A heterogeneous vehicle intelligent scheduling method for large city supermarket distribution

Abstract

The application discloses a kind of heterogeneous vehicle intelligent scheduling methods for large city supermarket distribution, belong to logistics field, the intelligent scheduling method is established by step one scheduling optimization model clear optimization goal, determines vehicle scheduling optimization variable, fully considers the influence of heterogeneous vehicle to total cost, and the scope of application is greater than traditional logistics optimization algorithm.Step two gives the various possibilities of the shortest path distance after order determination, establishes foundation for later path optimization.Step three utilizes the multiverse algorithm improved by differential evolution algorithm to obtain optimal scheduling scheme, and the algorithm has fast convergence speed, high convergence accuracy and good robustness, effectively improves the efficiency and economy of large supermarket supermarket distribution, and realizes the rationalization of city scheduling.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of logistics, and more particularly relates to a heterogeneous vehicle intelligent scheduling method for large city supermarket distribution. BACKGROUND

[0002] With the change of people's lifestyle, the rapid development of e-commerce and the rapid rise of urban freight, intelligent scheduling has gradually become the development direction and research focus of the city logistics industry. As an important part of city logistics, large city supermarket distribution based on manual operation and traditional communication means scheduling mode needs to be changed urgently.

[0003] The traditional scheduling system mainly uses two types of algorithms: one is the exact algorithm; the second is the heuristic algorithm. At present, the related research only considers the shortest distance of vehicle scheduling from the traveling salesman problem, and cannot consider the transportation cost of different heterogeneous vehicles and other problems, resulting in differences between the scheduling scheme and the actual application, and unable to meet the actual needs of large city supermarket distribution.

[0004] Due to the characteristics of large city supermarket distribution such as large flow, multiple flow direction, many variable nodes, wide distribution, small radius, etc., the constraint conditions are more complex than general vehicle scheduling, and the related research is less, so it is very necessary to have a heterogeneous vehicle intelligent scheduling system for city supermarket distribution. SUMMARY

[0005] Therefore, the present application provides a heterogeneous vehicle intelligent scheduling system for large city supermarket distribution to solve the problems of low standardization level in large city supermarket distribution, low customer satisfaction in the distribution process, and low distribution efficiency. The system designs and establishes a heterogeneous vehicle scheduling optimization model based on city supermarket distribution with clear objective function, controllable decision variables and complete constraint conditions under the current city supermarket distribution scheduling mode, so as to improve the efficiency and economy of large supermarket distribution, and realize the rationalization of city scheduling.

[0006] In order to achieve the above purpose, the present application adopts the following technical scheme:

[0007] Step one, establish a heterogeneous vehicle scheduling optimization model based on city supermarket distribution;

[0008] Step two, according to the order requirement, set the number of vehicle passing through the receiving point interval, use Hamilton loop algorithm to calculate the shortest distance of each path, match the shortest path distance with each vehicle to generate order distribution set, and take the order distribution set as the initial population;

[0009] Step three, using the improved differential evolution algorithm mixed with the multi-universe algorithm to optimize the initial population, search for the optimal scheduling scheme.

[0010] Further, the heterogeneous vehicle scheduling optimization mathematical model based on city supermarket distribution in step one is as follows:

[0011] Objective function:

[0012] (1)

[0013] (1) In the formula is the fixed total cost of all vehicles, is the total cost of all vehicle transportation, is the total cost of all drivers' wages, is the utilization rate of all vehicles, is a 0-1 variable, indicating whether vehicle A is in operation, 1 for operation and 0 for non-operation;

[0014] Constraint conditions: vehicle maximum load constraint, vehicle maximum distribution point constraint, vehicle travel time window limit.

[0015] Further, the objective function calculation formula is as follows:

[0016] Vehicle fixed cost of vehicle:

[0017] (2)

[0018] (2) In the formula, indicates the vehicle model;

[0019] indicates the vehicle code;

[0020] indicates the starting price of the vehicle with model ;

[0021] indicates the point fee of the vehicle with model ;

[0022] is a 0-1 variable, indicating whether vehicle A reaches the delivery address , 1 for reaching and 0 for not reaching;

[0023] Fixed total cost of all vehicles ( ):

[0024] (3)

[0025] Vehicle transportation cost of vehicle ):

[0026] (4)

[0027] (4) In the formula, since the vehicles are divided into bulky goods and heavy goods according to their loading attributes, the bulky-to-weight ratio of a certain vehicle is set as the ratio of the weight to the volume of the goods. To indicate;

[0028] (5)

[0029] (5) In the formula Indicates vehicle Delivered to the delivery address The weight of the goods;

[0030] Indicates vehicle Delivered to the delivery address The volume of the goods;

[0031] For vehicles When the loaded goods are heavy goods ( The price required for the vehicle to transport a unit weight of goods per unit distance.

[0032] For vehicles When the loaded goods are bulky goods ( The price required for the vehicle to transport a unit weight of goods per unit distance.

[0033] Indicates the distance between each receiving point;

[0034] Indicates the distance between the distribution center and the receiving point;

[0035] Total transportation cost of all vehicles ( ):

[0036] (6)

[0037] Total cost of wages for all drivers ( ):

[0038] (7)

[0039] (5) In the formula, G represents the wages paid to workers per unit of time;

[0040] Indicates model number vehicles Driving on the middle section of the entire delivery route Time consumed;

[0041] Vehicle loading utilization rate ):

[0042] (8)

[0043] (8) wherein, denotes the weight of the goods delivered to the delivery address;

[0044] denotes the volume of the goods delivered to the delivery address;

[0045] Loading utilization rate of all delivery vehicles ):

[0046] (9)

[0047] Further, the constraint condition is calculated as follows:

[0048] Vehicle maximum loading capacity constraint:

[0049] (10)

[0050] (11)

[0051] (10), (11) wherein, denotes the weight of the goods delivered to the delivery address;

[0052] denotes the volume of the goods delivered to the delivery address;

[0053] denotes the maximum weight of the goods delivered by the vehicle;

[0054] denotes the maximum volume of the goods delivered by the vehicle;

[0055] Vehicle maximum delivery point constraint:

[0056] (12)

[0057] (12) wherein is the total number of delivery vehicles;

[0058] is the total number of delivery addresses;​​​​

[0059] Vehicle travel time window limit:

[0060] (13)

[0061] (13) wherein is the average delivery time, is the total delivery distance to the delivery address, is the average delivery speed, represents the vehicle delivery to the delivery address the weight of the goods; is the loading and unloading speed; is the maximum set time for delivery to the delivery address. Further, the second step is specifically implemented as follows:

[0062] Further, the second step is specifically implemented as follows:

[0063] Step 2.1, set the total number of orders and order number, set the maximum number of vehicles that can be assigned by the distribution center ;

[0064] Step 2.2, calculate the total number of orders , generate a random integer between 1 and for each order , representing the vehicle carrying this order;

[0065] Step 2.3, initialize the city matrix to obtain the distance between each delivery point, set the number of delivery points passed by the vehicle, and use the Hamilton circuit algorithm to calculate the shortest distance of each path;

[0066] Step 2.4, assign each shortest path distance to each vehicle to form an order distribution set, add orders that meet the constraint conditions to the initial population until the number of individuals in the initial population reaches the upper limit of the population size. Further, the Hamilton circuit is specifically calculated as follows:

[0067] From any starting point, pass through all other nodes, and finally return to the starting point, let the vertex , edge , be the distance between point , , and define the decision variable:

[0068] (14)

[0069] (14) wherein, 1 represents a point not passed through, and 0 represents a point already passed through; it is determined that each edge can only be accessed once;​

[0070] Objective function:

[0071] (15)

[0072] (15) wherein, is the position of each receiving point, is the sum of the minimum distance of Hamiltonian circuit;

[0073] Constraint condition:

[0074] (16)

[0075] (16) indicates that the symmetric circuit is;

[0076] (17)

[0077] (17) indicates that all edges are visited only once;

[0078] (18)

[0079] (18) indicates that the total length of the Hamiltonian circuit composed of N receiving points, is the position of the distribution center, is the position of the th receiving point, is the starting receiving point, is the ending receiving point.

[0080] Further, step 3.1, initialize algorithm parameters: including the maximum number of iterations T; The maximum number of chaotic searches S; The number of universes NP; The dimension of the universe D; Wormhole existence probability WEP; Travel distance rate TDR; TDR related constant C and p; Upper limit of search space ; Lower limit of search space ; Initial population ;

[0081] Step 3.2, update the global extreme value and individual extreme value of the multiverse, set the optimal individual of the initial universe as the global extreme value; If the individual universe does not satisfy the constraint condition, set the expansion rate of the universe to negative infinity or artificially change the matter of the universe;

[0082] Step 3.3, for individual universes, generate a random number between 0 and 1, if is less than the standard expansion rate of the individual, select the universe that produces a black hole according to the roulette method, and exchange matter between the selected universe and the roulette selected universe;

[0083] (19)

[0084] (19) where, is the kth variable of the ith universe; is the standard expansion rate of the individual universe; is a random number between 0 and 1;

[0085] Step 3.4, the universe receives the material sent from the optimal universe;

[0086] If is greater than the standard expansion rate of the universe, a random number between 0 and 1 is generated If is less than the wormhole existence probability WEP, the optimal universe sends material to the universe through the wormhole;

[0087] A random number between 0 and 1 is generated , which will determine the search direction of the universe;

[0088] The mutation operator in the differential evolution algorithm is used to further improve the operation of following the optimal universe search;

[0089] (20)

[0090] Save the universe after following the optimal universe search;

[0091] Step 3.5, chaotic search

[0092] If the of the universe is greater than the wormhole existence probability WEP, chaotic search is performed on the universe;

[0093] A random number is generated in the interval , and the minimum search step :

[0094] (21)

[0095] A random number between -0.5 and 0.5 is generated , and the universe chaotic search strategy is:

[0096] (22)

[0097] Update the search step:

[0098] (23)

[0099] Calculate the universe expansion degree, and update the individual universe according to the greedy strategy:

[0100] (24)

[0101] Step 3.6, merge the original individual universe, the individual universe of optimization and the chaotic disturbance individual universe, sort them from large to small according to the expansion degree, and eliminate a part of the universe by using the roulette method, so that the total number of the universe does not exceed the upper limit of the number of the universe;

[0102] Step 3.7, update the wormhole existence probability:

[0103] (25)

[0104] Update the travel distance rate:

[0105] (26)

[0106] Step 3.7, update the global extreme value and the individual extreme value, and set the global optimal universe as So that other universes follow the optimal universe for optimization in the next iteration;

[0107] Step 3.8, judge whether the maximum iteration number is reached, if yes, exit the optimization; otherwise, return to step 3.3;

[0108] Step 3.9, output the scheduling scheme meeting the constraint condition and the objective function.

[0109] Advantages of the present application:

[0110] The scheduling optimization model established by step one clearly defines the optimization target, determines the vehicle scheduling optimization variable, fully considers the influence of heterogeneous vehicles on the total cost, and has a larger application range than traditional logistics optimization algorithms. Step two gives various possibilities of the shortest path distance after the order is determined, which establishes a foundation for later path optimization. Step three uses the improved multi-universe algorithm of differential evolution algorithm to obtain the optimal scheduling scheme, and the algorithm has fast convergence speed, high convergence precision and good robustness, which effectively improves the efficiency and economy of large supermarket distribution, and realizes the rationalization of city scheduling. BRIEF DESCRIPTION OF DRAWINGS

[0111] Fig. 1 Flowchart of the present application

[0112] Fig. 2 Flowchart of the improved multi-universe algorithm of differential evolution algorithm in step three. DETAILED DESCRIPTION

[0113] In order to make the purpose, technical scheme and beneficial effects of the present application clearer, the preferred embodiments of the present application will be described in detail below with reference to the drawings, so as to facilitate the understanding of the skilled in the art.

[0114] As Figs. 1-2As shown, a heterogeneous vehicle intelligent scheduling method for large city supermarket distribution, characterized in that, comprising the following steps:

[0115] Step one, establish a heterogeneous vehicle scheduling optimization model based on city supermarket distribution;

[0116] The heterogeneous vehicle scheduling optimization mathematical model based on city supermarket distribution in step one is as follows:

[0117] Objective function:

[0118] (1)

[0119] (1) In the formula is the fixed total cost of all vehicles, is the total cost of all vehicle transportation, is the total cost of all drivers, is the utilization rate of all vehicle loading, is a 0-1 variable, indicating whether vehicle A is for delivery, 1 for delivery, and 0 for no delivery.

[0120] The optimization goal of city supermarket distribution is different according to different applications, and the present application takes the lowest distribution cost and the maximum vehicle loading rate as the goal.

[0121] Constraint conditions: vehicle maximum loading capacity constraint, vehicle maximum distribution point constraint, vehicle driving time window limit.

[0122] Further, the objective function calculation formula is as follows:

[0123] Vehicle fixed cost of vehicle:

[0124] (2)

[0125] (2) In the formula, indicates the vehicle model;

[0126] indicates the vehicle code;

[0127] indicates the starting price of the vehicle with model ;

[0128] indicates the point fee of the vehicle with model ;

[0129] is a 0-1 variable, indicating whether vehicle A reaches the delivery address , 1 for reaching, and 0 for not reaching; each vehicle fixed cost can be calculated by formula (2).

[0130] Fixed total cost of all vehicles

[0131] (3)

[0132] The fixed total cost of all vehicles is calculated by accumulating the fixed cost of each vehicle using equation (3).

[0133] Vehicle Vehicle freight cost

[0134] (4)

[0135] In equation (4), the vehicle is divided into bubble goods and heavy goods according to the different loading properties of the vehicle, so the bubble-heavy ratio of a vehicle is set as the ratio of the weight and volume of the goods, which is represented by

[0136] (5)

[0137] In equation (5) represents the weight of the goods sent by the vehicle to the delivery address;

[0138] represents the volume of the goods sent by the vehicle to the delivery address;

[0139] is the price required by the vehicle to transport a unit weight of goods to travel a unit distance when the goods loaded by the vehicle are heavy goods;

[0140] is the price required by the vehicle to transport a unit weight of goods to travel a unit distance when the goods loaded by the vehicle are bubble goods;

[0141] represents the distance between each receiving point;

[0142] represents the distance between the distribution center and the receiving point;

[0143] The transportation cost of each vehicle transporting bubble goods and heavy goods can be calculated by equation (5).

[0144] Total transportation cost of all vehicles

[0145] ​​​​​​​​​​​​(6)

[0146] The total transportation cost of all vehicles is calculated by summing up the transportation costs of each vehicle using equation (6).

[0147] Total cost of wages for all drivers ( ):

[0148] (7)

[0149] (5) In the formula, G represents the wages paid to workers per unit of time;

[0150] Indicates model number vehicles Driving on the middle section of the entire delivery route Time consumed;

[0151] Vehicle load utilization rate ( ):

[0152] (8)

[0153] In formula (8), Indicates vehicle Delivered to the delivery address The weight of the goods;

[0154] Indicates vehicle Delivered to the delivery address The volume of the goods;

[0155] Load utilization rate of all delivery vehicles ( ):

[0156] (9)

[0157] The formula for calculating the constraints is as follows:

[0158] Maximum vehicle load capacity constraint:

[0159] (10)

[0160] (11)

[0161] In equations (10) and (11), Indicates vehicle Delivered to the delivery address The weight of the goods;

[0162] Indicates vehicle Delivered to the delivery address The volume of the goods;

[0163] This indicates the maximum delivery weight of the vehicle;

[0164] This indicates the maximum delivery volume of the vehicle;

[0165] By constraining the maximum load capacity of each vehicle using equations (10) and (11), we can avoid the optimization results from causing vehicle overloading.

[0166] Maximum delivery point constraint for vehicles:

[0167] (12)

[0168] (12) In the formula Total number of delivery vehicles; Total number of delivery addresses;

[0169] Vehicle driving time window restrictions:

[0170] (13)

[0171] In formula (13) This is the average delivery time. To the delivery address Total delivery distance For average delivery speed, Indicates vehicle Delivered to the delivery address The weight of the goods; For loading and unloading speed; To the delivery address The maximum set time.

[0172] By using formula (13) to constrain the vehicle's driving time window, delivery time can be avoided from being overdue.

[0173] Step 2: Based on the order requirements, set the range of the number of delivery points the vehicle will pass through, use the Hamiltonian cycle algorithm to calculate the shortest distance for each path, and match the shortest path distance with each vehicle to generate individual data and generate the initial population.

[0174] Step 2.1: Set the total number of orders and order numbers, and set the maximum number of vehicles that can be assigned to the distribution center. ;

[0175] Step 2.2: Calculate the total number of orders. Generate a number from 1 to 2 for each order. random integers between , representing the The vehicle will transport this order;

[0176] Step 2.3, initialize the city matrix, get the distance between each pickup point, set the number of vehicle passing through the pickup point interval, use Hamilton loop algorithm to calculate the shortest distance of each path;

[0177] By setting the number of vehicle passing through the pickup point interval, the possible paths of vehicle distribution passing through the pickup point can be obtained. The path of vehicle between each pickup point is equivalent to the traveling salesman problem, which can be expressed as Where V represents the position of each pickup point, and E represents the edge set between each pickup point. By Hamilton loop algorithm, the shortest path of vehicle passing through all pickup points, each pickup point passing through only once, and finally returning to the starting position of distribution center can be calculated.

[0178] The Hamilton loop specific calculation formula is as follows:

[0179] From any starting point, passing through all other nodes, and finally returning to the starting point, let the vertex , the edge , be the distance between points , , and define the decision variable:

[0180] (14)

[0181] In formula (14), 1 represents the point not passed, and 0 represents the point passed; it is determined that each edge can only be accessed once;

[0182] Objective function:

[0183] (15)

[0184] In formula (15), is the position of each pickup point, is the sum of the minimum distance of Hamilton loop;

[0185] Constraint condition:

[0186] (16)

[0187] Formula (16) represents the symmetric loop;

[0188] (17)

[0189] (17) represents that all edges can only be accessed once;

[0190] (18)

[0191] (18) The total length of the Hamilton circuit composed of N pickup points is represented by the formula, The location of the distribution center is represented by (18) formula, The location of the first pickup point is represented by (18) formula, The starting pickup point is represented by (18) formula, The ending pickup point is represented by (18) formula, and the delivery vehicle returns to the distribution center after completing the delivery.

[0192] Step 2.4, assign each shortest path distance to each vehicle, form an order distribution set, and add orders that meet the constraint conditions to the initial population until the number of individuals in the initial population reaches the upper limit of the population size.

[0193] Initial population (27)

[0194] In the formula (27) , the first vehicle delivers the shortest distance order.

[0195] Step three, use the improved differential evolution algorithm to search for the optimal scheduling scheme.

[0196] Step 3.1, initialize algorithm parameters: including maximum iteration number T; maximum chaotic search number S; universe number NP; universe dimension D; wormhole existence probability WEP; travel distance rate TDR; TDR related constant C and p; search space upper limit ; search space lower limit ; initial population ;

[0197] Substitute the initial population obtained in step two into the improved differential evolution algorithm of the multi-universe algorithm to search for the optimal scheduling scheme.

[0198] Step 3.2, update the global extreme value of the multi-universe and the individual extreme value, set the optimal individual of the initial universe as the global extreme value; if the individual universe does not meet the constraint condition, set the expansion rate of the universe to negative infinity or artificially change the material of the universe;

[0199] Step 3.3, for individual universe, generate a random number between 0 and 1, if is less than the standard expansion rate of the individual, select the universe that produces black holes according to the roulette method, and exchange the material of the universe with the universe selected by the roulette method;

[0200] (19)

[0201] ​(19) where, is the kth variable of the ith universe; is the standard expansion rate of the individual universe; is a random number between 0 and 1;

[0202] Step 3.4, the universe receives the material sent from the optimal universe;

[0203] If is greater than the standard expansion rate of the universe, a random number between 0 and 1 is generated If is less than the wormhole existence probability WEP, the optimal universe sends material to the universe through the wormhole;

[0204] A random number between 0 and 1 is generated , which will determine the search direction of the universe;

[0205] The mutation operator in the differential evolution algorithm is used to further improve the operation of following the optimal universe search;

[0206] (20)

[0207] Save the universe after following the optimal universe search;

[0208] Step 3.5, chaotic search

[0209] If the of the universe is greater than the wormhole existence probability WEP, chaotic search is performed on the universe;

[0210] A random number is generated in the interval , and the minimum search step :

[0211] (21)

[0212] A random number between -0.5 and 0.5 is generated , and the universe chaotic search strategy is:

[0213] (22)

[0214] Update the search step:

[0215] (23)

[0216] Calculate the universe expansion degree, and update the individual universe according to the greedy strategy:

[0217] (24)

[0218] Step 3.6, merge the original individual universe, the individual universe of optimization and the individual universe of chaos disturbance, sort them according to the expansion degree from large to small, eliminate a part of the universe by using the roulette method, so that the total number of universes does not exceed the upper limit of the number of universes;

[0219] Step 3.7, update the wormhole existence probability:

[0220] (25)

[0221] Update the travel distance rate:

[0222] (26)

[0223] Step 3.7, update the global extreme value and individual extreme value, set the global optimal universe as In order to follow the optimal universe for optimization in the next iteration;

[0224] Step 3.8, judge whether the maximum iteration number is reached, if yes, exit the optimization; otherwise, return to step 3.3;

[0225] Step 3.9, output the scheduling scheme meeting the constraint condition and the objective function.

[0226] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any change or replacement without creative labor should be covered in the protection scope of the present application.

Claims

1. A method for intelligent scheduling of heterogeneous vehicles for delivery to large urban supermarkets, characterized in that, Includes the following steps: Step 1: Establish a heterogeneous vehicle dispatch optimization model based on urban supermarket delivery; Step 2: Based on the order requirements, set the range of the number of delivery points that the vehicle will pass through, use the Hamiltonian cycle algorithm to calculate the shortest distance for each path, match the shortest path distance with each vehicle to generate an order allocation set, and use the order allocation set as the initial population. Step 3: Optimize the initial population using a hybrid improved multiverse algorithm based on differential evolution to find the optimal scheduling scheme. The specific implementation steps of Step 3 are as follows: Step 3.1: Initialize algorithm parameters, including maximum number of iterations T; maximum number of chaotic searches S; number of universes NP; universe dimension D; wormhole existence probability WEP; travel distance rate TDR; TDR-related constants C and p; and upper limit of the search space. Search space lower bound Initial population ; Step 3.2: Update the global and individual extreme values ​​of the multiverse, and set the optimal individual in the initial universe as the global extreme value; if an individual universe does not meet the constraints, set the expansion rate of that universe to negative infinity or artificially change the matter in that universe. Step 3.3: For an individual universe, generate a random number between 0 and 1. ,like If the expansion rate is less than the standard expansion rate of the individual, then the universe that produces the black hole is selected according to the roulette wheel method, and the universe exchanges matter with the universe selected by the roulette wheel. （19） (19) In the formula, Let k be the variable in the i-th universe; The standard expansion rate of an individual universe; A random number between 0 and 1; Step 3.4: The universe receives matter sent from the optimal universe; like If the expansion rate is greater than the standard expansion rate of the universe, then a random number between 0 and 1 is generated. ,like If the probability of a wormhole's existence is less than the probability of its existence (WEP), then the optimal universe sends matter to that universe through the wormhole. Generate random numbers between 0 and 1 This random number will determine the direction of the search in that universe; Further improvements were made to the operation of following the optimal universe search using the mutation operator in the differential evolution algorithm; （20）； Save the universe after following the optimal universe search; Step 3.5, Chaotic Search If that universe If the probability of a wormhole's existence is greater than the probability of its existence (WEP), then a chaotic search is performed on that universe. The interval is generated at random numbers Let the minimum search step size be : （21） Generate random numbers between -0.5 and 0.

5. The cosmic chaos search strategy is as follows: （22） Update search step size: （23） Calculate the expansion of the universe and update individual universes using a greedy strategy: （24） Step 3.6: Merge the original individual universes, the optimal individual universes, and the chaotic perturbation individual universes, sort them from largest to smallest expansion degree, and use a roulette wheel to eliminate some universes so that the total number of universes does not exceed the upper limit of the number of universes; Step 3.7: Update the probability of wormhole existence: （25） Update travel distance rate: （26） Step 3.7: Update the global and individual extrema, and set the globally optimal universe as... This is so that other universes can follow the optimal universe in the next iteration to find the best one; Step 3.8: Determine if the maximum number of iterations has been reached. If yes, exit the optimization; otherwise, return to step 3.

3. Step 3.9: Output the scheduling scheme that satisfies the constraints and objective function.

2. The method for intelligent scheduling of heterogeneous vehicles for delivery to large urban supermarkets according to claim 1, characterized in that, The mathematical model for heterogeneous vehicle scheduling optimization based on urban supermarket delivery described in step one is as follows: Objective function: （1） (1) In the formula For the total fixed cost of all vehicles, Total transportation cost for all vehicles The total cost of wages for all drivers For the load utilization rate of all vehicles, It is a 0-1 variable, indicating whether vehicle A is transporting goods; 1 indicates transporting goods, and 0 indicates not transporting goods. Constraints: Maximum vehicle load capacity, maximum number of delivery points per vehicle, and vehicle travel time window limit.

3. The method for intelligent scheduling of heterogeneous vehicles for delivery to large urban supermarkets according to claim 2, characterized in that, The formula for calculating the objective function is as follows: Vehicle fixed costs: （2） (2) In the formula, Indicates the vehicle model; Indicates the vehicle code; Indicates model number The starting price for vehicles; Indicates model number Vehicle location fee; This is a 0-1 variable, indicating whether vehicle A has arrived at the delivery address. Arrival is represented by 1, and non-arrival by 0; Total fixed cost of all vehicles ( ): （3） vehicle Vehicle freight costs ( ): （4） (4) In the formula, since the vehicles are divided into bulky goods and heavy goods according to their loading attributes, the bulky-to-weight ratio of a certain vehicle is set as the ratio of the weight to the volume of the goods. To indicate; （5） (5) In the formula Indicates vehicle Send to delivery address The weight of the goods; Indicates vehicle Send to delivery address The volume of the goods; For vehicles When the loaded goods are heavy goods ( The price required for the vehicle to transport a unit weight of goods per unit distance. For vehicles When the loaded goods are bulky goods ( The price required for the vehicle to transport a unit weight of goods per unit distance. Indicates the distance between each receiving point; Indicates the distance between the distribution center and the receiving point; Total transportation cost of all vehicles ( ): （6） Total cost of wages for all drivers ( ): （7） (5) In the formula, G represents the wages paid to workers per unit of time; Indicates model number vehicles Driving on the middle section of the entire delivery route Time consumed; Vehicle load utilization rate ( ): （8） In formula (8), Indicates vehicle Send to delivery address The weight of the goods; This indicates the maximum delivery weight of the vehicle; This indicates the maximum delivery volume of the vehicle; Indicates vehicle Send to delivery address The volume of the goods; Load utilization rate of all delivery vehicles ( ): （9）。 4. The method for intelligent scheduling of heterogeneous vehicles for delivery to large urban supermarkets according to claim 2, characterized in that, The formula for calculating the constraints is as follows: Maximum vehicle load capacity constraint: （10） （11） In equations (10) and (11), Indicates vehicle Send to delivery address The weight of the goods; Indicates vehicle Send to delivery address The volume of the goods; This indicates the maximum delivery weight of the vehicle; This indicates the maximum delivery volume of the vehicle; Maximum delivery point constraint for vehicles: （12） (12) In the formula Total number of delivery vehicles; Total number of delivery addresses; Vehicle driving time window restrictions: （13） In formula (13) Average delivery time To the delivery address Total delivery distance For average delivery speed, Indicates vehicle Send to delivery address The weight of the goods; For loading and unloading speed; To the delivery address The maximum set time.

5. The method for intelligent scheduling of heterogeneous vehicles for delivery to large urban supermarkets according to claim 1, characterized in that, The specific implementation steps of step two are as follows: Step 2.1: Set the total number of orders and order numbers, and set the maximum number of vehicles that can be assigned to the distribution center. ; Step 2.2: Calculate the total number of orders. Generate a number from 1 to 2 for each order. random integers between , representing the The vehicle will transport this order; Step 2.3: Initialize the city matrix, obtain the distance between each delivery point, set the range of the number of delivery points passed by the vehicle, and use the Hamiltonian cycle algorithm to calculate the shortest distance of each path. Step 2.4: Assign the shortest path distance to each vehicle to form an order allocation set. Add orders that meet the constraints to the initial population until the number of individuals in the initial population reaches the population limit.

6. The method for intelligent scheduling of heterogeneous vehicles for delivery to large urban supermarkets according to claim 5, characterized in that, The specific calculation formula for the Hamiltonian circuit is as follows: Starting from any given point, traversing all other nodes, and finally returning to the starting point, let's define a vertex. ,side , For point , The distance between them, define the decision variable: （14） In equation (14), 1 represents a point that has not been visited, and 0 represents a point that has been visited; this determines that each edge can only be visited once. Objective function: （15） In equation (15), For the locations of each receiving point, It is the sum of the minimum distances of the Hamiltonian circuit; Constraints: （16） Equation (16) represents a symmetrical circuit; （17） (17) indicates that all edges can be visited exactly once; （18） Equation (18) represents the total length of the Hamiltonian loop formed by N receiving points. For the location of the distribution center, Indicates the first Location of each receiving point Indicates the originating receiving point. Indicates the final receiving point.

Images