Analysis method of an improved dynamic analysis model for regional vehicle route planning

Abstract

The present invention discloses an analysis method for an improved dynamic analysis model of regional vehicle route planning, which has the following steps: Select a distribution transfer center according to the existing order volume, divide the goods according to regions, and concentrate them in the corresponding distribution transfer centers. Subsequently, through the centralized planning of the goods in the transfer centers, use the 3L-CVRP combined optimization model to establish a loading model for the goods. In the third stage, perform route planning for the trucks that have successfully loaded goods. Because of the slime mold path optimization algorithm based on two factors of time window and material cost, it avoids the local optimum problem brought by traditional heuristic algorithms and effectively improves the transportation efficiency of the logistics center in distributing goods.

Description

Technical Field

[0001] The present invention relates to an analysis method for an improved dynamic analysis model of regional vehicle path planning, belonging to the technical field of vehicle transportation scheduling optimization. Background Art

[0002] In modern enterprises, with the advent of the era of intelligentization and mechanization, modern logistics includes advanced organizational management methods, intelligent science and technology, and diversified information means, which undoubtedly opens up new sources of profit for enterprises.

[0003] Among numerous logistics production activities, distribution is at the core and plays a crucial role in the logistics system. Distribution is the final link of logistics, realizing the ultimate allocation of resources and being an important link in modern logistics. Since the delivery destinations of goods are fixed, a reasonable vehicle-cargo assembly sequence can not only effectively improve the space utilization rate of vehicles, reduce and avoid space waste and overweight phenomena, ensure the safety of drivers, vehicles and goods, but also effectively reduce logistics costs and improve loading and transportation efficiency; while reasonable vehicle path optimization can effectively reduce the driving paths of vehicles, lower the time cost and material cost of distribution, improve vehicle utilization rate and enhance customer satisfaction of distribution. Nowadays, many logistics distribution enterprises only consider one aspect separately during the distribution process. However, the problems of cargo loading and path planning are closely linked and restrict each other. If only one aspect of the problem is considered alone, the logistics distribution cannot be optimized to the greatest extent. Therefore, to truly improve the efficiency and effectiveness of the distribution link, it is necessary to combine the key transaction nodes of the distribution link, fully consider and study the connections between the transaction nodes, and perform unified solutions. Summary of the Invention

[0004] To solve the problems existing in the prior art, the present invention proposes an analysis method for an improved dynamic analysis model of regional vehicle path planning. During the path planning of vehicles, a more scientific calculation of fuel consumption cost is carried out, considering the wind resistance of different vehicles and the friction resistance factors brought by the load, making the fuel consumption cost accounting further scientific and reasonable. Finally, the slime mold algorithm introducing a time window is used for path planning and solving, enhancing the practical applicability of the algorithm.

[0005] To achieve the above object, the technical solution adopted by the present invention is: an analysis method for an improved dynamic analysis model of regional vehicle path planning, characterized in that:

[0006] It includes the following specific steps:

[0007] Step 1: Establish a location model for the distribution transfer center, establish a minimum transportation construction cost function as the objective function for location selection, and solve the location selection problem of the distribution transfer center for regional distribution according to the cuckoo algorithm in combination with the proposed location model for the distribution transfer center;

[0008] Step 2: Regarding the loading problem of the carriage, the concept of the weight-bubble ratio is proposed, and the 3L-CVRP algorithm is improved. Instead of focusing on single-objective solution, the problem of loading goods into boxes is combined with the delivery destinations of the goods and their weight-bubble ratios to determine whether to load multiple goods and the loading order of the loaded goods. The influencing factors at least include the load, volume, and frontal area of the truck;

[0009] Step 3: Regarding the actual road conditions in the city, the vehicle congestion factor and the delivery time window limit are introduced, and an improved slime mold algorithm is established. On the basis of the existing algorithm, a time review model is established. By combining the congestion time of the road with the ideal driving time, the final actual arrival time is obtained, which is further compared with the time window, and the paths that exceed the time limit are deleted. Under the condition of meeting the preset delivery time, a dynamic and changing distribution path that matches the actual situation is generated.

[0010] Further, in the above Step 1, establishing the location model for the distribution transfer center at least involves the following factors: civil engineering costs, transportation costs, management costs, and cost accounting.

[0011] The civil engineering costs include: infrastructure construction costs and resettlement costs;

[0012] The transportation costs include: fuel consumption costs, vehicle depreciation costs, driver commissions, and toll fees from the general warehouse to the distribution transfer center and from the distribution transfer center to the specific demand locations;

[0013] The management costs include: ground rent costs, fixed costs for configuring the management system, office supplies expenses, office staff expenses, and equipment purchase expenses;

[0014] Therefore, the final transportation cost function is:

[0015]

[0016] Among them, m: the number of alternative logistics centers, n: the number of demand points, Z: the total cost of building the distribution transfer center, K i : the civil engineering costs of the distribution transfer center, H i : the management costs of the distribution transfer center, D i : the distance from the general warehouse to the distribution transfer center, T i : the freight from the general warehouse to the distribution transfer center, in tons·kilometers, D ij : the distance from the alternative distribution transfer center to the demand point, hi : Alternative delivery transfer center shipment volume, P i : Shipping volume from the main warehouse to each delivery transfer center, Q ij : Shipping volume from the breeding transfer center to the demand point.

[0017] Further, in step 2, the specific steps of the improved 3L-CVRP algorithm are as follows:

[0018] Classify the goods from a three-dimensional perspective according to the actual weight and volume of the goods. They are respectively "heavy goods" and "bulky goods", and the two adopt different pricing methods. The vehicle constraints are calculated according to the classification of heavy and bulky goods. Finally, the constraint condition regarding the vehicle load is that both heavy goods and bulky goods are carried on the vehicle, achieving the goal of minimizing the cost under the condition of carrying heavy and bulky goods, avoiding the problem that the goods are not fully loaded due to too many heavy goods, and also avoiding the problem that even if the cargo box is full, the cost cannot be recovered due to too low load capacity;

[0019] The volume weight of heavy and bulky goods can be determined by the following formula,

[0020] Volume weight = (length cm × width cm × height cm) ÷ 5000 cm 3 / kg

[0021] According to the result calculated by the formula, goods with a weight greater than 200 kg per cubic meter are heavy goods, and those less than 200 kg are bulky goods; bulky goods are calculated according to the space occupied by the goods, and heavy goods are calculated according to the actual weight of the goods.

[0022] Further, the final transportation revenue of a single truck is: introducing the labor transportation cost, the final net profit function of a single truck's distribution is the transportation revenue of the two parts of goods minus the fuel consumption cost, labor transportation cost, toll loss, and depreciation loss, and its formula is expressed as follows:

[0023] maxW i =g i p+v j q-Q s ·y-S-R-f

[0024] Among them, g i is the weight of goods i; v i is the volume of the goods; p is the price per ton for transporting heavy goods; q is the price per cubic meter for transporting bulky goods; W i represents the revenue of a single truck; Q s is the fuel consumption per 100 kilometers; y is the current oil price; S is the driver's transportation fee; R is the toll loss; f is the vehicle depreciation loss.

[0025] Furthermore, a multi-vehicle loading model that conforms to actual transportation is established in the improved 3L-CVRP algorithm, breaking through the constraints of a single objective, establishing a multi-objective constraint model. Based on the concept of heavy and light goods, the order of placing goods is controlled. Heavy goods are placed at the bottom of the truck, and light goods are placed on top of the heavy goods. The key constraint conditions are as follows:

[0026] The ratio of the weight to volume of the u-th item of the i-th customer in the k-th vehicle of type M;

[0027] a: Support surface coefficient;

[0028] When At this time, x1>x2, y1>y2, H(j, v) ∈ B.

[0029] Furthermore, in step 3, a slime mold algorithm with a dynamic time window is introduced to solve the optimal path planning, and the specific steps are as follows:

[0030] Step 3.1. Initialization: Establish a two-dimensional plane coordinate system for the urban roads, and at the same time calibrate the starting points, passing points, and ending points of each order through coordinates. The starting point of each intersection in the path is set to 0;

[0031] Step 3.2. Calculate the pressure values of each intersection;

[0032] Step 3.3. Calculate the traffic flow of each road,

[0033] Assume that the pressures of the delivery points I and J are p i 、p j , the path length connecting the two delivery points is L ij , its width is r ij , and the overall traffic flow is Q ij ; Assume that the flow velocity is uniform and the flow pattern is laminar, then we get:

[0034]

[0035] η is the viscosity coefficient, πr 4 / 8ηL ij is used to measure the conductivity of the conduit;

[0036] Substitute the pressure values obtained above into the following formula to solve the vehicle passing flow of each section of the road;

[0037]

[0038] Step 3.4. Calculate the street conductivity of the next stage,, and the average value of the node pressures After substituting into the above formula and combining with Formula 3, calculate the traffic flow of each road in the next stage.

[0039]

[0040] Step 3.5: Conduct iterative judgment. If there is a road R ij satisfying the following formula, the iteration ends, and further judgment is made on the time window. If the following inequality is not satisfied, continue the iteration, return to Step 3, and increment the iteration count by one;

[0041] |D ij (N + 1) - D ij (N)| ≤ 10 -3

[0042] Step 3.6: Calculate the time for the currently obtained optimal path and check whether it is within the required time window. If the destination cannot be reached within the specified time, this path cannot be output as the optimal path. It is necessary to return to Step 3, exclude the road when the first output result is obtained, and select the sub-optimal path for time window verification until the time window requirement is met.

[0043]

[0044] Step 3.7: End the calculation and obtain the optimal delivery path corresponding to this order.

[0045] Furthermore, in Step 3.2, the pipeline flow in the slime mold is fused with the vehicle congestion status of the road. The pipeline flow represents the vehicle congestion situation. If the pipeline flow is large, that is, the road is relatively unobstructed and the conductivity is large, it means that there are more passing people and it belongs to an economical and applicable route;

[0046] The road conductivity is simplified to:

[0047]

[0048] Among them, D i,j is the conductivity, Q i,j is the flow rate, L i,j is the road length between two points in a single block;

[0049] Combined with the flow factor in the algorithm, regardless of where the initial state of the slime mold algorithm starts, the conductivity of the shortest path will definitely converge to 1, and the conductivity of the non-shortest path will converge to 0.

[0050] Compared with the prior art, the beneficial effects of the present invention are:

[0051] 1. This invention breaks through the limitation of only considering a single problem in the domestic cargo loading and path planning problems. It establishes a cost accounting model for the site selection from the distribution center to the distribution transit center, and uniformly calculates the costs required to establish a transit center in reality. At the same time, the cost from the distribution transit center to the actual distribution point is estimated according to the regional order volume, and the line between the general warehouse, transit center and distribution point is opened up, and the cost is calculated in an integrated manner, making the site selection of the distribution transit center more scientific and reasonable.

[0052] 2. The present invention introduces the cost of fuel consumption and vehicle wear and tear in the cost accounting of vehicle distribution. Compared with simply estimating fuel consumption based on distance, the fuel consumption cost in the model updates the profit function for different vehicle models and different load masses. The updated profit model classifies and charges heavy goods and bulk goods. While reasonably allocating the ratio and packing location of heavy goods and bulk goods, it reduces vehicle losses.

[0053] 3. The present invention improves the recently popular slime mold algorithm in the path planning link, combines the traffic flow conditions on the road, and adds a time window restriction, which helps to avoid delays and losses caused by untimely delivery in real life and reduce the transportation losses of enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a flow chart of the present invention.

[0055] Figure 2 This is a flowchart of obtaining the optimal path after adding the concept constraint of the time window in the present invention. DETAILED DESCRIPTION

[0056] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0057] An analysis method of an improved dynamic analysis model for regional vehicle route planning proposed in this embodiment disassembles the goods distribution transportation problem into three main processes through modularization: the selection of the goods transfer center, the goods loading problem, and the goods distribution route planning problem. The three problems are solved separately using unified order data, and finally, a transportation cost model is established based on the unified vehicle loading result, and the slime mold algorithm based on the time window is used to solve the loading route, so as to achieve the optimal route planning solution. Moreover, the present invention takes more detailed considerations of the actual environment, assigns different weights to the road types around the location of the goods transfer center, models the road surface traffic conditions and the time window of customer orders, and can solve a distribution plan that is more in line with the actual situation.

[0058] It includes the following steps:

[0059] Step 1: Based on past order data, establish a location model for the distribution transfer center, analyze and model according to the three major factors involved in the location: civil engineering costs, transportation costs, and management costs, build a minimum construction cost function, and at the same time, weight the final cost function according to the road conditions and infrastructure near the distribution transfer center. The final cost function combines the improved cuckoo algorithm to calculate the location of the distribution transfer center.

[0060] Among them, the specific refinement factors of the three major factors are as follows:

[0061] Civil engineering costs include: infrastructure construction, resettlement costs;

[0062] Transportation costs include: fuel consumption costs, vehicle depreciation costs, driver commissions, and tolls from the general warehouse to the distribution transfer center and from the distribution transfer center to the specific demand location;

[0063] Management costs include: ground rent costs, fixed costs of configuring the management system, office supplies expenses, office staff expenses, and equipment purchase expenses.

[0064] Finally, the total construction cost of a single distribution transfer center can be expressed as:

[0065]

[0066] Take the surrounding environment of the distribution transfer center into consideration, introduce a weighting factor into the above formula (1), and the specific content is shown in Table 1

[0067] Table 1: Cost impact weights of main infrastructure on the location of the distribution center

[0068]

[0069] The final weight is:

[0070] Step 2: In the link of vehicle cargo distribution, combined with the actual situation, aiming at the load problem of the vehicle, calculate the load, volume and frontal area of trucks of different models. On the basis of the above data, establish a three-dimensional packing model. Goods are usually divided into "heavy goods" and "bulky goods", and different pricing mechanisms and placement orders should be adopted for the two in the distribution process. Bulky goods occupy a large space, while heavy goods will have a certain limit on the specified loading weight of the vehicle. Therefore, in the process of distribution and loading, the concept of "heavy and bulky goods" is introduced in this embodiment. Based on the actual situation of land transportation, the main formula is as follows:

[0071] Volume weight = (length in cm × width in cm × height in cm) ÷ 5000 cm 3 / kg

[0072] For irregular items, the above length, width and height are calculated according to the longest, highest and widest data.

[0073] According to the result calculated by the formula, goods with a weight of more than 200 kg per 1 cubic meter are called heavy goods, and those less than 200 kg are called bulky goods. Generally, they are defined as "bulky goods", and the pricing is charged according to the space occupied by the goods. On the contrary, they are "heavy goods", and the pricing is charged according to the actual weight of the goods.

[0074] Based on the calculation results of heavy and bulky goods, the final single-vehicle distribution revenue function is further improved in the ordinary revenue function part, where the fuel consumption cost Q s combines the calculation results of heavy and bulky goods,

[0075]

[0076]

[0077]

[0078] The depreciation loss f is shown in Table 2:

[0079]

[0080] Table 2

[0081] The final single-vehicle transportation revenue is: the sum of the heavy goods fee and the bulky goods fee minus the labor cost, fuel consumption cost, depreciation loss and toll fee.

[0082] maxW i =g i p+v j q-Q s ·y-S-R-f

[0083] Meanwhile, for the delivery objective function of bicycles, the load capacity of the vehicle and the volume utilization rate are maximized. The objective functions for both are as follows:

[0084] Maximize the load utilization rate of all vehicles:

[0085] Maximize the volume utilization rate of all vehicles:

[0086] Secondly, based on the already allocated packing results, the packing order of the goods is further planned, and a multi - vehicle packing model that matches the actual transportation is established. At the same time, based on the concept of heavy and bulky goods, the order of placing the goods is controlled. Heavy goods are placed at the bottom of the truck, and bulky goods are placed on top of the heavy goods. The key constraints are:

[0087] The weight - to - bulk ratio of the u - th item of the i - th customer on the k - th vehicle of type M;

[0088] When then

[0089] 0 ≤ x kiu , 0 ≤ y kiu , 0 ≤ z kiu ,

[0090] ( x kiu , y kiu , z kiu ) represents the coordinates of the lower left - rear corner of the u - th item in the carriage of the i - th order on the k - th vehicle

[0091] represents the coordinates of the upper right - front corner of the u - th item in the carriage of the i - th customer on the k - th vehicle

[0092] A kiu 、B kiu 、H kiu : represent the length, width, and height of the u - th item of the i - th order on the k - th vehicle;

[0093] A, B, H: represent the length, width, and height of the carriage;

[0094] Among them, H(i, v) ∈ B represents the set of all goods.

[0095] Calculate the total delivery weight of the packed goods, plug the result into the fuel consumption formula, digitally measure the fuel consumption of individual vehicles and vehicle wear and tear, and calculate the cost of a single transport.

[0096] maxW i =g i p+v j qQ s ·ySR

[0097] Among them, S is the driver's transportation cost, R is the toll loss, and f is the vehicle depreciation loss.

[0098] Under the condition of lowest cost, the objective function constraints of vehicle loading volume and weight are added to form a multi-objective packing constraint model.

[0099] Step 3: According to the actual situation of the road, the congestion factors, road conditions and delivery time window are taken into consideration, and a new improved slime mold algorithm is established. By calculating the "flow" size of the road, the travel time and conditions of the road, a dynamically changing delivery path that is consistent with the actual situation is generated, such as Figure 2 shown.

[0100] The existing delivery route is checked and calculated to see if it is within the time window scheduled by the customer. If it cannot be delivered within the specified time, the route is not considered the optimal route, and a route with the second-best cost needs to be selected from the iteration to continue the time window test until the planned route can deliver the goods to the destination within the specified time. The specific process is as follows:

[0101] Step 3.1: Establish a two-dimensional plane coordinate system for urban roads, and calibrate the starting point, passing point, and end point of each order with coordinates. Among the possible paths, the initial conductivity value of each road is 1, the pipeline flow is 0, the starting pressure of each intersection is set to 0, and the number of iterations is N = 0.

[0102] Step 3.2: Calculate the pressure value of each intersection, combine the congestion status of the road in reality with the algorithm, and digitize the congestion degree of the road into the pipeline flow in the slime mold algorithm. If the road is in good condition, vehicles can pass quickly in a short time, and there are many lanes available on the road, then the conductivity of the pipeline is good and the conductivity index is large. At the same time, the conductivity also indicates that there are many vehicles passing through this section of the road, which is an economical route. The conductivity of the road can be expressed as

[0103]

[0104] Where D i,j is conductivity, Q i,j is the flow rate, Li,j The length of the road directly between two points in a single block, γ is the extinction rate of the pipeline, which represents the frequency of passing through this section of the road in the actual path. f is a monotonically increasing function, and f(0) = 0.

[0105] Combined with the flow factor in the algorithm, regardless of where the initial state of the slime mold algorithm starts, the conductivity of the shortest path will converge to 1, and the conductivity of the non - shortest path will converge to 0.

[0106] Substitute the current conductivity and road length into the following formula to calculate the pressure values of each intersection node

[0107]

[0108] Step 3.3: Calculate the flow of each road. Substitute the pressure values obtained in the second step above into the following formula to solve the vehicle passing flow of each section of the road.

[0109]

[0110] Step 3.4: Calculate the road conductivity of the next stage. After substituting the average value of the node pressure into the above formula and combining with Formula 3, calculate the flow of each road in the next stage.

[0111]

[0112] Step 3.5: Conduct an iterative judgment. If there is a road R ij that satisfies the following formula, the iteration ends, and further judge the time window. If the following inequality is not satisfied, continue the iteration, return to Step 3, and increment the iteration count by one

[0113] |D ij (N + 1)-D ij (N)|≤10 -3

[0114] Step 3.6: Calculate the time for the currently obtained optimal path to detect whether it is within the required time window. If it cannot reach the destination within the specified time, this path cannot be output as the optimal path. It is necessary to return to Step 3, exclude the roads in the first output result, select the sub - optimal path for time window verification until the time window requirement is met.

[0115]

[0116] Step 3.7: End the calculation and finally obtain the optimal delivery path corresponding to this order.

[0117] The symbol meanings in this embodiment are as follows:

[0118] m: The number of alternative logistics centers

[0119] n: The number of demand points

[0120] Z: The total cost of building distribution transfer centers

[0121] K i : The civil engineering cost of the distribution transfer center

[0122] H i : The management cost of the distribution transfer center

[0123] D i : The distance from the general warehouse to the distribution transfer center

[0124] T i : The freight (ton·kilometer) from the general warehouse to the distribution transfer center

[0125] D ij : The distance from the alternative distribution transfer center to the demand point

[0126] h i : The shipment volume of the alternative distribution transfer center

[0127] P i : The quantity of goods transported from the general warehouse to each distribution transfer center

[0128] Q ij : The quantity of goods transported from the breeding transfer center to the demand point

[0129] E: The influence of the surrounding environment factor

[0130] g i : The weight of cargo i;

[0131] v i : The volume of the cargo;

[0132] p: The price per ton for transporting heavy goods;

[0133] q: The price per cubic meter for transporting light goods;

[0134] W i : Represents the income of a single truck;

[0135] Q s : The fuel consumption per 100 kilometers

[0136] y: The current oil price

[0137] S: The driver's transportation cost

[0138] R: The toll loss

[0139] f: The vehicle depreciation loss

[0140] Q S : Indicates fuel consumption per 100 kilometers

[0141] G a : Indicates vehicle weight

[0142] U a : Indicates vehicle speed

[0143] p: Indicates the resistance power of vehicle nature

[0144] g e : Indicates fuel consumption rate

[0145] γ: Indicates fuel specific weight

[0146] η T : Indicates the mechanical efficiency of the system

[0147] Indicates the resistance coefficient of the road

[0148] C D : Indicates the air resistance coefficient

[0149] A: Indicates the vehicle frontal area

[0150] The weight-bubble ratio of the u-th item of the i-th customer of the k-th vehicle of type M;

[0151] a: Support surface coefficient

[0152] D i,j : Is conductivity

[0153] Q i,j : Is the flow rate

[0154] L i,j : The direct road length between two points in a single block

[0155] Based on the existing domestic research, this embodiment is more in line with actual factors, breaks through the limitations of research, and combines the three from the overall perspective of the site selection of the logistics distribution transfer center, vehicle loading, and path planning, forming an improved dynamic analysis model for regional vehicle path planning. The model makes a more detailed division of the site selection factors of the distribution transfer center and forms a new cost accounting model.

[0156] In the vehicle path planning, the fuel consumption cost is calculated more scientifically, considering the windward resistance of different vehicles and the friction resistance factor caused by the load, making the fuel consumption cost accounting more scientific and reasonable. Finally, the slime mold algorithm with time window is introduced to solve the path planning, enhancing the practical applicability of the algorithm.

[0157] It should be noted that the above embodiments are illustrative of the present invention rather than restrictive of the present invention, and those skilled in the art can design alternative embodiments without departing from the scope of the appended claims.

Claims

1. An analysis method for an improved dynamic analysis model of regional vehicle route planning, characterized in that: It includes the following specific steps: Step 1: Establish a location model for the distribution transfer center. The location model for the distribution transfer center involves at least the following factors: civil engineering costs, transportation costs, management costs, and conduct cost accounting; The civil engineering costs include: infrastructure construction costs and resettlement costs; the transportation costs include: fuel consumption costs, vehicle depreciation costs, driver commissions, and toll fees from the general warehouse to the distribution transfer center and from the distribution transfer center to the specific demand location; the management costs include: ground rent costs, fixed costs for configuring the management system, office supplies expenses, office staff expenses, and equipment purchase expenses; Therefore, the final transportation cost function is: Among them, m: the number of alternative logistics centers, n: the number of demand points, Z: the total cost of building distribution transfer centers, K i : the civil engineering cost of the distribution transfer center, H i : the management cost of the distribution transfer center, D i : the distance from the general warehouse to the distribution transfer center, T i : the freight from the general warehouse to the distribution transfer center, in tons·kilometers, D ij : the distance from the alternative distribution transfer center to the demand point, h i : the shipment volume of the alternative distribution transfer center, P i : the volume of goods transported from the general warehouse to each distribution transfer center, Q ij : the volume of goods transported from the breeding transfer center to the demand point; Establish a minimum transportation construction cost function as the objective function for site selection. According to the cuckoo algorithm, combined with the proposed location model for the distribution transfer center, solve the problem of locating the distribution transfer center for regional distribution; Step 2: For the problem of vehicle loading, the concept of the weight-to-volume ratio is proposed, and the 3L-CVRP algorithm is improved. Combine the problem of loading goods into boxes with the delivery destinations of the goods and their weight-to-volume ratios to determine whether to load multiple goods and the loading order of the loaded goods; the influencing factors include at least the load capacity, volume, and frontal area of the truck; The specific steps of the improved 3L-CVRP algorithm are: Classify the types of goods according to the actual situation, and establish a new cost accounting model in three dimensions, that is, classify the goods according to the weight-to-volume ratio. By calculating the ratio of the actual weight to the volume of the goods, implement a separate pricing mechanism for heavy goods and light goods; The constraints of vehicle loading are calculated according to the classification of heavy and light goods. Finally, the constraint is proposed: in each carriage, heavy goods and light goods need to exist at the same time, to avoid the problem that the goods are not fully loaded due to too many heavy goods, and also to avoid the problem that the cost cannot be recovered due to too low load capacity even if the cargo box is full; The volume weight of heavy and light goods is determined by the following formula, Volumetric weight = (Length in cm × Width in cm × Height in cm) ÷ 5000 cm 3 / kg According to the results calculated by the formula, goods with a weight of more than 200 kilograms per cubic meter are heavy goods, and those less than 200 kilograms are light goods; light goods are calculated according to the space occupied by the goods, and heavy goods are calculated according to the actual weight of the goods; Step 3: Considering the actual road conditions in the city, taking into account the factors of vehicle congestion and the time window limit for cargo distribution, establish an improved slime mold algorithm; on the basis of the original algorithm, add the verification of the vehicle arrival time. If the vehicle cannot arrive within the expected time, the algorithm will discard the existing results and calculate and judge the sub-optimal path until it meets the expected time of distribution, so as to generate a dynamic distribution path that matches the actual situation; Step 3.1: Initialization: Establish a two-dimensional plane coordinate system for the urban roads, and at the same time calibrate the starting points, passing points, and ending points of each order through coordinates. The starting points of each intersection in the path are set to 0; Step 3.2: Calculate the pressure values of each intersection; Step 3.3: Calculate the traffic flow of each road, Suppose the pressures at delivery points I and J are p i and p j respectively. The path length connecting the two delivery points is L ij and its width is r ij . The overall flow rate is Q ij ; assuming that the flow velocity is uniform and the flow pattern is laminar, we obtain: η is the viscosity coefficient, and πr 4 / 8ηL ij is used to measure the conductivity of the catheter; Substitute the already obtained pressure values into the following formula to solve the vehicle passing flow of each section of the road; Step 3.4: Calculate the street conductivity of the next stage, and substitute the average value Q of the node pressure ij into the above formula to calculate the traffic flow of each road in the next stage. Step 3.5, perform iterative judgment. If there is a road R ij that satisfies the following formula, the iteration ends, and further judgment is made on the time window. If the following inequality is not satisfied, continue the iteration, return to Step 3, and increment the iteration count by one; |D ij (N + 1)-D ij (N)|≤10 -3 Step 3.6: Calculate the time of the currently obtained optimal path and check whether it is within the required time window. If the destination cannot be reached within the specified time, this path cannot be output as the optimal path. It is necessary to return to Step 3, exclude the roads in the first output result, select the sub-optimal path for time window verification until the time window requirement is met; Step 3.7: End the calculation and obtain the optimal delivery path corresponding to this order.

2. The analysis method for the improved dynamic analysis model of regional vehicle route planning according to claim 1, characterized in that: The final transportation revenue of a single truck is as follows: The artificial transportation cost is introduced. The net profit function of the delivery of a single truck is the transportation revenue of the two parts of goods minus the fuel consumption cost, the artificial transportation cost, the toll loss, and the depreciation loss. The formula is expressed as follows: maxW i = g i p + v j q - Q s · - y - S - R - f where g i is the weight of cargo i; v i is the volume of the cargo; p is the price per ton for transporting heavy cargo; q is the price per cubic meter for transporting light cargo; W i represents the revenue of a single truck; Q s is the fuel consumption per 100 kilometers; y is the current oil price; S is the driver's transportation cost; R is the toll loss; f is the vehicle depreciation loss.

3. The analysis method for the improved dynamic analysis model of regional vehicle route planning according to claim 1, characterized in that: In the improved 3L-CVRP algorithm, a multi-type vehicle loading model that conforms to actual transportation is also established. At the same time, based on the concept of heavy and light goods, the order of placing goods is controlled. Heavy goods are placed at the bottom of the truck, and light goods are placed on top of the heavy goods. The key constraints are: The weight-bubble ratio of the u-th item of the i-th customer of the k-th vehicle of type M; a: Support surface coefficient; When then x1 > x2, y1 > y2, H(j, v) ∈ B.

4. The analysis method of the improved regional vehicle path planning dynamic analysis model according to claim 1, characterized in that: In Step 3.2, the pipeline flow in the slime mold is fused with the vehicle congestion state on the road. The pipeline flow represents the vehicle congestion situation; The road conductivity after simplification is: Among them, D i,j is conductivity, Q i,j is the flow rate, L i,j is the road length directly between two points in a single block; Combined with the flow factor in the algorithm, no matter where the initial state of the slime mold algorithm starts, the conductivity of the shortest path will definitely converge to 1, and the conductivity of the non-shortest path will converge to 0.

Images