Tray three-dimensional loading optimization method based on adjustable supporting structure

Abstract

The invention discloses an adjustable tray three-dimensional loading optimization method based on a supporting structure, and belongs to the technical field of container loading. Based on a telescopic three-dimensional tray, constructing a three-dimensional loading model of a delivery cargo tray; and solving the three-dimensional loading model of the delivery cargo tray by using a hybrid heuristic algorithm based on a simulated annealing-genetic algorithm to obtain an optimal loading scheme, and completing the three-dimensional loading optimization of the adjustable tray based on the supporting structure. According to the invention, the problems of complex cargo loading and unloading and low distribution efficiency of a loading scheme when a cargo three-dimensional loading problem and a distribution service problem are superposed are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of cargo box loading, and particularly relates to a three-dimensional loading optimization method for a tray based on adjustable support structures. Background Art

[0002] In existing research on urban logistics distribution optimization, existing solutions focus on factors such as the total cost of logistics distribution, spatio-temporal characteristics of customer demand, and vehicle load, while three-dimensional loading constraints are rarely considered. At the same time, there are also significant characteristics such as heterogeneous service demands of customers and volume differences in cargo types in practical applications. Therefore, when optimizing the logistics distribution route, in addition to necessary constraints such as logistics distribution costs and distribution facilities, the three-dimensional loading factors of the distribution goods should also be comprehensively considered according to the actual needs of customers.

[0003] Currently, three-dimensional loading solutions have been widely applied to the loading process of distribution vehicles, and a complete cargo distribution plan is formed by combining with distribution route optimization. This type of problem is to allocate a certain number of goods with different lengths, widths, and heights, and at the same time allocate a certain number of cargo distribution vehicles, requiring all the goods to be loaded into the carriage space according to rules and satisfying constraints such as the total cost of logistics distribution, vehicle volume utilization rate, and customer point time window requirements, so as to minimize the number of required distribution vehicles. The solutions to this type of problem often only consider improving the volume utilization rate of the vehicle to improve the distribution efficiency. However, in the actual distribution process, the links that can improve the distribution efficiency include not only the three-dimensional loading operation of the carriage cargo, but also its loading and unloading operations. In addition, the existing three-dimensional loading technology solutions still cannot solve the problems of rapid loading, unloading, and distribution of a large number of small goods.

[0004] The existing three-dimensional loading technology uses a series of heuristic algorithms to layout and plan the goods in the carriage, that is, given a certain number of goods with different lengths, widths, and heights, and at the same time given a certain number of distribution vehicles, requiring all the goods to be loaded into the three-dimensional carriage in sequence and satisfying certain constraints to reduce the number of distribution vehicles used. In the process of urban logistics distribution, a planned and organized vehicle loading plan can improve the volume and load utilization rate of distribution vehicles, and thus improve the efficiency of distribution services.

[0005] The functions of logistics include transportation, storage, loading and unloading, handling, packaging, distribution processing, and information processing. The current three-dimensional loading technology often only focuses on whether more volume of goods can be accommodated in the carriage, while often ignoring its impact on other functions of logistics, mainly involving the following two points: The existing three-dimensional loading scheme has problems in implementation during the loading and unloading process. It has high economic benefits for some long-distance transports such as ports and multimodal transports, but is not applicable to transports with certain timeliness requirements for logistics such as express delivery and supermarkets. On the one hand, for the scheme with the single goal of maximizing the loading rate, there will surely be confusion in the distinction of the ownership of goods, bringing certain pressure to the unloading work at the arrival station; on the other hand, the quantity of express goods is huge. Without using specific tools or containers for preprocessing the goods, the subsequent loading and unloading efficiency will be greatly reduced.

[0006] The three-dimensional loading scheme brings difficulties to the use of automated equipment. On the one hand, the carriage is a closed space, and it is difficult for large automated loading and unloading equipment to operate in the internal space; on the other hand, without an intermediate carrier for loading, the volume will be very large. Generally, only a single device can be used for loading, and the advantages of automated equipment cannot be fully utilized. Summary of the Invention

[0007] Aiming at the above deficiencies in the prior art, a three-dimensional loading optimization method for a tray based on an adjustable support structure provided by the present invention solves the problems of complex loading and unloading of the loading scheme and low distribution efficiency when the three-dimensional loading problem of goods is superimposed with the distribution service problem.

[0008] In order to achieve the above invention purpose, the technical solution adopted by the present invention is: a three-dimensional loading optimization method for a tray based on an adjustable support structure, including the following steps: S1. Construct a retractable three-dimensional tray; S2. Based on the retractable three-dimensional tray, construct a three-dimensional loading model for the distribution goods tray; S3. Use a hybrid heuristic algorithm based on the simulated annealing-genetic algorithm to solve the three-dimensional loading model of the distribution goods tray, obtain the optimal loading scheme, and complete the three-dimensional loading optimization of the tray based on the adjustable support structure.

[0009] The beneficial effects of the present invention are as follows: The retractable three-dimensional tray proposed by the present invention, on the one hand, solves the problem of the support surface in the three-dimensional loading problem of the tray; on the other hand, taking the tray as the minimum unloading unit for the demand customer point can improve the loading and unloading efficiency and reduce the time cost caused by low loading and unloading efficiency; The tray loading and the carriage loading of multiple vehicle types save distribution resources. The present invention can use as high a loading rate as possible to improve the full utilization of the vehicle volume and avoid the waste of transport capacity resources caused by the excessive use of vehicles multiple times. In addition, the present invention combines the three-dimensional loading with the use of multiple vehicle types to form a new solution, which can call appropriate vehicle types according to the demand and avoid the waste of resources caused by improper vehicle selection; at the same time, the use of the standard tray and the carriage proposed by the present invention can reduce the complexity of the optimization process and the actual operation process.

[0010] Further, the retractable three-dimensional tray in step S1 includes a goods placement tray, four support rods, and a stacking surface; the top view of the goods placement tray is quadrilateral; each of the support rods is vertically embedded in the four corners of the goods placement tray; each of the support rods can be retracted through an adjustment switch in the middle of the support rod; the stacking surface is a plane formed by connecting the tops of the support rods.

[0011] The beneficial effects of the above further solution are as follows: The retractable three-dimensional tray proposed by the present invention solves the problem of the support surface in the three-dimensional loading problem on the one hand. On the other hand, taking the tray as the minimum unloading unit for customer demand points can improve the loading and unloading efficiency and reduce the time cost caused by low loading and unloading efficiency.

[0012] Further, the objective function of the three-dimensional loading model of the distribution goods tray is:

[0013]

[0014]

[0015]

[0016]

[0017] Where, is the objective function of the three-dimensional loading model of the distribution goods tray; is the cost conversion weight coefficient; is the cost of the distribution vehicle; is the average loading rate of all distribution vehicles; is the total number of vehicles; is the distribution vehicle number; is the set of distribution vehicles; is the maximum value function; is the th vehicle from customer point to to during the service cycle is the distribution cost per unit distance during vehicle distribution; and are both customer point numbers; is the set of customer points; represents whether the th vehicle transports from customer point to to during the service cycle is the service cycle; is the set of service cycles; is the distance from the customer point to ; is the usage cost of the vehicle; is within the service cycle the unit vehicle usage cost of the vehicle model used by the th vehicle; represents whether the th vehicle is used within the service cycle decision variable; is the time cost that the vehicle has to bear for arriving at the unloading site too early; is the unit time cost for arriving too early; is the customer point starting point of the required service time window; is within the service cycle the time when the th vehicle arrives at the th customer point; is the time cost that the vehicle has to bear for arriving at the unloading site too late; is the unit time cost for arriving too late; is the customer point ending point of the required service time window; is within the service cycle the time when the th vehicle arrives at the th customer point; is the unloading time of the unit retractable three-dimensional pallet; is the set of all goods; is within the service cycle the number of retractable three-dimensional pallets of the th vehicle at the th customer point; is the distance from the customer point to ; is the vehicle driving speed; is the retractable three-dimensional pallet number; is the set of all retractable three-dimensional pallets; is within the service cycle the volume of the th retractable three-dimensional pallet of the th vehicle; is within the service cycle the length of the carriage of the th vehicle; is within the service cycle the width of the carriage of the th vehicle; During the service period the height of the carriage of the th vehicle; is the goods number; represents that during the service period the th vehicle's th customer point's th piece of goods is a decision variable indicating whether it is loaded onto the th retractable three-dimensional pallet; During the service period the th vehicle's th retractable three-dimensional pallet's th piece of goods' length; During the service period the th vehicle's th retractable three-dimensional pallet's th piece of goods' width; During the service period the th vehicle's th retractable three-dimensional pallet's th piece of goods' height.

[0018] The beneficial effect of the above further solution is that the three-dimensional loading model of the distribution goods pallet proposed by the present invention establishes a dual-objective model starting from the loading rate and loading cost, which can improve the loading rate while minimizing the loading cost as much as possible, thereby enhancing the practicality of the present invention.

[0019] Furthermore, the cost conversion weight coefficient is determined by the entropy method, which specifically includes the following steps: A1. Construct a set of solutions and a set of objectives based on the situations of using and not using pallets; A2. Establish an evaluation matrix according to the set of solutions and the set of objectives:

[0020] wherein, is the evaluation matrix; is the first objective; is the second objective; is the first solution; is the second solution; is the evaluation value of the first objective corresponding to the first solution; is the evaluation value of the first objective corresponding to the second solution; is the evaluation value of the second objective corresponding to the first solution; is the evaluation value of the second objective corresponding to the second solution; is the solution The corresponding target evaluation value; is the th solution; is the solution number; is the th target; is the target number; is the total number of solutions; is the total number of targets; A3. According to the evaluation matrix, the entropy value is obtained:

[0021]

[0022]

[0023] Among them, is the entropy value of the th target; is the reciprocal of the natural logarithm of; is the solution corresponding to the th target the sum of all evaluation values; is the natural logarithm; A4. According to the entropy value, the cost conversion weight coefficient is obtained:

[0024]

[0025]

[0026] Among them, is the cost conversion weight coefficient; is the cost conversion weight coefficient of the first target; is the cost conversion weight coefficient of the second target; is the th target's cost conversion weight coefficient; is the th target's consistency degree of solution contribution.

[0027] The beneficial effect of the above further solution is: calculating the cost conversion weight coefficient to prepare for the construction of the three-dimensional loading model of the distribution goods pallet.

[0028] Furthermore, the constraint conditions of the three-dimensional loading model of the distribution goods pallet include:

[0029]

[0030]

[0031]

[0032]

[0033]

[0034]

[0035]

[0036]

[0037]

[0038]

[0039]

[0040]

[0041]

[0042]

[0043]

[0044]

[0045]

[0046] wherein, is the abscissa coordinate of the th piece of cargo on the th retractable three-dimensional pallet of the th vehicle within the service cycle; is the length of the th retractable three-dimensional pallet of the th vehicle within the service cycle; is the length of the th piece of cargo on the th retractable three-dimensional pallet of the th vehicle within the service cycle; is the length of the th piece of cargo on the th retractable three-dimensional pallet of the th vehicle within the service cycle; is the abscissa coordinate of the th piece of cargo on the The vertical axis coordinate of the nth piece of goods; During the service cycle n, the width of the nth retractable three-dimensional pallet of the During the service cycle n, the width of the nth piece of goods of the nth retractable three-dimensional pallet of the During the service cycle n, the vertical axis coordinate of the nth piece of goods of the nth retractable three-dimensional pallet of the During the service cycle n, the height of the nth retractable three-dimensional pallet of the During the service cycle n, the height of the nth piece of goods of the nth retractable three-dimensional pallet of the During the service cycle n, the horizontal axis coordinate of the nth retractable three-dimensional pallet of the During the service cycle n, the length of the During the service cycle n, the vertical axis coordinate of the nth retractable three-dimensional pallet of the During the service cycle n, the width of the During the service cycle n, the vertical axis coordinate of the nth retractable three-dimensional pallet of the During the service cycle n, the height of the horizontal axis coordinate of the goods ; horizontal axis coordinate of the goods ; horizontal axis coordinate of the goods ; is the decision variable for overlapping goods and goods in the X-axis direction; and are both goods numbers; is the set of all goods; is the vertical axis coordinate of goods ; is the vertical axis coordinate of goods ; is the width of goods ; is the decision variable for overlapping goods and goods in the Y-axis direction; is the vertical axis coordinate of goods ; is the vertical axis coordinate of goods ; is the height of goods ; is the decision variable for overlapping goods and goods in the Z-axis direction; is the horizontal axis coordinate of the retractable three-dimensional pallet ; is the horizontal axis coordinate of the retractable three-dimensional pallet ; is the length of the th retractable three-dimensional pallet of the th vehicle within the service cycle ; is the decision variable for overlapping the retractable three-dimensional pallet and the retractable three-dimensional pallet in the X-axis direction; is the vertical axis coordinate of the retractable three-dimensional pallet ; is the vertical axis coordinate of the retractable three-dimensional pallet ; is the width of the th retractable three-dimensional pallet of the th vehicle within the service cycle ; is the decision variable for overlapping the retractable three-dimensional pallet and the retractable three-dimensional pallet in the Y-axis direction; is the vertical axis coordinate of the retractable three-dimensional pallet ; is the vertical axis coordinate of the retractable three-dimensional pallet ; For the service cycle Within, the height of the th retractable three-dimensional tray of the th vehicle; is the decision variable for overlapping the retractable three-dimensional tray and the retractable three-dimensional tray in the Z-axis direction; is the distribution vehicle number; is the set of distribution vehicles; is to judge whether, within the service cycle the th vehicle serves customer point i; is the service cycle; is the set of service cycles; and are both customer point numbers; is the set of customer points; is the set of all transportation nodes; is to represent whether, within the service cycle the th vehicle transports from customer point to ; is the number of customer points served by the th vehicle within the service cycle ; is the weight of the goods carried by the th vehicle within the service cycle ; is the maximum load capacity of the vehicle; is to represent whether, within the service cycle the th vehicle is used; is the decision variable for whether the th item of goods on the th retractable three-dimensional tray of the th vehicle within the service cycle is loaded; is the retractable three-dimensional tray number; is the set of all retractable three-dimensional trays; is the goods number.

[0047] The beneficial effect of the above further solution is that the length, width and height of the goods, trays and carriages are constrained from three-dimensional coordinates. Based on the constructed three-dimensional loading model of the distribution goods trays, the goods loading efficiency can be improved and the utilization rate of the carriage space can be increased as much as possible.

[0048] Furthermore, the specific step S3 is as follows: S301. Obtain vehicle information, three-dimensional cargo information, cargo weight information, customer point coordinates, customer demand information, and demand time window; S302. Set the termination temperature and the maximum number of iterations without improvement; S303. Perform three-dimensional loading of pallets based on the vehicle information, three-dimensional cargo information, customer point coordinates, customer demand information, termination temperature, and the maximum number of iterations without improvement, to obtain the pallet loading results at each customer point; S304. Perform three-dimensional loading of the carriage based on the cargo weight information, demand time window, customer point coordinates, and the pallet loading results at each customer point, to obtain the optimal loading plan.

[0049] The beneficial effects of the above further solution are as follows: The retractable three-dimensional pallet proposed by the present invention, on the one hand, solves the problem of the support surface in the three-dimensional loading problem. On the other hand, taking the pallet as the minimum unloading unit for demand customer points can improve the loading and unloading efficiency and reduce the time cost caused by low loading and unloading efficiency; The pallet loading and the carriage loading of multiple vehicle types save distribution resources. The present invention can make the best use of the loading rate as much as possible to improve the full utilization of the vehicle volume and avoid the waste of transport capacity resources caused by the excessive use of vehicles with multiple frequencies. In addition, the present invention combines three-dimensional loading with the use of multiple vehicle types to form a new solution, which can call appropriate vehicle types according to the demand volume and avoid the waste of resources caused by improper vehicle selection; At the same time, the use of standard pallets and carriages proposed by the present invention can reduce the complexity of the optimization process and the actual operation process.

[0050] Further, the step S303 is specifically as follows: S3031. Arrange the goods to be loaded at the current customer point in descending order of volume according to the three-dimensional cargo information, customer point coordinates, and customer demand information, to obtain the initial loading sequence; S3032. Based on the vehicle information and the initial loading sequence, use the wall-building theory to perform three-dimensional loading with the carriage height as the height of the retractable three-dimensional pallet, to obtain a feasible solution under the current loading sequence including loadable goods and unloadable goods; S3033. According to the feasible solution under the current loading sequence, the termination temperature, and the maximum number of iterations without improvement, determine whether the termination condition is satisfied. If so, perform the annealing operation, and use the optimal solution in the feasible solution under the current loading sequence as the optimal loading solution, and enter step S3034. Otherwise, use the optimal solution in the feasible solution under the current loading sequence as the first optimal loading solution, and perform the annealing operation, selection operation, crossover operation, and mutation operation, and return to step S3032; The termination condition is that the current annealing temperature is not greater than the termination temperature, or the number of times without improvement of the fitness function of the feasible solution under the current loading sequence compared with the fitness function of the first optimal loading solution reaches the maximum number of iterations without improvement; S3034. According to the optimal loading solution, determine whether the tray loading rate of the current retractable three-dimensional tray at the current customer point is less than the threshold. If so, lower the height of the retractable three-dimensional tray and return to step S3032. Otherwise, record the specifications of the current retractable three-dimensional tray and proceed to step S3035; S3035. Determine whether all the goods to be loaded at the current customer point have been completely loaded. If so, obtain the tray loading result of the current customer point according to the specifications of each retractable three-dimensional tray, and return to step S3031 to perform the tray loading of the next customer point. Otherwise, arrange the unloadable goods in the optimal loading solution in descending order of volume to obtain a new loading sequence, update the initial loading sequence to the new loading sequence, and return to step S3032 to perform the loading of the next retractable three-dimensional tray.

[0051] The beneficial effects of the above further solution are as follows: The retractable three-dimensional tray proposed by the present invention solves the problem of the support surface of the tray in the three-dimensional loading problem on the one hand. On the other hand, taking the tray as the minimum unloading unit of the demand customer point can improve the loading and unloading efficiency and reduce the time cost caused by low loading and unloading efficiency; the tray loading saves distribution resources. The present invention can make full use of the vehicle volume with as high a loading rate as possible to avoid the waste of transport capacity resources caused by the excessive use of vehicles multiple times.

[0052] Further, the specific steps of step S304 are as follows: S3041. Cluster according to the customer point coordinates to obtain several clustered customer groups; S3042. According to the vehicle information, randomly select a vehicle to load the current clustered customer group into the carriage; S3043. Obtain the tray loading result of the current clustered customer group according to the tray loading results of each customer point; S3044. Sort the tray loading result of the current clustered customer group in descending order of volume to obtain the initial tray loading sequence; S3045. According to the initial tray loading sequence, use the wall-building theory for three-dimensional loading to obtain a feasible solution including loadable trays and unloadable trays under the current tray loading sequence; S3046. Determine whether the vehicle loading termination condition is satisfied according to the feasible solution, termination temperature, and maximum non-improving iteration count under the current pallet loading sequence. If so, perform the annealing operation, use the optimal solution in the feasible solution under the current pallet loading sequence as the optimal pallet loading solution, and proceed to step S3047. Otherwise, use the optimal solution in the feasible solution under the current pallet loading sequence as the first optimal pallet loading solution, and perform the annealing operation, selection operation, crossover operation, and mutation operation, then return to step S3045. The vehicle loading termination condition is that the current annealing temperature is not greater than the termination temperature, or the number of non-improving times reaches the maximum non-improving iteration count when comparing the fitness function of the feasible solution under the current pallet loading sequence with the fitness function of the first optimal pallet loading solution. S3047. According to the optimal pallet loading solution, determine whether the loading rate of the current vehicle is less than the threshold. If so, select a smaller carriage for loading and return to step S3045. Otherwise, record the model of the current vehicle and proceed to step S3048. S3048. Determine whether all the pallets to be loaded in the current clustered customer group have been loaded. If so, obtain the vehicle loading result of the current clustered customer group according to the specifications of each vehicle, and return to step S3042 to perform the vehicle loading of the next clustered customer group. Otherwise, arrange the non-loadable pallets in the optimal pallet loading solution in descending order of volume to obtain a new pallet loading sequence, update the initial pallet loading sequence to the new pallet loading sequence, randomly select a vehicle model as the vehicle model for the next loading, and return to step S3045 to perform the loading of the next vehicle. S3049. Obtain the vehicle loading information according to the vehicle loading results of each clustered customer group. S30410. Perform path optimization for each vehicle respectively according to the cargo weight information, demand time window, and vehicle loading information to obtain the optimal loading plan.

[0053] The beneficial effects of the above further solution are as follows: The vehicle loading of multiple vehicle models saves distribution resources. The present invention can make full use of the vehicle volume as much as possible with a high loading rate, avoiding the waste of transport capacity resources caused by the excessive use of vehicles multiple times. In addition, the present invention combines three-dimensional loading with the use of multiple vehicle models to form a new solution, which can call appropriate vehicle models according to the demand volume, avoiding the waste of resources caused by improper vehicle selection. At the same time, the use of standard pallets and carriages proposed by the present invention can reduce the complexity of the optimization process and the actual operation process.

[0054] Further, the specific steps of step S30410 are as follows: S304101. According to the vehicle loading information, obtain the coordinates of the customer points to which each retractable three-dimensional pallet in the current transport vehicle belongs. S304102. Randomly generate a path solution based on the coordinates of the customer points to which the retractable three-dimensional pallets in the current transport vehicle belong; S304103. According to the path solution, determine whether the termination condition for path optimization is satisfied. If so, use the result with the shortest path in the path solution as the optimal path result of the current vehicle, and return to step S304102 to perform path optimization for the next transport vehicle until the path optimization of all transport vehicles is completed to obtain a loading plan, and then enter step S304104. Otherwise, use the result with the shortest path in the path solution as the first optimal path, and perform selection, crossover, and mutation iterations, and return to step S304102; the termination condition for path optimization is that the current annealing temperature is not greater than the termination temperature, or the number of times without improvement in the optimal length of the current path solution compared with the first optimal path reaches the maximum number of non-improving iterations; S304104. According to the cargo weight information, required time window, customer point coordinates, and loading plan, determine whether the termination condition for the overall plan is satisfied. If so, obtain the optimal loading plan. Otherwise, use the loading plan as the original optimal plan, and return to step S3041 to perform iteration on the overall plan; the termination condition for the overall plan is that the current annealing temperature is not greater than the termination temperature, or the number of times without improvement in the value of the objective function of the three-dimensional loading model of the distribution cargo pallets based on the current loading plan compared with the value of the objective function of the three-dimensional loading model of the distribution cargo pallets of the original optimal plan reaches the maximum number of non-improving iterations.

[0055] The beneficial effects of the above further solution are as follows: Through path optimization, the transportation efficiency is improved, and at the same time, the overuse of vehicles is reduced, and the vehicle life is prolonged. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a flowchart of the method of the present invention.

[0057] Figure 2 It is a distribution map of the distribution center and customer points in the embodiment of the present invention.

[0058] Figure 3 It is a schematic structural diagram of the retractable three-dimensional pallet of the present invention.

[0059] Figure 4 It is a schematic diagram of pallet loading by the "wall building" method in the embodiment of the present invention.

[0060] Figure 5 It is a schematic diagram of the process of adjusting the pallet height in the embodiment of the present invention.

[0061] Figure 6 It is a schematic diagram of pallet stacking in the optimized carriage in the embodiment of the present invention.

[0062] Figure 7This is a bar chart comparing relevant indicators before and after the optimization of the multi-vehicle loading plan under the use of pallets in the embodiments of the present invention. Detailed implementation manners

[0063] The following describes the detailed implementation manners of the present invention to facilitate those skilled in the art of the present technology to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the detailed implementation manners. For those of ordinary skill in the art of the present technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions created using the concept of the present invention are within the scope of protection.

[0064] As Figure 1 shown, in an embodiment of the present invention, a three-dimensional loading optimization method for a pallet based on adjustable support structure includes the following steps: S1. Construct a retractable three-dimensional pallet; S2. Based on the retractable three-dimensional pallet, construct a three-dimensional loading model for the pallet of the distributed goods; S3. Use a hybrid heuristic algorithm based on the simulated annealing-genetic algorithm to solve the three-dimensional loading model of the pallet of the distributed goods, obtain the optimal loading plan, and complete the three-dimensional loading optimization of the pallet based on the adjustable support structure.

[0065] In this embodiment, case analysis is used to further illustrate the present invention. As Figure 2 shown in the distribution map of the distribution center and customer points. Taking a certain express delivery company's distribution center delivering to 23 surrounding customer points as an example, the express delivery company has three commonly used distribution vehicle models. The pallet is 1.2m long and 1m wide, and the carriage specifications are 3.6×2×2.2m, 4.8×2×2.2m, and 6×2×2.2m respectively. Among them, the truck load capacities are 6 tons, 8 tons, and 10 tons respectively, and the vehicle usage costs are 100 yuan, 150 yuan, and 200 yuan respectively. Each vehicle model is vehicle model 1, 2, and 3 respectively, and the vehicle driving speed is 15km / h. In order to quantify the benefits of the pallet in improving the loading and unloading efficiency, it is set that the time for unloading each pallet of the goods using the three-dimensional loading plan of the pallet at each customer point is 5 minutes, while the time required for unloading using the ordinary three-dimensional loading plan is 30 minutes at each customer point. The penalty cost for violating the customer service time window during transportation is 100 yuan / h, and the transportation cost is 5 yuan / km.

[0066] There is no specific specification standard for the packaging in the express delivery industry. In the example, 8 commonly used packaging specifications of goods with different sizes are used as the objects, as shown in Table 1 specifically. The demand for goods at each distribution point is also different. Information such as the coordinates, time windows, and goods demands of the customer points is shown in Table 2, and Table 3 gives the initial settings of the relevant parameters of the optimization algorithm.

[0067] Table 1

[0068] Table 2

[0069] Table 3

[0070] As Figure 3 shown, the retractable three-dimensional tray in step S1 includes a cargo placement tray, four support rods, and a stacking surface; the top view of the cargo placement tray is quadrilateral; each of the support rods is vertically embedded at the four corners of the cargo placement tray; each of the support rods can be telescoped through an adjustment switch in the middle of the support rod; the stacking surface is a plane formed by connecting the tops of the support rods.

[0071] The objective function of the three-dimensional loading model of the distribution cargo tray is as follows:

[0072]

[0073]

[0074]

[0075]

[0076] Among them, is the objective function of the three-dimensional loading model of the distribution cargo tray; is the cost conversion weight coefficient; is the cost of the distribution vehicle; is the average loading rate of all distribution vehicles; is the total number of vehicles; is the distribution vehicle number; is the set of distribution vehicles; is the maximum value function; is the th vehicle from customer point within the service cycle to distribution cost; is the distribution cost per unit distance during vehicle distribution; and are both customer point numbers; is the set of customer points; is to indicate that within the service cycle the th vehicle whether from customer point is transported to decision variable; is the service cycle; is the set of service cycles; is the distance from the customer point to ; is the usage cost of the vehicle; is within the service cycle , the unit vehicle usage cost of the vehicle model used by the th vehicle; represents whether the th vehicle is used within the service cycle ; decision variable of the vehicle; is the time cost that needs to be borne when the vehicle arrives at the unloading site prematurely; is the unit time cost of premature arrival; is the start point of the required service time window of the customer point ; is within the service cycle , the time when the th vehicle arrives at the th customer point; is the time cost that needs to be borne when the vehicle arrives at the unloading site too late; is the unit time cost of late arrival; is the end point of the required service time window of the customer point ; is within the service cycle , the time when the th vehicle arrives at the th customer point; is the unloading time of a unit retractable three-dimensional pallet; is the set of all goods; is within the service cycle , the number of retractable three-dimensional pallets of the th vehicle at the th customer point; is the distance from the customer point to ; is the vehicle driving speed; is the retractable three-dimensional pallet number; is the set of all retractable three-dimensional pallets; is within the service cycle , the volume of the th vehicle's th retractable three-dimensional pallet; is within the service cycle , the length of the carriage of the th vehicle; is within the service cycle , the Width of the carriage of the vehicle; During the service period the height of the carriage of the vehicle; is the goods number; Indicates that during the service period the the th decision variable indicating whether the th piece of goods at the During the service period the the th length of the During the service period the the th width of the During the service period the the th height of the

[0077] The cost conversion weight coefficient is determined by the entropy method, which specifically includes the following steps: A1. Construct a set of solutions and a set of objectives based on the situations of using pallets and not using pallets; A2. Establish an evaluation matrix based on the set of solutions and the set of objectives:

[0078] Among them, is the evaluation matrix; is the first objective; is the second objective; is the first solution; is the second solution; is the evaluation value of the first objective corresponding to the first solution; is the evaluation value of the first objective corresponding to the second solution; is the evaluation value of the second objective corresponding to the first solution; is the evaluation value of the second objective corresponding to the second solution; is the evaluation value of the th objective corresponding to the solution; is the th solution; is the solution number; is the th target; is the target number; is the total number of solutions; is the total number of targets; A3. According to the evaluation matrix, the entropy value is obtained:

[0079]

[0080]

[0081] Among them, is the th target's entropy value; is the reciprocal of the natural logarithm of; is the solution corresponding to the th target 's sum of all evaluation values; is the natural logarithm; A4. According to the entropy value, the cost conversion weight coefficient is obtained:

[0082]

[0083]

[0084] Among them, is the cost conversion weight coefficient; is the cost conversion weight coefficient of the first target; is the cost conversion weight coefficient of the second target; is the th target's cost conversion weight coefficient; is the th target's consistency degree of solution contribution degree.

[0085] The constraint conditions of the three-dimensional loading model of the distribution goods pallet include:

[0086]

[0087]

[0088]

[0089]

[0090]

[0091]

[0092]

[0093]

[0094]

[0095]

[0096]

[0097]

[0098]

[0099]

[0100]

[0101]

[0102]

[0103] wherein, is the abscissa of the th piece of cargo on the th retractable three-dimensional pallet of the th vehicle within the service cycle; is the length of the th retractable three-dimensional pallet of the th vehicle within the service cycle; is the length of the th piece of cargo on the th retractable three-dimensional pallet of the th vehicle within the service cycle; is the ordinate of the th piece of cargo on the th retractable three-dimensional pallet of the th vehicle within the service cycle; is the width of the th retractable three-dimensional pallet of the th vehicle within the service cycle; is the width of the th retractable three-dimensional pallet of the th vehicle within the service cycle; is the width of the th retractable three-dimensional pallet of the is the width of the th retractable three-dimensional pallet of the The width of the th piece of cargo on the th retractable three-dimensional pallet of the vehicle; During the service cycle the th retractable three-dimensional pallet of the th vehicle's vertical axis coordinate of the th piece of cargo; During the service cycle the height of the th retractable three-dimensional pallet of the th vehicle; During the service cycle the height of the th piece of cargo on the th retractable three-dimensional pallet of the th vehicle; During the service cycle the th retractable three-dimensional pallet of the length of the th vehicle's carriage; During the service cycle the vertical axis coordinate of the th retractable three-dimensional pallet of the th vehicle; During the service cycle the width of the th vehicle's carriage; During the service cycle the vertical axis coordinate of the th retractable three-dimensional pallet of the th vehicle; During the service cycle the horizontal axis coordinate of the cargo; During the service cycle length of the cargo; During the service cycle the decision variable for overlapping the cargo and the cargo in the X-axis direction; Both and are cargo numbers; During the service cycle the The vertical axis coordinate; is the goods width; is for the goods and the goods decision variable for Y-axis direction overlap; is the goods vertical axis coordinate; is the goods vertical axis coordinate; is the goods height; is for the goods and the goods decision variable for Z-axis direction overlap; is the horizontal axis coordinate of the retractable three-dimensional pallet ; is the horizontal axis coordinate of the retractable three-dimensional pallet ; is the service cycle within, the th vehicle's th retractable three-dimensional pallet length; is for the retractable three-dimensional pallet and the retractable three-dimensional pallet decision variable for X-axis direction overlap; is the vertical axis coordinate of the retractable three-dimensional pallet ; is the vertical axis coordinate of the retractable three-dimensional pallet ; is the service cycle within, the th vehicle's th retractable three-dimensional pallet width; is for the retractable three-dimensional pallet and the retractable three-dimensional pallet decision variable for Y-axis direction overlap; is the vertical axis coordinate of the retractable three-dimensional pallet ; is the vertical axis coordinate of the retractable three-dimensional pallet ; is the service cycle within, the th vehicle's th retractable three-dimensional pallet height; is for the retractable three-dimensional pallet and the retractable three-dimensional pallet decision variable for Z-axis direction overlap; is the distribution vehicle number; is the set of distribution vehicles; is to judge the service cycle within, the decision variable indicating whether the vehicle serves customer point i; is the set of service cycles; and are both customer point numbers; is the set of customer points; is the set of all transportation nodes; is to represent within the service cycle within, the vehicle whether transports from customer point to is within the service cycle within, the number of customer points served by the vehicle; is within the service cycle weight of the goods carried by the vehicle; is to represent within the service cycle whether to use the vehicle; is within the service cycle within, the vehicle, the number of the retractable three-dimensional pallet, whether the goods is loaded; is the set of retractable three-dimensional pallets; is the goods number.

[0104] Specifically, the step S3 is as follows: S301. Obtain vehicle information, three-dimensional information of goods, goods weight information, customer point coordinates, customer demand information, and demand time window; S302. Set the termination temperature and the maximum number of iterations without improvement; S303. According to the vehicle information, three-dimensional information of goods, customer point coordinates, customer demand information, termination temperature, and the maximum number of iterations without improvement, perform three-dimensional loading of pallets to obtain the pallet loading results of each customer point; S304. According to the goods weight information, demand time window, customer point coordinates, and the pallet loading results of each customer point, perform three-dimensional loading of the carriage to obtain the optimal loading plan.

[0105] Specifically, the step S303 is as follows: S3031. Arrange the goods to be loaded at the current customer point in descending order of volume according to the three-dimensional information of the goods, the coordinates of the customer point, and the customer demand information to obtain an initial loading sequence; S3032. According to the vehicle information and the initial loading sequence, use the height of the carriage as the height of the retractable three-dimensional pallet, and perform three-dimensional loading using the wall-building theory to obtain a feasible solution under the current loading sequence including loadable goods and unloadable goods; S3033. According to the feasible solution under the current loading sequence, the termination temperature, and the maximum number of non-improving iterations, determine whether the termination condition is met. If so, perform the annealing operation, and use the optimal solution in the feasible solution under the current loading sequence as the optimal loading solution, and enter step S3034. Otherwise, use the optimal solution in the feasible solution under the current loading sequence as the first optimal loading solution, and perform the annealing operation, selection operation, crossover operation, and mutation operation, and return to step S3032; the termination condition is that the current annealing temperature is not greater than the termination temperature, or the number of non-improving times of the fitness function of the feasible solution under the current loading sequence compared with the fitness function of the first optimal loading solution reaches the maximum number of non-improving iterations; S3034. According to the optimal loading solution, determine whether the pallet loading rate of the current retractable three-dimensional pallet at the current customer point is less than the threshold. If so, reduce the height of the retractable three-dimensional pallet and return to step S3032. Otherwise, record the specifications of the current retractable three-dimensional pallet and enter step S3035; S3035. Determine whether all the goods to be loaded at the current customer point have been loaded. If so, obtain the pallet loading result of the current customer point according to the specifications of each retractable three-dimensional pallet, and return to step S3031 to perform pallet loading for the next customer point. Otherwise, arrange the unloadable goods in the optimal loading solution in descending order of volume to obtain a new loading sequence, update the initial loading sequence to the new loading sequence, and return to step S3032 to perform loading for the next retractable three-dimensional pallet.

[0106] In this embodiment, the "wall-building" theory is used to determine whether the goods can be loaded into the space. If so, the goods code is 1, otherwise it is 0 until all the goods at the customer point have been judged to obtain a feasible solution containing (0, 1).

[0107] In terms of loading logic, the present invention uses the "wall-building" theory for loading. The specific method is to use a vertical surface inside the pallet as the starting surface and perform "wall-building" operations along this surface. It should be noted that each layer of the wall is regarded as a whole, and the thickness of the first goods of the wall is the thickness of this layer of the wall. When entering the next layer of "wall-building" operation, a new thickness will be obtained until the thickness (length) constraint of the three-dimensional loading of the pallet is reached, that is, it does not exceed the other side of the pallet. As Figure 4 shown.

[0108] The specific steps are as follows: (1) First, define the three dimensions of the goods to be loaded in the order of .

[0109] (2) Sort the volumes of the goods from largest to smallest to generate an initial traversal sequence.

[0110] (3) Give the pallet a height constraint of up to the height of the carriage. Take the innermost left side of the pallet as the coordinate origin, and detect whether the three dimensions of the goods satisfy .

[0111] (4) For the goods that do not exceed the pallet constraints, first detect . If it exceeds, then detect the next good. If it does not exceed the pallet height, then detect . If it exceeds, then update the y coordinate, that is, load it onto the upper layer. If it does not exceed, then continue to detect whether the length of the good can fit into the current x-axis, that is . If it exceeds, update the x coordinate and load it onto the next layer. If it does not exceed, place the good at the current position and re-update the coordinates of the x, y, and z axes. For example, when placing the first good in an empty pallet and detecting that it does not exceed the three dimensions of the pallet, its new detection points are updated to , , .

[0112] The use of the adjustable support structure pallet in the present invention and an alternative to the matching "wall-building" theory are as follows.

[0113] In terms of equipment use, there are multi-layer packaging uses similar to the use principle of this solution. The specific method is to load small goods into larger boxes and then load the large boxes into the carriage. However, this method does not involve the use of pallets, and the loading and unloading efficiency may be reduced, and the actual effect remains to be verified. Existing ordinary pallet solutions can also improve the loading and unloading efficiency, but do not consider the problem of supporting the top goods, and its improvement of the carriage loading rate also remains to be investigated.

[0114] In terms of the loading logic, this solution uses the "wall-building" theory for loading. Existing three-dimensional loading methods also include the remaining space theory, the "layer" theory, etc. The remaining space theory requires re-planning the space inside the container every time a good is loaded. Its loading rate is relatively high, but the program running time is very long, and the computational complexity is relatively high. The height of each layer in the "layer" theory is fixed. Therefore, there are more gaps between layers, and the support of the goods is relatively unreasonable.

[0115] The above are other alternative solutions that can achieve this purpose.

[0116] In this embodiment, the process of adjusting the height of the pallet is as shown in Figure 5As shown in the figure, the initial height of the pallet is 2 m (the height of the carriage). All the goods can be loaded onto the pallet, but the loading rate is only 41.47%. At this time, it is necessary to lower the height of the pallet to save space. When the height of the pallet is 1.8 m, all the goods can still be loaded onto the pallet, and the loading rate is 46.07%. Therefore, it is necessary to continue to lower the height of the pallet. By analogy, when the height of the pallet is 1 m in the figure, the goods can be completely loaded, and the loading rate is greater than 70%, which is 82.89%. Therefore, when the quantity of this goods is determined, the height of the pallet is 1 m.

[0117] The specific steps of step S304 are as follows: S3041. Cluster according to the customer point coordinates to obtain several clustered customer groups; S3042. Randomly select a vehicle according to the vehicle information to load the current clustered customer group into the carriage; S3043. Obtain the pallet loading result of the current clustered customer group according to the pallet loading results of each customer point; S3044. Sort in descending order of volume according to the pallet loading result of the current clustered customer group to obtain the initial pallet loading sequence; S3045. According to the initial pallet loading sequence, use the wall-building theory for three-dimensional loading to obtain a feasible solution for the current pallet loading sequence including loadable pallets and unloadable pallets; S3046. According to the feasible solution for the current pallet loading sequence, the termination temperature, and the maximum number of non-improving iterations, determine whether the carriage loading termination condition is satisfied. If so, perform the annealing operation, and use the optimal solution in the feasible solution for the current pallet loading sequence as the optimal pallet loading solution, and enter step S3047. Otherwise, use the optimal solution in the feasible solution for the current pallet loading sequence as the first optimal pallet loading solution, and perform the annealing operation, selection operation, crossover operation, and mutation operation, and return to step S3045; the carriage loading termination condition is that the current annealing temperature is not greater than the termination temperature, or the number of non-improving times of the fitness function of the feasible solution for the current pallet loading sequence compared with the fitness function of the first optimal pallet loading solution reaches the maximum number of non-improving iterations; S3047. According to the optimal pallet loading solution, determine whether the loading rate of the current vehicle is less than the threshold. If so, select a smaller carriage for loading and return to step S3045. Otherwise, record the model of the current vehicle and enter step S3048; S3048. Determine whether all the pallets to be loaded in the current clustered customer group have been fully loaded. If so, obtain the carriage loading results of the current clustered customer group according to the specifications of each vehicle, and return to step S3042 to perform the carriage loading of the next clustered customer group. Otherwise, arrange the non-loadable pallets of the optimal pallet loading solution in descending order of volume to obtain a new pallet loading sequence. Update the initial pallet loading sequence to the new pallet loading sequence. Randomly select a vehicle model as the vehicle model for the next loading, and return to step S3045 to perform the loading of the next vehicle; S3049. Obtain the vehicle loading information according to the carriage loading results of each clustered customer group; S30410. Perform path optimization for each vehicle respectively according to the cargo weight information, demand time window, and vehicle loading information to obtain the optimal loading plan.

[0118] The specific steps of step S30410 are as follows: S304101. Obtain the coordinates of the customer points to which each retractable three-dimensional pallet in the current transport vehicle belongs according to the vehicle loading information; S304102. Randomly generate a path solution according to the coordinates of the customer points to which each retractable three-dimensional pallet in the current transport vehicle belongs; S304103. According to the path solution, determine whether the termination condition for path optimization is satisfied. If so, use the result with the shortest path in the path solution as the optimal path result of the current vehicle, and return to step S304102 to perform the path optimization of the next transport vehicle until the path optimization of all transport vehicles is completed to obtain the loading plan, and enter step S304104. Otherwise, use the result with the shortest path in the path solution as the first optimal path, and perform selection, crossover, and mutation iterations, and return to step S304102; the termination condition for path optimization is that the current annealing temperature is not greater than the termination temperature, or the number of non-improving times when the optimal length of the current path solution is compared with the first optimal path reaches the maximum number of non-improving iterations; S304104. According to the cargo weight information, demand time window, customer point coordinates, and loading plan, determine whether the overall plan termination condition is satisfied. If so, obtain the optimal loading plan. Otherwise, use the loading plan as the original optimal plan, and return to step S3041 to perform iterations on the overall plan; the overall plan termination condition is that the current annealing temperature is not greater than the termination temperature, or the number of non-improving times when the value of the objective function of the three-dimensional loading model of the distribution goods pallets based on the current loading plan is compared with the value of the objective function of the three-dimensional loading model of the distribution goods pallets of the original optimal plan reaches the maximum number of non-improving iterations.

[0119] Before optimization, the loading plan was to load with the maximized loading rate and then optimize the route. Among them, no intermediate carrier was used for cargo loading, nor was multi-vehicle type used for vehicle-cargo matching. This plan has a relatively high loading rate when the cargo volume is sufficient. The results of the plan before optimization are shown in Table 4.

[0120] Table 4

[0121] As can be seen from Table 4, the average vehicle loading rate of the plan before optimization is 65.51%. The main reason is that multi-vehicle types are not used, resulting in the problem of vehicle-cargo mismatch in the subsequent vehicles. At the same time, due to the lack of standardized loading and unloading auxiliary equipment before optimization, the internal cargo sorting and loading and unloading efficiency are lower. The average time cost is 3929.31 yuan, and the average cost is 4297.69 yuan. And because the internal cargo aims at the highest loading rate, there is a situation where the cargo of a certain customer is in multiple carriages at the same time. Therefore, in this case, there is a situation of serving multiple vehicles at the same customer point, which is not reasonable.

[0122] In order to improve the loading and unloading efficiency and save time costs, the optimized loading plan uses a pallet with an adjustable support structure as an intermediate carrier, and cooperates with the use of multi-vehicle types to improve the vehicle-cargo matching degree. The results of the optimized plan obtained through algorithm optimization are shown in Table 5.

[0123] Table 5

[0124] The average loading rate of the optimized plan is 78.15%. Using pallets can speed up the loading and unloading speed and save time costs. Therefore, the average time cost of using pallets is 31.92 yuan, indicating that the present invention has a faster loading and unloading efficiency. The average cost of the plan is 251.76 yuan. Considering the cost and loading rate comprehensively, the multi-vehicle type loading plan under the use of pallets in the present invention is more excellent.

[0125] The schematic diagram of pallet stacking in the optimized carriage is as Figure 6 shown. Its number represents the customer point number, and the internal square represents the three-dimensional space of the pallet. As can be seen from Figure 6 it, the telescopic structure of the pallet can adjust the pallet height to facilitate better stacking in the vehicle. At the same time, in order to solve the problem that the last vehicle is prone to vehicle-cargo mismatch, the algorithm distributes the cargo to three smaller carriages to ensure the full use of the vehicle space and avoid waste.

[0126] To more intuitively display the optimized effect, a bar chart of the comparison of relevant indicators before and after the optimization of the multi-vehicle type loading plan under the use of pallets is drawn here, as Figure 7 shown. From Figure 7It can be seen that after the optimization of the multi-vehicle loading plan under the use of pallets, the average loading rate of vehicles has increased significantly, and the time consumed for loading and unloading has also been greatly reduced.

[0127] The solution process of the embodiments of the present invention has been described in detail above. However, the present invention is not limited to the above embodiments and can also be used for vehicle routing problems. By combining time windows to reduce distribution costs, relevant distribution routes and loading plans can be optimized and designed.

[0128] The beneficial effects of the present invention are as follows: (1) The use of pallets improves the loading and unloading efficiency. In terms of three-dimensional loading, previous studies focused on effectively increasing the volume loading rate of the carriage. The three-dimensional loading method of goods is one of the important solutions to this problem. When the three-dimensional loading problem of goods is superimposed on the distribution service problem, problems such as complex loading and unloading of goods, low distribution efficiency, and great difficulty in differentiating goods in previous solutions become more prominent. The pallet with adjustable support structure height proposed in this solution solves the support surface problem of the pallet in the three-dimensional loading problem on the one hand. On the other hand, taking the pallet as the minimum unloading unit of the demand customer point can improve the loading and unloading efficiency and reduce the time cost caused by low loading and unloading efficiency. Through laboratory data comparison tests, this solution can achieve a loading rate of 78.15% and can save 99.18% of the time, that is, it improves the distribution efficiency.

[0129] (2) The three-dimensional loading of pallets and the loading of multi-vehicles save distribution resources. Under this research, the three-dimensional loading can make the best use of the loading rate to improve the full utilization of the vehicle volume and avoid the waste of transport capacity resources caused by the excessive use of vehicles multiple times. In addition, this solution combines the three-dimensional loading with the use of multi-vehicles to form a new solution, which can call appropriate vehicle types according to the demand and avoid the waste of resources caused by improper vehicle selection.

[0130] (3) Standardization reduces the operation difficulty. In the process of logistics exploration for up to a century, standardization has always been an important factor in promoting the process of modern logistics. The use of standard pallets and carriages proposed in this solution can reduce the complexity of the optimization process and the actual operation process. In the future, standardization of goods packaging and operation equipment should also be used to further improve the production operation efficiency.

Claims

1. A three-dimensional loading optimization method for a tray based on adjustable support structure, characterized in that, It includes the following steps: S1. Construct a retractable three-dimensional pallet; S2. Based on the retractable three-dimensional pallet, build a three-dimensional loading model for the distribution goods pallet; S3. Use a hybrid heuristic algorithm based on the simulated annealing-genetic algorithm to solve the three-dimensional loading model of the distribution goods pallet, obtain the optimal loading plan, and complete the three-dimensional loading optimization of the pallet with adjustable support structure.

2. The three-dimensional loading optimization method of the tray adjustable based on the support structure according to claim 1, wherein, In step S1, the retractable three-dimensional pallet includes a goods placement tray, four support rods, and a stacking surface; the top view of the goods placement tray is a quadrilateral; each of the support rods is vertically embedded in the four corners of the goods placement tray; each of the support rods can be extended or retracted through an adjustment switch in the middle of the support rod; the stacking surface is a plane formed by connecting the tops of the support rods.

3. The three-dimensional loading optimization method of the tray adjustable based on the support structure according to claim 2, wherein, The objective function of the three-dimensional loading model of the distribution goods pallet is: Among them, F3 is the objective function of the three-dimensional loading model for the distribution goods pallet; β is the cost conversion weight coefficient; F1 is the cost of the distribution vehicle; F2 is the average loading rate of all distribution vehicles; S is the total number of vehicles; k is the distribution vehicle number; K is the set of distribution vehicles; max(·) is the maximum value function; is the distribution cost of the k-th vehicle from customer point i to j within the service period T; c d is the distribution cost per unit distance during the vehicle distribution process; both i and j are customer point numbers; J is the set of customer points; is the decision variable indicating whether the k-th vehicle transports from customer point i to j within the service period t; t is the service period; T is the set of service periods; d ij is the distance from customer point i to j; is the vehicle usage cost; is the unit vehicle usage cost of the vehicle type used by the k-th vehicle within the service period t; is the decision variable indicating whether the k-th vehicle is used within the service period t; is the time cost that the vehicle needs to bear for arriving at the unloading site too early; is the unit time cost for arriving too early; is the starting point of the demand service time window of customer point j; is the time when the k-th vehicle arrives at the j-th customer point within the service period t; is the time cost that the vehicle needs to bear for arriving at the unloading site too late; is the unit time cost for arriving too late; is the ending point of the demand service time window of customer point j; is the time when the k-th vehicle arrives at the (j - 1)-th customer point within the service period t; is the unloading time of the unit retractable three-dimensional pallet; N is the set of all goods; is the number of retractable three-dimensional pallets of the k-th vehicle at the j-th customer point within the service period t; d (j-1)j is the distance from customer point (j - 1) to j; v is the vehicle driving speed; m is the retractable three-dimensional pallet number; M is the set of all retractable three-dimensional pallets; is the volume of the m-th retractable three-dimensional pallet of the k-th vehicle within the service period t; is the length of the carriage of the k-th vehicle within the service period t; is the width of the carriage of the k-th vehicle within the service period t; is the height of the carriage of the k-th vehicle within the service period t; n is the goods number; is the decision variable indicating whether the n-th piece of goods at the i-th customer point of the k-th vehicle is loaded onto the m-th retractable three-dimensional pallet within the service period t; is the length of the nth piece of cargo on the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the width of the nth piece of cargo on the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the height of the nth piece of cargo on the mth retractable three-dimensional tray of the kth vehicle within the service cycle t.

4. The three-dimensional loading optimization method of the tray adjustable based on the support structure according to claim 3, wherein The cost conversion weight coefficient is determined by the entropy method, which specifically includes the following steps: A1. Construct a set of solutions and a set of objectives according to the situations of using and not using pallets; A2. Establish an evaluation matrix based on the set of solutions and the set of objectives; Among them, is the evaluation matrix; O1 is the first objective; O2 is the second objective; P1 is the first solution; P2 is the second solution; r 11 is the evaluation value of the first objective corresponding to the first solution; r 12 is the evaluation value of the first objective corresponding to the second solution; r 21 is the evaluation value of the second objective corresponding to the first solution; r 22 is the evaluation value of the second objective corresponding to the second solution; is the evaluation value of the nth objective; is the nth solution; is the nth objective; is the total number of solutions; is the total number of objectives; A3. Obtain the entropy value according to the evaluation matrix; wherein, is the entropy value of the th target; κ is the reciprocal of the natural logarithm of ; is the sum of all evaluation values of the th target corresponding to the solution; ln is the natural logarithm; ​ A4. Obtain the cost conversion weight coefficient according to the entropy value; β = w1, (1 - β) = w2 Among them, β is the cost conversion weight coefficient; w1 is the cost conversion weight coefficient of the first objective; w2 is the cost conversion weight coefficient of the second objective; is the cost conversion weight coefficient of the th objective; is the degree of consistency of the contribution of the solution under the th objective.

5. The three-dimensional loading optimization method of the tray adjustable based on the support structure according to claim 1, wherein The constraint conditions of the three-dimensional loading model of the distribution goods pallet include: Among them, is the abscissa of the nth piece of cargo on the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the length of the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the length of the nth piece of cargo on the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the ordinate of the nth piece of cargo on the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the width of the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the width of the nth piece of cargo on the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the vertical axis coordinate of the nth piece of cargo on the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the height of the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the height of the nth piece of cargo on the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the abscissa of the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the length of the compartment of the kth vehicle within the service cycle t; is the ordinate of the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the width of the compartment of the kth vehicle within the service cycle t; is the vertical axis coordinate of the mth retractable three-dimensional tray of the kth vehicle within the service cycle t; is the height of the compartment of the kth vehicle within the service cycle t; x a is the abscissa of cargo a; x b is the abscissa of cargo b; l b is the length of cargo b; q xab is the decision variable for overlapping cargo a and cargo b in the X-axis direction; a and b are both cargo numbers; N is the set of all cargos; y a is the ordinate of cargo a; y b is the ordinate of cargo b; w b is the width of cargo b; q yab is the decision variable for overlapping cargo a and cargo b in the Y-axis direction; z a is the vertical axis coordinate of cargo a; z b is the vertical axis coordinate of cargo b; h b is the height of cargo b; q zab is the decision variable for overlapping cargo a and cargo b in the Z-axis direction; is a retractable three-dimensional tray is the horizontal axis coordinate of is a retractable three-dimensional tray is the horizontal axis coordinate of Within the service cycle t, for the k-th vehicle, the length of the k-th retractable three-dimensional tray For the retractable three-dimensional tray and the retractable three-dimensional tray is the decision variable for overlapping in the X-axis direction is the vertical axis coordinate of the retractable three-dimensional tray is the vertical axis coordinate of the retractable three-dimensional tray Within the service cycle t, for the k-th vehicle, the width of the k-th retractable three-dimensional tray For the retractable three-dimensional tray and the retractable three-dimensional tray is the decision variable for overlapping in the Y-axis direction is the vertical axis coordinate of the retractable three-dimensional tray is the vertical axis coordinate of the retractable three-dimensional tray Within the service cycle t, for the k-th vehicle, the height of the k-th retractable three-dimensional tray For the retractable three-dimensional tray and the retractable three-dimensional tray is the decision variable for overlapping in the Z-axis direction; k is the delivery vehicle number; K is the set of delivery vehicles; is the decision variable for determining whether the k-th vehicle serves customer point i within the service cycle t; t is the service cycle; T is the set of service cycles; i and j are both customer point numbers; J is the set of customer points; is the set of all transportation nodes; is the decision variable indicating whether the k-th vehicle transports from customer point i to j within the service cycle t; is the number of customer points served by the k-th vehicle within the service cycle t; is the weight of the goods carried by the k-th vehicle within the service cycle t; G is the maximum load capacity of the vehicle; is the decision variable indicating whether the k-th vehicle is used within the service cycle t; is the decision variable indicating whether the n-th item of goods in the m-th retractable three-dimensional tray of the k-th vehicle is loaded within the service cycle t; m is the retractable three-dimensional tray number; M is the set of all retractable three-dimensional trays; n is the goods number.​​​​ 6. The three-dimensional loading optimization method of the tray adjustable based on the support structure according to claim 1, wherein Step S3 is specifically: S301. Obtain vehicle information, three-dimensional information of goods, goods weight information, customer point coordinates, customer demand information, and demand time window; S302. Set the termination temperature and the maximum number of non-improving iterations; S303. According to the vehicle information, three-dimensional information of goods, customer point coordinates, customer demand information, termination temperature, and the maximum number of non-improving iterations, perform three-dimensional loading of the pallet to obtain the pallet loading results at each customer point; S304. According to the goods weight information, demand time window, customer point coordinates, and the pallet loading results at each customer point, perform three-dimensional loading of the carriage to obtain the optimal loading plan.

7. The three-dimensional loading optimization method of the tray adjustable based on the support structure according to claim 6, characterized in that, Step S303 is specifically: S3031. According to the three-dimensional information of goods, customer point coordinates, and customer demand information, arrange the goods to be loaded at the current customer point in descending order of volume to obtain an initial loading sequence; S3032. According to the vehicle information and the initial loading sequence, use the height of the carriage as the height of the retractable three-dimensional pallet and perform three-dimensional loading using the wall-building theory to obtain a feasible solution under the current loading sequence including loadable goods and unloadable goods; S3033. According to the feasible solution under the current loading sequence, termination temperature, and the maximum number of non-improving iterations, determine whether the termination condition is met. If so, perform annealing operation, and use the optimal solution in the feasible solution under the current loading sequence as the optimal loading solution, and enter step S3034. Otherwise, use the optimal solution in the feasible solution under the current loading sequence as the first optimal loading solution, and perform annealing operation, selection operation, crossover operation, and mutation operation, and return to step S3032; the termination condition is that the current annealing temperature is not greater than the termination temperature, or the number of non-improving times of the fitness function of the feasible solution under the current loading sequence compared with the fitness function of the first optimal loading solution reaches the maximum number of non-improving iterations; S3034. According to the optimal loading solution, determine whether the pallet loading rate of the current retractable three-dimensional pallet at the current customer point is less than the threshold. If so, lower the height of the retractable three-dimensional pallet and return to step S3032. Otherwise, record the specifications of the current retractable three-dimensional pallet and proceed to step S3035; S3035. Determine whether all the goods to be loaded at the current customer point have been loaded. If so, obtain the pallet loading result at the current customer point according to the specifications of each retractable three-dimensional pallet, and return to step S3031 to perform pallet loading for the next customer point. Otherwise, arrange the non-loadable goods in the optimal loading solution in descending order of volume to obtain a new loading sequence, update the initial loading sequence to the new loading sequence, and return to step S3032 to perform loading for the next retractable three-dimensional pallet.

8. The three-dimensional loading optimization method of the tray adjustable based on the support structure according to claim 6, wherein, The specific steps of S304 are as follows: S3041. Cluster according to the customer point coordinates to obtain several clustered customer groups; S3042. According to the vehicle information, randomly select a vehicle to load the current clustered customer group into the carriage; S3043. Obtain the pallet loading result of the current clustered customer group according to the pallet loading results of each customer point; S3044. Sort in descending order of volume according to the pallet loading result of the current clustered customer group to obtain an initial pallet loading sequence; S3045. According to the initial pallet loading sequence, use the wall-building theory for three-dimensional loading to obtain a feasible solution including loadable pallets and non-loadable pallets under the current pallet loading sequence; S3046. According to the feasible solution under the current pallet loading sequence, the termination temperature, and the maximum number of iterations without improvement, determine whether the carriage loading termination condition is satisfied. If so, perform the annealing operation, and use the optimal solution in the feasible solution under the current pallet loading sequence as the optimal pallet loading solution, and proceed to step S3047. Otherwise, use the optimal solution in the feasible solution under the current pallet loading sequence as the first optimal pallet loading solution, and perform the annealing operation, selection operation, crossover operation, and mutation operation, and return to step S3045; the carriage loading termination condition is that the current annealing temperature is not greater than the termination temperature, or the number of non-improved times reaches the maximum number of iterations without improvement when comparing the fitness function of the feasible solution under the current pallet loading sequence with the fitness function of the first optimal pallet loading solution; S3047. According to the optimal pallet loading solution, determine whether the loading rate of the current vehicle is less than the threshold. If so, select a smaller carriage for loading and return to step S3045. Otherwise, record the model of the current vehicle and proceed to step S3048; S3048. Determine whether all the pallets to be loaded in the current clustered customer group have been fully loaded. If so, obtain the carriage loading results of the current clustered customer group according to the specifications of each vehicle, and return to step S3042 to perform carriage loading for the next clustered customer group. Otherwise, arrange the non-loadable pallets in the optimal pallet loading solution in descending order of volume to obtain a new pallet loading sequence. Update the initial pallet loading sequence to the new pallet loading sequence. Randomly select a vehicle model as the vehicle model for the next loading, and return to step S3045 to perform loading for the next vehicle; S3049. Obtain vehicle loading information according to the carriage loading results of each clustered customer group; S30410. Perform path optimization for each vehicle respectively according to the cargo weight information, demand time window, and vehicle loading information to obtain the optimal loading plan.

9. The three-dimensional loading optimization method of the tray adjustable based on the support structure according to claim 8, characterized in that The specific steps of step S30410 are as follows: S304101. Obtain the customer point coordinates to which each retractable three-dimensional pallet in the current transport vehicle belongs according to the vehicle loading information; S304102. Randomly generate a path solution according to the customer point coordinates to which each retractable three-dimensional pallet in the current transport vehicle belongs; S304103. According to the path solution, determine whether the termination condition for path optimization is met. If so, use the result with the shortest path in the path solution as the optimal path result for the current vehicle, and return to step S304102 to perform path optimization for the next transport vehicle until the path optimization of all transport vehicles is completed to obtain a loading plan, and enter step S304104. Otherwise, use the result with the shortest path in the path solution as the first optimal path, and perform selection, crossover, and mutation iterations, and return to step S304102; The termination condition for path optimization is that the current annealing temperature is not greater than the termination temperature, or the number of times without improvement in the comparison between the optimal length of the current path solution and the first optimal path reaches the maximum number of non-improving iterations; S304104. According to the cargo weight information, demand time window, customer point coordinates, and loading plan, determine whether the overall plan termination condition is met. If so, obtain the optimal loading plan. Otherwise, use the loading plan as the original optimal plan, and return to step S3041 to perform iteration on the overall plan; the overall plan termination condition is that the current annealing temperature is not greater than the termination temperature, or the number of times without improvement in the comparison between the value of the objective function of the three-dimensional loading model of the distribution goods pallet based on the current loading plan and the value of the objective function of the three-dimensional loading model of the distribution goods pallet of the original optimal plan reaches the maximum number of non-improving iterations.

Images