0097]

[0098] according to Figure 3As shown in Table 2, the FNSGWO algorithm converges faster than the NSGWO and GWO algorithms, and obtains better objective function values. The distance objective function value is 3%-4% lower than that of the NSGWO algorithm, with an average decrease of 3.27%; 4-5% lower than that of the GWO algorithm, with an average decrease of 4.28%; 4-5% lower than that of the PSO algorithm, with an average decrease of 4.85%. The backpack value objective function value is 7%-10% higher than that of the NSGWO algorithm, with an average increase of 8.21%; 7-13% higher than that of the GWO algorithm, with an average increase of 8.21%; 6-12% higher than that of the PSO algorithm, with an average increase of 9.05%. Customer satisfaction is 6%-10% higher than that of the NSGWO algorithm, with an average increase of 7.42%; 8-15% higher than that of the GWO algorithm, with an average increase of 12.36%; 6-13% higher than that of the PSO algorithm, with an average increase of 10.54%.

[0099] In the present invention, unless otherwise clearly stipulated and limited, the terms such as "installation", "setting", "connection", "fixation" and "rotation" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral one; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium; it can be the internal connection of two elements or the interaction relationship between two elements. Unless otherwise clearly defined, ordinary technicians in this field can understand the specific meanings of the above terms in the present invention according to the specific circumstances.

[0100] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent trajectory planning of a dynamic logistics backpack group, characterized in that: The following steps are involved: S1. Construct a logistics terminal distribution model, which includes N+1 distribution points and K vehicles. N+1 distribution points consist of 1 distribution center and N customer terminals. S2. According to the logistics terminal distribution model, a dynamic logistics backpack optimization model is constructed with the joint optimization objectives of minimizing the total distribution distance cost, maximizing the total customer satisfaction and maximizing the total vehicle backpack value rate; S3. Use the improved grey wolf optimization algorithm to solve the dynamic logistics backpack optimization model and obtain the optimal distribution plan.

2. According to claim 1, a dynamic logistics backpack group intelligent trajectory planning method is characterized in that: The dynamic logistics backpack optimization model is expressed as L=MinZ1+MaxZ2+MaxZ3 Among them, Z1 represents the total delivery distance cost objective function, Z2 represents the total customer satisfaction objective function, and Z3 represents the total vehicle backpack value rate objective function; c ij represents the delivery distance cost from delivery point i to delivery point j; x ijk is a binary variable. If vehicle k is transported from distribution point i to distribution point j, then x ijk =1, otherwise x ijk =0; f(t i ) represents the customer satisfaction of distribution point i, X i represents the customer demand for goods at distribution point i; M represents the quantity of goods, p m represents the value of goods m; y ik is a binary variable. If vehicle k performs loading and unloading activities at distribution point i, then y ik =1, otherwise y ik =0;ω m Indicates the weight of cargo m, W k represents the load limit of vehicle k; v m Represents the volume of cargo m, V k represents the capacity limit of vehicle k.

3. According to claim 2, a dynamic logistics backpack group intelligent trajectory planning method is characterized in that: Customer satisfaction f(t i ) is expressed as Among them, t i It represents the time from the start of delivery to the delivery point i, ET i represents the earliest delivery time that delivery point i can accept, LT i It represents the latest delivery time that delivery point i can accept.

4. According to claim 1, a dynamic logistics backpack group intelligent trajectory planning method is characterized in that: The improved grey wolf optimization algorithm is used to solve the dynamic logistics backpack optimization model including: S31. Initialize the gray wolf population P and the position of the gray wolf population P Indicate the position of individual i in the gray wolf population P, initialize the position operator; calculate the fitness of each individual in the gray wolf population P, and initialize the iteration number gen=1; S32. Execute the non-dominated sorting algorithm according to the gray wolf population P to generate the first generation gray wolf population Q; then set gen = gen + 1; S33. Merge the gray wolf population P with the first generation gray wolf population Q to form a new population R, where the new population R is located at represents the position of individual i in population R; S34. Perform fast non-dominated sorting according to the new population R to obtain a non-dominated population cluster F = {F1, F2, ..., F N’ }, F n represents the non-dominated individual population of level n=1, 2, …, N'; N' represents the level of the non-dominated individual population; calculate the crowding degree of each individual in each level of the non-dominated individual population; S35. The nth level non-dominated individual population F n Add a new population P' and determine whether the number of individuals in the new population P' is greater than p. If so, execute step S36. If not, set n=n+1 and repeat step S35. S36. Sort all individuals in the non-dominated population F according to the hierarchy and crowding degree, extract the first p-|P'| individuals and put them into the new parent gray wolf population P+1, where |P'| represents the number of individuals in the new population P'; determine whether the number of individuals in the new parent gray wolf population P+1 is greater than p, if so, execute step S37, otherwise repeat step S36; S37. Take the three best individuals after sorting as the alpha, beta and delta, and update the population according to the gray wolf algorithm to obtain a new offspring population Q+1; determine whether the maximum number of iterations has been reached. If so, output the optimal position in the new parent population P+1; if not, set gen=gen+1, merge the new parent population P+1 and the new offspring population Q+1 to form a new population R, and return to step S34.

5. According to claim 4, a dynamic logistics backpack group intelligent trajectory planning method is characterized in that: Step S32 executes a non-dominated sorting algorithm according to the gray wolf population P to generate a first generation gray wolf population Q, including: S321. Initialize the number of layers q = 1; for each individual in the population P, compare its dominance relationship with other individuals in the population P according to the Pareto dominance definition; S322. Traverse each individual in the population P. If the individual is not dominated by other individuals, store the individual in the offspring population Q. q ; S323. Calculate and update the offspring population Q q The fitness of each individual in , and then execute step S324; S324. Eliminate the offspring population Q from the population P q , and then determine whether the population P is empty. If so, merge all the offspring populations into the population Q and execute step S325. If not, set q = q + 1 and re-execute step S322. S324. Update position operator; arrange all individuals in population Q in descending order of fitness, select the first individual as the alpha wolf and update its position, select the second individual as the alpha wolf and update its position, select the third individual as the alpha wolf and update its position; calculate and update the positions of the remaining individuals according to the position of the alpha wolf, and obtain the position Represents the position of individual i in population Q.

6. According to claim 4, a dynamic logistics backpack group intelligent trajectory planning method is characterized in that: Step S34 performs a fast non-dominated sort according to the new population R, including: S341. Determine the dominating set and the number of dominated for each individual, including: For individual i∈R, initialize the dominant individual set S of individual i i is an empty set, the number of dominated i = 0; compare individual i with the rest of the individuals in population R one by one. If individual i dominates individual j, j∈R and i≠j, then store individual j in the dominated individual set S i , otherwise let n i =n i +1; S342. Store all individuals in population R whose number of dominated individuals is 0 into the non-dominated individual population F of level n=1, 2, ..., N' n ; S343. Let the nth level non-dominated individual population F n The non-dominated value of each individual in is n; S344. For the nth level non-dominated individual population F n For each individual in the set, reduce the number of individuals dominated by each individual in the set of individuals it dominates by 1, and store all individuals whose number of dominated individuals is 0 after reducing it by 1 into the set H; S345. Let n = n + 1, and denote the set H as the nth level non-dominated individual population F n , and returns to step S343; the loop ends when all individuals in population R are graded.

7. The method for intelligent trajectory planning of a dynamic logistics backpack group according to claim 4 is characterized in that: The formula for calculating congestion is: id=(f m (i+1)-f m (i-1)) Among them, i d represents the crowding degree of individual i, f m (i+1) represents the objective function value of individual i+1, f m (i-1) represents the objective function value of individual i-1.

8. The method for intelligent trajectory planning of a dynamic logistics backpack group according to claim 4 is characterized in that: Step S36 sorts all individuals in the non-dominated species cluster F according to level and crowding degree, including: The more excellent individuals in the non-dominated individual population are in the front, that is, the individuals with smaller non-dominated values ​​are in the front; If two individuals have the same non-dominated population, the one with the higher crowding degree comes first.

Images