A Multi-Temperature Co-Allocation Path Planning Method Based on Enhanced Multi-Strategy Spider Wasp Algorithm

Abstract

The present invention discloses a multi-temperature co-distribution path planning method based on an enhanced multi-strategy spider wasp algorithm, which is applied to the multi-temperature co-distribution technology field in the field of logistics and supply chain management. The method includes: obtaining the vehicle resources, customer demands and location information of a distribution center, and establishing a refrigeration cost calculation function based on the change of environmental temperature; based on the refrigeration cost calculation function, constructing a multi-temperature co-distribution path planning model with the goal of minimizing the total distribution cost based on flexible compartments and environmental temperature, and constructing the constraint conditions of the multi-temperature co-distribution path planning model based on the vehicle resources and customer information; using the enhanced multi-strategy spider wasp algorithm combined with chaotic mapping, adaptive elite strategy and reciprocating population reduction strategy to solve the objective of the model, and obtaining the optimal departure time and distribution path. The present invention effectively improves the energy efficiency utilization rate in the cold chain transportation process, and while minimizing the number of vehicles used, improves the full load rate of the vehicles.

Description

Technical Field

[0001] The present invention relates to the field of multi-temperature co-distribution technology in the field of logistics and supply chain management, and particularly relates to a multi-temperature co-distribution path planning method based on an enhanced multi-strategy spider bee algorithm. Background Art

[0002] With the acceleration of the urbanization process and the pace of life, consumers' demand for convenient diets has promoted the rapid development of the prefabricated food industry. As a key support, cold chain logistics needs to optimize the path and reduce costs in multi-temperature zone distribution while ensuring food quality. However, traditional path optimization methods have limitations in dealing with time windows, order splitting, and temperature control complexity, and it is difficult to meet the high-efficiency requirements of multi-temperature co-distribution.

[0003] Flexible compartment technology provides greater flexibility for multi-temperature co-distribution, improving the loading rate and distribution efficiency by dynamically adjusting the compartment size. However, existing research mostly focuses on specific cargo types and does not fully consider the diverse temperature requirements and environmental changes in prefabricated food distribution. Optimizing the vehicle departure time is also an important strategy to improve distribution efficiency, but most studies assume a fixed departure time and ignore the impact of long-term high-temperature exposure on refrigeration energy consumption considering environmental temperature factors, limiting the optimization effect.

[0004] Therefore, how to provide a multi-temperature co-distribution path optimization model that can comprehensively consider environmental temperature, flexible compartments, time windows, and load factors, combine flexible compartments and departure time optimization, and design an improved multi-strategy spider bee optimization algorithm to improve the solution quality and convergence speed through chaos mapping, adaptive mutation, and population reduction strategies, so as to improve the energy efficiency utilization rate during cold chain transportation, and while minimizing the number of vehicles used, improve the full load rate of vehicles is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention proposes a multi-temperature co-distribution path planning method based on an enhanced multi-strategy spider wasp algorithm. First, obtain the vehicle resources, customer demands, and location information of the distribution center; secondly, calculate the environmental temperature at time t, and establish a refrigeration cost calculation function based on the change of the environmental temperature; finally, under the vehicle resource constraints of the distribution center, construct a multi-temperature co-distribution optimization model that minimizes the total distribution cost, and use an enhanced multi-strategy spider wasp algorithm that combines chaotic mapping, adaptive elite strategy, and reciprocating population reduction strategy to solve the objective of the multi-temperature co-distribution path planning model, and obtain the optimal departure time and distribution path. The present invention combines the flexible compartment and the dynamic change of the environmental temperature, and improves the energy efficiency utilization rate in the cold chain transportation process by adaptively optimizing the vehicle departure time. While minimizing the number of vehicles used, it also improves the full load rate of the vehicles, providing a more efficient energy-saving and consumption-reducing and quality assurance solution for enterprises.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A multi-temperature co-distribution path planning method based on an enhanced multi-strategy spider wasp algorithm, comprising:

[0008] Step 1: Obtain the vehicle resources, customer demands, and location information of the distribution center, and establish a refrigeration cost calculation function based on the change of the environmental temperature;

[0009] Step 2: Based on the refrigeration cost calculation function, construct a multi-temperature co-distribution path planning model based on flexible compartments and environmental temperature with the goal of minimizing the total distribution cost, and construct the constraint conditions of the multi-temperature co-distribution path planning model based on vehicle resources and customer information;

[0010] Step 3: Use an enhanced multi-strategy spider wasp algorithm that combines chaotic mapping, adaptive elite strategy, and reciprocating population reduction strategy to solve the objective of the multi-temperature co-distribution path planning model, and obtain the optimal departure time and distribution path.

[0011] Optionally, in Step 1, establishing a refrigeration cost calculation function based on the change of the environmental temperature is specifically:

[0012] Divide the environmental temperature by hour, take each hour as a time zone, and calculate the environmental temperature of each time zone as follows:

[0013]

[0014] Wherein, represents the environmental temperature of the mth time zone; m represents the number of the time zone's belonging moment; represents the highest temperature within a day; represents the lowest temperature within a day; represents the simulation of each hour temperature values; is a fixed parameter;

[0015] Assume that the temperature within each hour is constant. Consider the vehicle as passing through one or more time zones. Let vehicle k depart from point i, and the departure time is denoted as , the next node visited is denoted as j, and the travel time from point i to point j is The service time at point j is denoted as , the heat conduction area of the refrigerated and frozen compartments is as follows:

[0016]

[0017] Among them, represents the distance from node i to node j; v represents the speed of vehicle k; N represents the set of nodes composed of customer points and the distribution center, i, j ∈ N, when i = 0, it represents the distribution center; p represents the p-type compartment of the distribution vehicle, p = 1 represents the refrigerated type, p = 2 represents the frozen type; represents the demand for refrigerated goods at node j; represents the demand for p-type goods at node j; Z represents the total heat conduction area of the compartment; represents the demand for frozen goods at node j;

[0018] Time The time zone m to which it belongs is denoted as , among which, , are respectively the start time point and end time point corresponding to the time zone m to which the departure time belongs. also represents the start time point of time zone m + 1. Since the end time point of time zone m is the same as the start time point of time zone m + 1, the end time point corresponding to time zone m is represented by the start time point of time zone m + 1; then the driving and service time in time zone is denoted as , and judge whether holds;

[0019] If it holds, then from time to leaving point j, the vehicle is driving at a temperature of . Ignoring the influence of opening and closing the doors during service, the refrigeration cost during this period is expressed as:

[0020]

[0021] Among them, represents the temperature of time zone m; represents the heat conduction coefficient of the compartment; represents the temperature of the refrigerated compartment; represents the temperature of the frozen compartment; represents the unit fuel supply level; K represents the set of distribution vehicles; represents the refrigeration cost; represents the heat generated by the refrigeration unit during the process of distribution vehicle k traveling from node i to node j; E represents the heat generated by unit gasoline;

[0022] If not, then in the time period the vehicle drives at temperature during this period, is expressed as:

[0023]

[0024] Update , , re-judge whether it holds to obtain the refrigeration cost.

[0025] Optionally, in step 2, based on the refrigeration cost calculation function, construct a multi-temperature co-distribution path planning model with the goal of minimizing the total distribution cost, as follows:

[0026]

[0027]

[0028] Among them, represents the fixed cost; K represents the set of distribution vehicles; represents the fixed resource consumption for starting distribution vehicle k; represents whether the vehicle departs from the distribution center; represents the transportation cost; N represents the set of nodes composed of customer points and the distribution center, i, j ∈ N, when i = 0, it represents the distribution center; represents whether distribution vehicle k travels from node i to the next node j; represents the capacity of the distribution vehicle; represents the initial quantity of goods loaded when distribution vehicle k departs from the distribution center; represents the fuel consumption rate when the distribution vehicle is fully loaded; represents the fuel consumption rate when the distribution vehicle is empty; represents the distance from node i to node j; represents the unit fuel supply level; represents the loss cost; p represents the p - type compartment of the distribution vehicle, p = 1 represents the refrigerated type, p = 2 represents the frozen type; represents the loss rate of p - type goods; L represents the unit price of goods per unit quantity; represents the time required to travel from node i to node j; Denote the service time of node j; Denote the demand quantity of type-p goods at node j; Denote the carriage of vehicle k Whether it can meet the demand of type-p goods of customer i; Denote the time window penalty cost; Denote the time window of customer point i, where Denote the earliest arrival time required for the distribution vehicle to reach node i, Denote the latest arrival time required for the distribution vehicle to reach node i; Denote the arrival time of distribution vehicle k at node i; 、 Denote the penalty coefficients for the vehicle arriving early and late respectively; Denote the refrigeration cost.

[0029] Optionally, in step 2, the constraint conditions for constructing the multi-temperature co-distribution path planning model based on vehicle resources and customer information are as follows:

[0030]

[0031]

[0032] Among them, N represents the node set composed of customer points and the distribution center, i, j ∈ N, when i = 0, it represents the distribution center; Denote whether to depart from the distribution center to node j; Denote whether to return from a certain node i to the distribution center; K represents the set of distribution vehicles; Denote whether distribution vehicle k travels from node i to the next node j; Denote whether the type-p carriage of vehicle k can meet the demand of type-p goods of customer j; p represents the type-p carriage of the distribution vehicle, p = 1 represents the refrigerated type, and p = 2 represents the frozen type; Denote whether vehicle k travels from node j to node i; Denote the demand quantity of type-p goods at node i; Denote the capacity of the distribution vehicle; Denote the departure time of distribution vehicle k from point j; Denote the departure time of distribution vehicle k from point i; Denote the time required to travel from node i to node j; Denote the service time of node j; S represents the set of path vertices; Denote the departure time of distribution vehicle k from the distribution center; Denote the distribution time window of the distribution center, Denote the earliest time required by the distribution center for distribution vehicle k; Represents the latest time; m represents the number of customers; Indicates whether the delivery vehicle k travels from node i to the next node j; Indicates whether the p - type carriage of the delivery vehicle k meets the demand for p - type goods of customer i; Represents the time of departure from the last customer point of the path; Represents the latest time to return to the distribution center.

[0033] Optionally, in step 3, an enhanced multi - strategy spider - bee algorithm combining chaotic mapping, adaptive elite strategy, and reciprocating population reduction strategy is used to solve the objective of the multi - temperature co - distribution path planning model. Specifically:

[0034] Initialize parameters, including: population size G, minimum population size , upper bound ub, lower bound nb, number of iterations , trade - off rate TR, crossover rate Cr, number of mutations of the current optimal individual p_m, number of mutations of the global optimal individual g_m; Define and initialize the current iteration number as T = 0;

[0035] Use the tent chaotic mapping initialization strategy to generate the initial spider - bee population, and perform encoding and decoding on the generated population. Calculate the fitness value set of the initial population using the multi - temperature co - distribution path planning model , from the fitness value set of the initial population select the spider - bee individual corresponding to the minimum fitness value , denoted as the initial population optimal spider - bee individual and the global optimal spider - bee individual , and the corresponding fitness value is the global optimal fitness value ;

[0036] Enter the main loop, generate a random number as the search agent index, , if , perform prey behavior to update the spider - bee population, otherwise perform mating behavior to update the spider - bee population; When performing prey behavior to update the spider - bee population, let , when , generate a random number p, , when , use search behavior to update the individuals in the spider - bee population, when , adopt following and escaping behaviors to update the individuals in the spider - bee population; When , use nesting behavior to update the individuals in the spider - bee population, generate random numbers , , , when , use the current optimal solution Update the individuals in the spider wasp population. When occurs, randomly select a spider wasp within the population and use an additional step size to update the individuals in the spider wasp population; when performing mating behavior to update the spider wasp population, randomly select spider wasp individuals 、 、 Generate parameters 、 , and use the parameters 、 to generate new spider wasp individuals. For each newly generated spider wasp individual in the population, generate a random number rand. When rand < Cr, update the corresponding individual in the population to the new spider wasp individual; otherwise, do not update.

[0037] Encode and decode the currently obtained new spider wasp individuals, and calculate the fitness value set of all individuals in the spider wasp population according to the multi-temperature co-mating path planning model , and respectively compare the fitness values in the current set with the fitness values of the corresponding individuals in the spider wasp population before the update. If the current one is better, update the corresponding individual in the population to the current spider wasp individual; otherwise, retain the original individual. Compare them in turn to update the spider wasp population ; Encode and decode the currently obtained new spider wasp individuals, calculate the fitness value set of all individuals in the updated spider wasp population, and compare the optimal fitness value with the global optimal fitness value . If the current fitness value is better, update the optimal spider wasp individual and the optimal fitness value to the current spider wasp individual and its corresponding fitness value;

[0038] Let the current iteration number be , and use the adaptive elite mutation strategy to perform elite mutation on each individual in the current spider wasp population p_m times. Check whether the new spider wasp positions exceed the upper and lower bounds ub and nb of the search space. If so, correct them to the boundaries; use the adaptive elite mutation strategy to perform elite mutation on the optimal spider wasp individual in the new population g_m times. If the fitness value of each mutated individual is better, update the corresponding spider wasp individual and the fitness value to the current spider wasp individual and its fitness value to obtain the new spider wasp population after mutation ;

[0039] Compare the sizes of and . If is better, update the optimal spider wasp individual and the optimal fitness value is the current spider wasp individual and the corresponding fitness value, and the population size G is updated using the reciprocating population reduction strategy until the number of reciprocating cycles is reached, and it is judged whether the current iteration number T is greater than , if so, output the corresponding optimal fitness value and the optimal spider wasp individual , otherwise return to the main loop to continue repeating the steps and update the solution.

[0040] Optionally, the tent chaos mapping initialization strategy is used to generate the initial spider wasp population, and the generated population is encoded and decoded. Specifically:[[]]

[0041] Randomly generate an initial matrix with 2M rows and G columns, and use the tent chaos mapping to update the initial matrix to generate the initial spider wasp population as follows:

[0042]

[0043] where represents the m-th value of the g-th spider wasp individual; the chaos parameter ; represents the m-th value of the g-th parameter in the initial matrix; represents the population size; represents the chaos variable; rand represents a random number; ub and lb represent the upper and lower bounds respectively; the generated initial spider wasp population is as follows:

[0044]

[0045] where each row represents a spider wasp individual , ; each spider wasp individual represents a set of solutions, and the spider wasp population is simplified to ;

[0046] For any spider wasp individual , represents the departure time, represents the customer node; the location-order encoding method is introduced to decode the delivery path of the last M nodes, and is sorted in ascending order, the numbers are recorded and then converted to natural number encoding, and this part of the encoding represents the customer point number; the path is allocated according to the vehicle capacity constraint. If the sum of the customer point demands is satisfied, the customer point is added to the path, otherwise a new path is generated until all nodes are allocated; for the vehicle departure time , it is directly allocated in order from front to back according to the number of generated paths.

[0047] Optionally, when updating the spider wasp population with prey behavior, use search behavior to update the individuals in the spider wasp population, specifically as follows:

[0048] Generate a random number 、 , ; When , generate parameter , , where rn1 is a random number; Randomly select two spider wasp individuals from the population 、 , and update the parameters in each spider wasp individual using the following formula:

[0049]

[0050] When , generate parameter , , where , , , rand represents a random number; Update the parameters in each spider wasp individual using the following formula:

[0051]

[0052] Use following and escaping behaviors to update the individuals in the spider wasp population, specifically as follows:

[0053] When , randomly select a spider wasp individual from the population , generate parameter C,

[0054] , and update the parameters in each spider wasp individual using the following formula:

[0055]

[0056] When , generate parameter matrix , , where , and update the parameters in each spider wasp individual using the following formula:

[0057]

[0058] Use nesting behavior to update the individuals in the spider wasp population, specifically as follows:

[0059] When , use the global optimal spider wasp individual to update the spider wasp individual as follows:

[0060]

[0061] When occurs, randomly select spider wasp individuals , , , and generate new spider wasp individuals using the following formula:

[0062]

[0063] where represents the parameter generated by Levy flight.

[0064] Optionally, when performing mating behavior to update the spider wasp population, randomly select spider wasp individuals , , to generate parameters , , specifically:

[0065] Calculate the matrix , randomly select a spider wasp individual , ;

[0066] Calculate , randomly select spider wasp individuals , , ;

[0067] Use the parameters , to generate new spider wasp individuals. For each newly generated spider wasp individual in the population, generate a random number rand. When rand < Cr, update the corresponding individual in the population to the new spider wasp individual using the following formula, otherwise do not update;

[0068]

[0069] Boundary check, check whether the new spider wasp position exceeds the upper and lower bounds ub, lb of the search space. If so, correct it to the boundary to obtain the updated spider wasp population at the T-th generation , .

[0070] Optionally, perform elite mutation on each individual in the current spider wasp population using the adaptive elite mutation strategy as follows:

[0071]

[0072] where , represent randomly selecting a spider wasp individual from the population.

[0073] Optionally, the population size G is updated using a reciprocating population reduction strategy as follows:

[0074]

[0075] Among them, % is the remainder operator, Indicates the number of reciprocating cycles.

[0076] It can be seen from the above technical solutions that, compared with the prior art, the present invention proposes a multi-temperature co-distribution path planning method based on the enhanced multi-strategy spider bee algorithm. The present invention enriches the research content of flexible compartments in multi-temperature co-distribution path optimization, and proposes a dynamic adjustment method based on flexible compartments, making distribution path planning more flexible and efficient. Flexible compartments can adjust space allocation according to the temperature control requirements of different goods, improve loading efficiency and resource utilization in a multi-temperature zone co-distribution environment, and thus provide a more intelligent and adaptable solution for cold chain distribution. By optimizing the vehicle departure time strategy, the present invention effectively reduces energy consumption during transportation, especially in the face of long-distance, multi-temperature co-distribution cold chain transportation scenarios. The optimized departure time can effectively avoid high temperature periods and reduce the vehicle's stay time in a high temperature environment, thereby reducing the refrigeration burden. The optimized departure time not only improves transportation efficiency, but also greatly reduces energy consumption, enhancing the environmental protection and economy of cold chain transportation. The present invention provides a comprehensive route optimization and energy efficiency management solution for the distribution of cold chain products such as pre-prepared meals, combining flexible compartments with departure time optimization, which not only improves transportation efficiency, but also reduces refrigeration energy waste by precisely controlling the temperature control environment. By implementing this solution, enterprises can improve transportation efficiency and resource utilization while ensuring the quality of cold chain distribution, and promote the industry to develop in a more efficient and environmentally friendly direction. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0078] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0079] Figure 2 It is a schematic diagram of the enhanced multi-strategy spider bee algorithm flow of the present invention.

[0080] Figure 3 It is a schematic diagram of the path optimization result of the present invention.

[0081] Figure 4Schematic diagram of the number of dispatched vehicles considering and not considering flexible compartments in the present invention.

[0082] Figure 5 Schematic diagram of the vehicle full load rate considering and not considering flexible compartments in the present invention.

[0083] Figure 6 Schematic diagram of the comparison of the total cost of optimizing and not optimizing the departure time in the present invention.

[0084] Figure 7 Schematic diagram of the iteration comparison between the algorithm of the present invention and other algorithms. Detailed implementation manners

[0085] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0086] Embodiment 1:

[0087] Embodiment 1 of the present invention discloses a multi-temperature co-distribution path planning method based on an enhanced multi-strategy spider bee algorithm, as Figure 1 shown, including:

[0088] It is stipulated that each delivery vehicle starts from the distribution center point 0 and finally returns to the distribution center point 0.

[0089] Step 1: Obtain the vehicle resources, customer demands and location information of the distribution center, and establish a refrigeration cost calculation function based on the change of environmental temperature.

[0090] The vehicle resources, customer demands and location information of the distribution center are specifically: the maximum load capacity Q of the delivery vehicle, the heat conduction area of the carriage, the location distribution of the distribution center point and customer points, the demand of each customer point and the time window requirement.

[0091] Establishing a refrigeration cost calculation function based on the change of environmental temperature is specifically:

[0092] The change of environmental temperature can be approximately regarded as a sine curve. The environmental temperature is divided by hour, and each hour is used as a time zone to calculate the environmental temperature of each time zone as follows:

[0093]

[0094] Among them, represents the environmental temperature of the mth time zone; m represents the number of the moment to which the time zone belongs; represents the highest temperature within a day; Represents the lowest temperature within a day; Represents the hourly simulation of temperature values; Is a fixed parameter, taking the value of 12;

[0095] Assume that the temperature within each hour is constant. Consider the vehicle as passing through one or more time zones. Let vehicle k depart from point i, and the departure time is represented as , the next node to be visited is represented as j, and the travel time from point i to point j is The service time at point j is represented as , and the heat conduction area of the refrigerated and frozen compartments is as follows:

[0096]

[0097] Among them, Represents the distance from node i to node j; v represents the speed of vehicle k; N represents the set of nodes composed of customer points and the distribution center, i, j ∈ N, when i = 0, it represents the distribution center; p represents the p - type compartment of the distribution vehicle, p = 1 represents the refrigerated type, p = 2 represents the frozen type; Represents the demand for refrigerated goods at node j; Represents the demand for p - type goods at node j; Z represents the total heat conduction area of the compartments; Represents the demand for frozen goods at node j;

[0098] Time The time zone m to which it belongs is represented as , among which, 、 Are respectively the start time point and the end time point corresponding to the time zone m to which the departure time belongs. Also represents the start time point of time zone m + 1. Since the end time point of time zone m is the same as the start time point of time zone m + 1, the end time point corresponding to time zone m is represented by the start time point of time zone m + 1; then the driving and service time that can still be carried out in time zone is represented as , and judge whether holds;

[0099] If it holds, then from the time to leaving point j, the vehicle is driving at a temperature of . Ignoring the influence of opening and closing the doors during service, the refrigeration cost during this period is represented as:

[0100]

[0101] Among them, Represents the temperature of time zone m; represents the heat conduction coefficient of the carriage; represents the temperature of the refrigerated carriage; represents the temperature of the frozen carriage; represents the unit fuel supply level; K represents the set of distribution vehicles; represents the refrigeration cost; represents the heat generated by the refrigerator during the process of the distribution vehicle k traveling from node i to node j; E represents the heat generated by unit gasoline;

[0102] If not, then in the time period the vehicle drives at the temperature and the of this period is expressed as:

[0103]

[0104] Update , , and re-judge whether it holds to obtain the refrigeration cost.

[0105] Step 2: Based on the refrigeration cost calculation function, construct a multi-temperature co-distribution path planning model with the goal of minimizing the total distribution cost, based on flexible compartments and ambient temperature, and construct the constraint conditions of the multi-temperature co-distribution path planning model based on vehicle resources and customer information.

[0106] Taking , as decision variables, where represents whether the distribution vehicle k travels from node i to the next node j; represents whether the p-type carriage of vehicle k meets the p-type commodity demand of customer j.

[0107] Based on the refrigeration cost calculation function, construct a multi-temperature co-distribution path planning model with the goal of minimizing the total distribution cost, based on flexible compartments and ambient temperature, as follows:

[0108]

[0109]

[0110] Among them, represents the fixed cost; K represents the set of distribution vehicles; represents the fixed resource consumption for starting the distribution vehicle k; represents whether the vehicle departs from the distribution center; represents the transportation cost; N represents the set of nodes composed of customer points and the distribution center, i, j ∈ N, when i = 0, it represents the distribution center; Indicate whether the delivery vehicle k travels from node i to the next node j; Indicate the capacity of the delivery vehicle; Indicate the quantity of the initially loaded goods when the delivery vehicle k departs from the distribution center; Indicate the fuel consumption rate when the delivery vehicle is fully loaded; Indicate the fuel consumption rate when the delivery vehicle is empty; Indicate the distance from node i to node j; Indicate the unit fuel supply level; Indicate the loss cost; p represents the p - type carriage of the delivery vehicle, p = 1 represents the refrigerated type, and p = 2 represents the frozen type; Indicate the loss rate of the p - type goods; L represents the unit price of the goods per quantity; Indicate the time required to travel from node i to node j; Indicate the service time of node j; Indicate the demand for the p - type goods at node j; Indicate the carriage of vehicle k Whether it meets the demand for the p - type goods of customer i; Indicate the time - window penalty cost; Indicate the time window of customer point i, where, Indicate the earliest arrival time required for the delivery vehicle to reach node i, Indicate the latest arrival time required for the delivery vehicle to reach node i; Indicate the arrival time of delivery vehicle k at node i; 、 Indicate the penalty coefficients for the vehicle arriving early and late respectively; Indicate the refrigeration cost.

[0111] Based on vehicle resources and customer information, the constraint conditions for constructing a multi - temperature co - distribution path planning model are as follows:

[0112]

[0113] Indicate that the vehicle departs from the distribution center, serves the customers, and finally returns to the distribution center;

[0114]

[0115] Indicate that when a vehicle passes through arc (i, j), customer j receives the delivery service;

[0116]

[0117] Indicate the vehicle flow balance of visiting customers;

[0118]

[0119] It means that the delivery demand of each customer's goods is not greater than the capacity of the corresponding delivery vehicle's carriage;

[0120]

[0121] It means that each customer's goods can only be delivered by one vehicle;

[0122]

[0123] It means the time relationship between the time predecessor and successor nodes when the vehicle arrives at customer j;

[0124]

[0125] It means that the customer demand is non - negative;

[0126]

[0127] It means that the customer demand cannot be split;

[0128]

[0129] It means to eliminate sub - circuits;

[0130]

[0131] It means that the vehicle departure time must be within the time window of the distribution center;

[0132]

[0133] It means that all vehicles must return to the distribution center within the specified time;

[0134]

[0135] It is a decision variable, which is 1 if vehicle k travels from customer i to customer j, otherwise 0;

[0136]

[0137] It is a decision variable, which means that if the carriage of vehicle k meets the demand for p - type goods of customer i, it is 1, otherwise 0;

[0138] Among them, N represents the node set composed of customer points and the distribution center, i, j ∈ N, when i = 0, it represents the distribution center; It means whether to depart from the distribution center to node j; It means whether to return to the distribution center from a certain node i; K represents the set of delivery vehicles; It means whether delivery vehicle k travels from node i to the next node j; Indicate whether the p - type carriage of vehicle k meets the p - type commodity demand of customer j; p represents the p - type carriage of the distribution vehicle, p = 1 represents the refrigerated type, and p = 2 represents the frozen type; Indicate whether vehicle k travels from node j to node i; Indicate the demand for p - type goods at node i; Indicate the capacity of the distribution vehicle; Indicate the departure time of distribution vehicle k from point j; Indicate the departure time of distribution vehicle k from point i; Indicate the time required to travel from node i to node j; Indicate the service time of node j; S represents the set of path vertexes; Indicate the departure moment of distribution vehicle k from the distribution center; Indicate the delivery time window of the distribution center; Indicate the earliest time required by the distribution center for distribution vehicle k; Indicate the latest time; m represents the number of customers; Indicate whether distribution vehicle k travels from node i to the next node j; Indicate whether the p - type carriage of distribution vehicle k meets the p - type commodity demand of customer i; Indicate the time of departure from the last customer point of the path; Indicate the latest time to return to the distribution center.

[0139] Step 3: Use the enhanced multi - strategy spider - bee algorithm that combines chaotic mapping, adaptive elite strategy, and reciprocating population reduction strategy to solve the objective of the multi - temperature co - distribution path planning model, and obtain the optimal departure time and distribution path. Figure 2 As shown, perform the objective solution of the multi - temperature co - distribution path planning model using the enhanced multi - strategy spider - bee algorithm that combines chaotic mapping, adaptive elite strategy, and reciprocating population reduction strategy to obtain the optimal departure time and distribution path.

[0140] Use the enhanced multi - strategy spider - bee algorithm that combines chaotic mapping, adaptive elite strategy, and reciprocating population reduction strategy to solve the objective of the multi - temperature co - distribution path planning model, specifically:

[0141] Initialize the parameters, including: population size G, minimum population size , upper bound ub, lower bound nb, number of iterations , trade - off rate TR, crossover rate Cr, number of mutations of the current optimal individual p_m, number of mutations of the global optimal individual g_m; Define and initialize the current iteration number as T = 0.

[0142] Adopt the tent chaotic mapping initialization strategy to generate the initial spider - bee population, and perform encoding and decoding on the generated population. Calculate the fitness value set of the initial population using the multi - temperature co - distribution path planning model , from the fitness value set of the initial population Select the smallest fitness value The corresponding spider wasp individual is denoted as the optimal spider wasp individual in the initial population and the global optimal spider wasp individual The corresponding fitness value is the global optimal fitness value .

[0143] Adopt the tent chaos mapping initialization strategy to generate the initial spider wasp population, and perform encoding and decoding on the generated population. Specifically:

[0144] Randomly generate an initial matrix with 2M rows and G columns, and use the tent chaos mapping to update the initial matrix to generate the initial spider wasp population as follows:

[0145]

[0146] Among them, represents the m-th value of the g-th spider wasp individual; the chaos parameter ; represents the m-th value of the g-th parameter in the initial matrix; represents the population size; represents the chaos variable; rand represents a random number; ub and nb represent the upper and lower bounds respectively; the generated initial spider wasp population is as follows:

[0147]

[0148] Among them, each row represents a spider wasp individual , ; each spider wasp individual represents a set of solutions, and the spider wasp population is simplified to ;

[0149] For any spider wasp individual , represents the departure time, represents the customer node; introduce the location-order encoding method to decode the delivery path of the last M nodes, and sort in ascending order, record the numbers and then convert them to natural number encoding. This part of the encoding represents the customer point numbers; allocate the paths according to the vehicle capacity constraint. If the sum of the customer point demands is satisfied, add the customer point to the path, otherwise generate a new path until all nodes are allocated; for the vehicle departure time , then directly allocate them in sequence from front to back according to the number of generated paths.

[0150] Enter the main loop, generate a random number as the search agent index, , if , execute the prey behavior to update the spider wasp population, otherwise execute the mating behavior to update the spider wasp population; when executing the prey behavior to update the spider wasp population, let , when , generate a random number p, , when , use the search behavior to update the individuals in the spider wasp population. When , adopt the following and escaping behaviors to update the individuals in the spider wasp population; when , use the nest-building behavior to update the individuals in the spider wasp population, generate random numbers , , , when , use the current optimal solution to update the individuals in the spider wasp population. When , randomly select a spider wasp in the population and use an additional step size to update the individuals in the spider wasp population; when executing the mating behavior to update the spider wasp population, randomly select spider wasp individuals , , to generate parameters , , use the parameters , to generate new spider wasp individuals. For each newly generated spider wasp individual in the population, generate a random number rand. When rand < Cr, update the corresponding individual in the population to the new spider wasp individual, otherwise do not update.

[0151] When executing the prey behavior to update the spider wasp population, use the search behavior to update the individuals in the spider wasp population. Specifically:

[0152] Generate random numbers , , ; when , generate a parameter , , where rn1 is a random number; randomly select two spider wasp individuals , from the population, and update the parameters in each spider wasp individual using the following formula:

[0153]

[0154] When , generate a parameter , , where , , , rand represents a random number; update the parameters in each spider wasp individual using the following formula:

[0155]

[0156] Use the following follow and escape behaviors to update the individuals in the spider wasp population:

[0157] When is true, randomly select an individual spider wasp from the population , generate the parameter C,

[0158] , and update the parameters in each individual spider wasp using the following equation:

[0159]

[0160] When is true, generate the parameter matrix , , where , and update the parameters in each individual spider wasp using the following equation:

[0161]

[0162] Use the following nesting behavior to update the individuals in the spider wasp population:

[0163] When is true, update the spider wasp individuals using the globally optimal spider wasp individual as follows:

[0164]

[0165] When is true, randomly select spider wasp individuals , , , and generate new spider wasp individuals using the following equation:

[0166]

[0167] where represents the parameter generated by Levy flight;

[0168] When performing mating behavior to update the spider wasp population, randomly select spider wasp individuals , , to generate the parameters , , specifically:

[0169] Calculate the matrix , randomly select a spider wasp individual , ;

[0170] Calculate , randomly select a spider wasp individual 、 , ;

[0171] Using parameters 、 to generate new spider wasp individuals. For each newly generated spider wasp individual in the population, generate a random number rand. When rand < Cr, update the corresponding individual in the population to the new spider wasp individual using the following formula, otherwise do not update;

[0172]

[0173] Boundary check. Check whether the new spider wasp position exceeds the upper and lower bounds ub and lb of the search space. If so, correct it to the boundary to obtain the updated spider wasp population in the T-th generation , .

[0174] Encode and decode the currently obtained new spider wasp individuals, and calculate the fitness value set of all individuals in the spider wasp population according to the multi-temperature co-allocation path planning model , and compare the fitness values in the current set with the fitness values of the corresponding individuals in the spider wasp population before update. If the current one is better, update the corresponding individual in the population to the current spider wasp individual, otherwise retain the original individual. Compare them in turn to update the spider wasp population ; Encode and decode the currently obtained new spider wasp individuals, calculate the fitness value set of all individuals in the updated spider wasp population, and compare the optimal fitness value with the global optimal fitness value . If the current fitness value is better, update the optimal spider wasp individual and the optimal fitness value to the current spider wasp individual and the corresponding fitness value

[0175] Let the current iteration number , and perform elite mutation p_m times on each individual in the current spider wasp population using the adaptive elite mutation strategy. Check whether the new spider wasp position exceeds the upper and lower bounds ub and lb of the search space. If so, correct it to the boundary; perform elite mutation g_m times on the optimal spider wasp individual in the new population . If the fitness value of each mutated individual is better, update the corresponding spider wasp individual and the fitness value to the current spider wasp individual and its fitness value to obtain the new spider wasp population after mutation 。

[0176] Perform elite mutation on each individual in the current spider wasp population using an adaptive elite mutation strategy as follows:

[0177]

[0178] Among them, 、 represents randomly selecting a spider wasp individual from the population.

[0179] Compare with If is better, then update the optimal spider wasp individual and the optimal fitness value to the current spider wasp individual and the corresponding fitness value, and update the population size G using the reciprocating population reduction strategy until the number of reciprocating cycles is reached. Then, determine whether the current iteration number T is greater than If so, output the corresponding optimal fitness value and the optimal spider wasp individual Otherwise, return to the main loop to continue repeating the steps for updated solution.

[0180] Update the population size G using the reciprocating population reduction strategy as follows:

[0181]

[0182] Among them, % is the remainder operator, represents the number of reciprocating cycles.

[0183] Example 2:

[0184] Example 2 of the present invention discloses a specific application of a multi-temperature co-allocation path planning method based on an enhanced multi-strategy spider wasp algorithm as follows:

[0185] Since there is no internationally recognized instance of MCVRPTW, in order to verify the effectiveness of the algorithm of the present invention, the VRPTW examples of Solomon (1987) are modified into available MCVRP examples. In the experiment, the refrigeration demand and freezing demand are obtained by splitting the customer demand according to a ratio of 2:1. In the experiment, the capacities of the refrigerated compartment and the freezer compartment with flexible compartments are not considered, and the capacities of the original compartments are divided according to a ratio of 3:1. In order to finely simulate the ambient temperature change corresponding to time, in view of the consistent time windows of the selected example distribution centers, here the time window of each distribution center is uniformly divided into 12 time periods, each time period represents 1 hour and corresponds to a temperature data, and on this basis, all customer time window data in the original example are modified accordingly to conform to the new time division scheme, and the vehicle speed is also changed to the value corresponding to one time period. In this experiment, the first 4 of the C1-class data sets with scales of 25, 50, and 100, a total of 12 examples, are selected for the modification experiment. The examples of each scale are respectively called S1-S4, R1-R4, and L1-L4. The algorithm is programmed using MATLAB R2023b, and all experiments are carried out on a server cluster with an Intel(R) Xeon(R) Gold 6258R CPU @ 2.70GHz, a memory of 256GB, and an operating system of CentOS Linux release 7.6.1810. The relevant parameters of the embodiment are shown in Table 1.

[0186] Table 1 Relevant parameters of the embodiment

[0187]

[0188] In order to verify the effect of the enhanced multi-strategy spider bee optimization algorithm proposed by the present invention in solving the path optimization problem, experimental tests are carried out, and the results are as Figure 3 shown. These calculation results further prove the superiority of the proposed algorithm. Figure 3 Intuitively shows the optimized path and service order. It can be clearly seen that all customer points have been successfully served, indicating that while improving the service quality, the optimization algorithm effectively reduces the overall distribution cost. These results show that the enhanced multi-strategy spider bee optimization algorithm not only achieves significant cost optimization on the premise of ensuring service quality, but also can stably obtain high-quality solutions in multiple experimental rounds, fully verifying the robustness and effectiveness of the algorithm.

[0189] In order to prove the advantage of considering flexible compartments compared with traditional fixed compartments, the algorithm of the present invention is run 10 times in 12 examples to obtain the optimal results. The comparison of the number of dispatched vehicles with and without considering flexible compartments is as Figure 4 shown. From Figure 4It can be seen that compared with the fixed compartment model, the flexible compartment model reduces the average number of vehicles used by 2, and can reduce up to 4 vehicles, which directly reduces the fixed costs of vehicle use. Secondly, in terms of the load factor, considering the comparison of the load factors of vehicles with and without flexible compartments, as Figure 5 shown. From Figure 5 it can be seen that the application of flexible compartments has also brought significant improvements. Whether it is small-scale, medium-scale or large-scale examples, the load factors have been greatly improved, from an average of 57.5%, 71.67%, 69.69% to 76.76%, 86%, 89.85% respectively. This improvement means that each vehicle can carry more goods in a single delivery, thus reducing the number of vehicle round trips and the empty running rate. At the same time, the increase in the load factor also enhances the stability and reliability of the distribution system.

[0190] To prove the advantages of optimizing the departure time, the algorithm was run 10 times in each example to obtain the optimal solution, and the comparison of the total costs with and without optimizing the departure time is shown in Figure 6 as shown. From Figure 6 the results, it can be seen that optimizing the departure time has a positive impact on the total cost. In examples of different scales, the optimized total cost is reduced by an average of 4.34% (large scale), 4.08% (medium scale), and 0.67% (small scale). This reduction is mainly due to the better matching of the external temperature changes encountered by the vehicle during transportation through precise adjustment of the departure time, thus reducing the refrigeration demand, optimizing the refrigeration efficiency, and avoiding the energy waste of over-refrigeration.

[0191] To verify the superiority and performance of the proposed enhanced multi-strategy spider wasp algorithm in model solving, it was compared with the newly developed heuristic algorithms in recent years, namely the adaptive spiral flight sparrow search algorithm (ASFSSA), tuna optimization algorithm (TSO), and moth swarm optimization algorithm (MSA), which are also widely used and perform excellently. At the same time, the particle swarm optimization algorithm (PSO) was selected as a classic benchmark for comparison to demonstrate the improvement space and advantages of the algorithm of the present invention compared with the classic algorithm. When each algorithm was run under example L4, the results are shown in Figure 7 as shown. It can be seen that although PSO and MSA converge relatively fast in the early stage, due to their weak initial solution generation ability and tendency to fall into local optima, it is difficult to further optimize in the later stage. In contrast, EMSWO, ASFSSA, and TSO have more advantages in the early stage and can quickly approach the optimal solution region. Among them, EMSWO shows a faster downward trend and exhibits significant convergence advantages in the initial stage of iteration. Compared with ASFSSA and TSO, the final convergence value of EMSWO is lower, indicating its stronger optimization ability. In addition, EMSWO tends to be stable in the later stage of iteration and does not show obvious fluctuations or rebounds, demonstrating excellent solution stability and convergence reliability.

[0192] The embodiment of the present invention discloses a multi-temperature co-distribution path planning method based on an enhanced multi-strategy spider bee algorithm. The present invention enriches the research content of flexible compartments in multi-temperature co-distribution path optimization, and proposes a dynamic adjustment method based on flexible compartments, so that distribution path planning is more flexible and efficient. Flexible compartments can adjust space allocation according to the temperature control requirements of different goods, improve loading efficiency and resource utilization in a multi-temperature zone co-distribution environment, and thus provide a more intelligent and adaptable solution for cold chain distribution. By optimizing the vehicle departure time strategy, the present invention effectively reduces energy consumption during transportation, especially in the face of long-distance, multi-temperature co-distribution cold chain transportation scenarios, optimizing the departure time can effectively avoid high temperature periods, reduce the vehicle's stay time in a high temperature environment, and thus reduce the refrigeration burden. The optimized departure time not only improves transportation efficiency, but also greatly reduces energy consumption, and enhances the environmental protection and economy of cold chain transportation. The present invention provides a comprehensive path optimization and energy efficiency management solution for the distribution of cold chain products such as pre-prepared dishes, combining flexible compartments with departure time optimization, which not only improves transportation efficiency, but also reduces refrigeration energy waste by accurately controlling the temperature control environment. By implementing this plan, companies can improve transportation efficiency and resource utilization while ensuring the quality of cold chain distribution, and promote the industry to develop in a more efficient and environmentally friendly direction.

[0193] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0194] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A multi-temperature co-distribution path planning method based on an enhanced multi-strategy spider bee algorithm, characterized in that: Including: Step 1: Obtain the vehicle resources, customer demands and location information of the distribution center, and establish a refrigeration cost calculation function based on the change of environmental temperature; Step 2: Based on the refrigeration cost calculation function, construct a multi-temperature co-distribution path planning model based on flexible compartments and environmental temperature with the goal of minimizing the total distribution cost, and construct the constraint conditions of the multi-temperature co-distribution path planning model based on vehicle resources and customer information; Step 3: Use an enhanced multi-strategy spider wasp algorithm that combines chaotic mapping, adaptive elite strategy and reciprocating population reduction strategy to solve the objective of the multi-temperature co-distribution path planning model, and obtain the optimal departure time and distribution path; In the enhanced multi-strategy spider wasp algorithm that combines chaotic mapping, adaptive elite strategy and reciprocating population reduction strategy, a tent chaotic mapping initialization strategy is used to generate the initial spider wasp population, and the generated population is encoded and decoded. Specifically: Randomly generate an initial matrix with 2M rows and G columns, and use the tent chaotic mapping to update the initial matrix to generate the initial spider wasp population as follows: x g,m =rand*(ub-nb)+nb; Among them, g，m represents the mth value of the gth spider bee individual; the chaos parameter δ∈(0,1]; x g，m represents the mth value of the gth parameter in the initial matrix; g = 1, 2, 3, ..., G represents the population size; m = 1, 2, ..., M represents the chaotic variable; rand represents a random number; ub and nb represent the upper and lower bounds respectively; the generated initial spider bee population is as follows: Each row represents a spider wasp individual. Each spider bee individual represents a set of solutions, simplifying the spider bee population to For any spider bee individual a g , o g，1 , o g，m ,…,o g，M Indicates departure time, o g，M+1 ,…,o g，M represents the customer node; the position-order coding method is introduced to decode the delivery path of the last M nodes, and o g，M+1 ,…,o g，M Sort in ascending order, record the number and then convert it into a natural number code, which represents the customer point number; assign the route according to the vehicle capacity constraint, if the sum of the customer point demand is satisfied, then add the customer point to the route, otherwise generate a new route until all nodes are assigned; for the vehicle departure time o g，1 , o g，m ,…,o g，M , then directly allocate them from front to back according to the number of generated paths; Use the adaptive elite mutation strategy to perform elite mutation on each individual in the current spider wasp population as follows: in, represents a random selection of a spider wasp individual from the population; Use the reciprocating population reduction strategy to update the population size G as follows: Where, % is the remainder operator, and ε represents the number of reciprocating cycles.

2. The multi-temperature co-distribution path planning method based on the enhanced multi-strategy spider bee algorithm according to claim 1 is characterized in that: In Step 1, a refrigeration cost calculation function is established based on the change of environmental temperature. Specifically: Divide the environmental temperature by hour, take each hour as a time zone, and calculate the environmental temperature of each time zone as follows: Among them, T m Indicates the ambient temperature of time zone m; m indicates the time of the time zone; T max Indicates the highest temperature in a day; T min Indicates the lowest temperature in a day; Temp indicates the simulated Temp temperature value per hour; Δ is a fixed parameter; Assume that the temperature is constant within each hour, and consider the vehicle to pass through one or more time zones. Suppose vehicle k departs from point i, and the departure time is denoted as t ik , the next node visited is denoted as j, and the travel time from point i to point j is The service time at point j is represented by w j , the heat transfer area of ​​the refrigerated and frozen compartments is as follows: Among them, c ij represents the distance from node i to node j; v represents the speed of vehicle k; N represents the node set consisting of the customer point set and the distribution center, i, j∈N, when i=0, it represents the distribution center; p represents the p-type compartment of the distribution vehicle, p=1 represents the refrigerated type, and p=2 represents the frozen type; represents the demand for refrigerated goods at node j; represents the demand for type p cargo at node j; Z represents the total heat transfer area of ​​the carriage; represents the demand for refrigerated goods at node j; Time t ik The time zone m is represented by Q m =[t m , t m+1 ], where t m ,t m+1 are the departure time t ik The start and end time points of the time zone m, t m+1 It also represents the start time of time zone m+1. Since the end time of time zone m coincides with the start time of time zone m+1, the end time of time zone m is represented by the start time of time zone m+1. m The driving and service time is denoted as t m+1 -t ik , judge f ij +w j ≤t m+1 -t ik whether it is established; If established, then from t ik From time to departure point j, the vehicles are at temperature T m The cooling cost during the period of driving is expressed as: H ijk =R1(S1(T m -T1)+S2(T m -T2))(f ij +w j ); Among them, T m represents the temperature in time zone m; R1 represents the heat transfer coefficient of the compartment; T1 represents the temperature of the refrigerated compartment; T2 represents the temperature of the frozen compartment; C f represents the unit fuel supply level; K represents the distribution vehicle set; C 制冷 represents the cooling cost; H ijk represents the heat generated by the refrigerator in the process of delivery vehicle k from node i to node j; E represents the heat generated by a unit of gasoline; If not, then in the time period [t ik , t m+1 ] The vehicle is at temperature T m Driving, H during this period ijk Expressed as: H ijk =R1(S1(T m -T1)+S2(T m -T2))(t m+1 -t ik ); Update f ij =f ij -(t m+1 -t ik ), t ik =t m+1 , re-judge f ij +w j ≤t m+1 -t ik Is it established to obtain the refrigeration cost.

3. The multi-temperature co-distribution path planning method based on the enhanced multi-strategy spider bee algorithm according to claim 1 is characterized in that: In Step 2, based on the refrigeration cost calculation function, construct a multi-temperature co-distribution path planning model based on flexible compartments and environmental temperature with the goal of minimizing the total distribution cost as follows: minf=C1+C2+C3+C4+C 制冷 ; Among them, C1 represents fixed cost; K represents the set of delivery vehicles; C k represents the fixed resource consumption of starting the delivery vehicle k; x 0jk Indicates whether the vehicle departs from the distribution center; C2 represents the transportation cost; N represents the node set consisting of the customer point set and the distribution center, i,j∈N, when i=0, it represents the distribution center; x kij Indicates whether the delivery vehicle k travels from node i to the next node j; Q represents the capacity of the delivery vehicle; Q k represents the initial quantity of cargo loaded when delivery vehicle k departs from the distribution center; Indicates the fuel consumption rate of the delivery vehicle when it is fully loaded; Indicates the fuel consumption rate of the delivery vehicle when it is unloaded; c ij represents the distance from node i to node j; c f represents the unit fuel supply level; C3 represents the loss cost; p represents the p-type compartment of the distribution vehicle, p=1 represents the refrigerated compartment, and p=2 represents the frozen compartment; λ p represents the damage rate of goods of category p; L represents the unit price of goods per unit quantity; f ij represents the time required to travel from node i to node j; w j represents the service time of node j; represents the demand for type p goods at node j; y ikp represents the compartment u of vehicle k k，p Whether the demand for p-type goods of customer i is met; C4 represents the time window penalty cost; [ET i , LT i ] represents the time window of customer point i, where ET i represents the earliest arrival time required by node i for the delivery vehicle to arrive, LT i represents the latest arrival time required by node i for the delivery vehicle to arrive; e ik represents the arrival time of delivery vehicle k to node i; α and β represent the penalty coefficients for early and late arrival of vehicles respectively; C 制冷 Represents the cooling cost.

4. The multi-temperature co-distribution path planning method based on the enhanced multi-strategy spider bee algorithm according to claim 1 is characterized in that: In Step 2, construct the constraint conditions of the multi-temperature co-distribution path planning model based on vehicle resources and customer information as follows: Where N represents the node set consisting of the customer point set and the distribution center, i,j∈N, when i=0, it represents the distribution center; x 0jk Indicates whether to start from the distribution center and rush to node j; x i0k Indicates whether to return to the distribution center from a certain i node; K represents the set of distribution vehicles; x ijk Indicates whether the delivery vehicle k travels from node i to the next node j; y jkp Indicates whether the p-type compartment of vehicle k meets the p-type commodity demand of customer j; p represents the p-type compartment of the delivery vehicle, p=1 represents refrigerated compartment, and p=2 represents frozen compartment; x jik Indicates whether vehicle k goes from node j to node i; represents the demand for type p goods at node i; Q represents the capacity of the delivery vehicle; t jk represents the time when delivery vehicle k departs from point j; t ik represents the time when the delivery vehicle k departs from point i; f ij represents the time required to travel from node i to node j; w j represents the service time of node j; S represents the set of path vertices; t 0k represents the departure time of delivery vehicle k from the distribution center; [ET0, LT0] represents the delivery time window of the distribution center, ET0 represents the earliest time when the distribution center requires delivery vehicle k; LT0 represents the latest time; m represents the number of customers; x kij Indicates whether the delivery vehicle k travels from node i to the next node j; y ikp Indicates whether the p-type compartment of delivery vehicle k meets the p-type commodity demand of customer i; t mk It indicates the departure time from the last customer point on the route; t0 indicates the latest return time to the distribution center.

5. The multi-temperature co-distribution path planning method based on the enhanced multi-strategy spider bee algorithm according to claim 1 is characterized in that: In Step 3, use an enhanced multi-strategy spider wasp algorithm that combines chaotic mapping, adaptive elite strategy and reciprocating population reduction strategy to solve the objective of the multi-temperature co-distribution path planning model. Specifically: Initialization parameters, including: population size G, minimum population size G min , upper bound ub, lower bound nb, number of iterations T max , trade-off rate TR, crossover rate Cr, current optimal individual mutation times p_m, global optimal individual mutation times g_m; define and initialize the current iteration times as T=0; The tent chaotic mapping initialization strategy is used to generate the initial spider bee population, and the generated population is encoded and decoded, and the fitness value set of the initial population is calculated using the multi-temperature co-matching path planning model. From the fitness value set F of the initial population * Select the minimum fitness value The corresponding spider bee individual is recorded as the optimal spider bee individual in the initial population and the global optimal spider bee individual C best , the corresponding fitness value is the global optimal fitness value S best ; Enter the main loop, generate a random number r6 as the search agent index, r6 ∈ (0, 1). If r6 < TR, execute the prey behavior to update the spider wasp population; otherwise, execute the mating behavior to update the spider wasp population. When executing the prey behavior to update the spider wasp population, let When g < G * K, generate a random number p, p ∈ (0, 1). When p < (1 - T / T max ), use the search behavior to update the individuals in the spider wasp population. When p > (1 - T / T max ), adopt the following and escape behaviors to update the individuals in the spider wasp population. When g > G * K, use the nesting behavior to update the individuals in the spider wasp population, generate random numbers r3, r4, r3, r4 ∈ (0, 1). When r3 < r4, use the current optimal solution to update the individuals in the spider wasp population. When r3 > r4, randomly select a spider wasp in the population and use an additional step size to update the individuals in the spider wasp population. When executing the mating behavior to update the spider wasp population, randomly select a spider wasp individual to generate parameters V1, V2, and use the parameters V1, V2 to generate a new spider wasp individual. For each newly generated spider wasp individual in the population, generate a random number rand. When rand < Cr, update the corresponding individual in the population to the new spider wasp individual; otherwise, do not update. The currently obtained new spider bee individuals are encoded and decoded, and the fitness value set of all individuals in the spider bee population is calculated according to the multi-temperature co-matching path planning model. Compare the current collection separately The fitness value of the spider bee population before the update is compared with the fitness value of the corresponding individual in the spider bee population. If the current one is better, the corresponding individual in the updated population is the current spider bee individual. Otherwise, the original individual is retained and compared in turn to update the spider bee population A. (T) ; Encode and decode the new spider bee individuals currently obtained, calculate the fitness value set of all individuals in the updated spider bee population, and take the optimal fitness value And the global optimal fitness value S best For comparison, if the current fitness value is better, then update the optimal spider bee individual C best And the optimal fitness value S best is the current spider bee individual and the corresponding fitness value; Let the current number of iterations T = T + 1, and use the adaptive elite mutation strategy to perform elite mutation p_m times on each individual in the current spider bee population. Check the boundary to see if the new spider bee position exceeds the upper and lower bounds ub and nb of the search space. If so, correct it to the boundary. Use the adaptive elite mutation strategy to modify the new population A (T) The best spider bee individual in the population undergoes elite mutation g_m times. If the fitness value of the individual after each mutation is better, the corresponding spider bee individual in the population is updated. and fitness value is the current spider bee individual and its fitness value, and the new spider bee population A after mutation is obtained (T) ; Compare With S best The size of Better, then update the optimal spider bee individual C best And the optimal fitness value S best is the current spider bee individual and its corresponding fitness value, and uses the reciprocating population reduction strategy to update the population size G until the number of reciprocating cycles is reached, and then determines whether the current number of iterations T is greater than T max , then the output corresponds to the optimal fitness value S best and the optimal spider bee individual C best , otherwise return to the main loop to repeat the steps and update the solution.

6. The multi-temperature co-distribution path planning method based on the enhanced multi-strategy spider bee algorithm according to claim 5 is characterized in that: When performing prey behavior to update the spider wasp population, use search behavior to update the individuals in the spider wasp population. Specifically: Generate random numbers r1, r2, r1, r2∈(0,1); when r1>r2, generate parameter m1, m1=|rn1|*r1, where rn1 is a random number; randomly select two spider wasp individuals from the population Update the parameters of each spider bee individual using the following formula: When r1<r2, the parameter m2 is generated, m2=B*cos(2lπ), where l=(α2-1)*rand+1,a2=-1-1*(T / T max ), rand represents a random number; the following formula is used to update the parameters of each spider bee individual: Use following and escaping behaviors to update the individuals in the spider wasp population. Specifically: When r1>r2, randomly select spider wasp individuals from the population Generate parameter C, a=2-2*(T / T max ), and update the parameters of each spider bee individual using the following formula: When r1<r2, the generated parameter matrix V={v,v2,v m , …, v M },v m ∈(-k, k), where Update the parameters of each spider bee individual using the following formula: Use nesting behavior to update the individuals in the spider wasp population. Specifically: When r1>r2, the global optimal spider bee individual C best Updated the spider bee individuals as follows: When r1<r2, randomly select spider bee individuals Generate new spider bee individuals using the following formula: Where, χ represents the parameter generated by levy flight.

7. The multi-temperature co-distribution path planning method based on the enhanced multi-strategy spider bee algorithm according to claim 5 is characterized in that: When performing mating behavior to update the spider bee population, randomly select spider bee individuals Generate parameters V1 and V2, specifically: Calculate the matrix V1 = {v1, ..., v m , …, v M }, randomly select a spider bee individual Calculate V2 = {z1, ..., z m , …, z M }, randomly select a spider bee individual Generate new spider wasp individuals using parameters V1 and V2. For each newly generated spider wasp individual in the population, generate a random number rand. When rand < Cr, update the corresponding individual in the population to the new spider wasp individual using the following formula, otherwise do not update; Boundary check: check whether the new spider bee position exceeds the upper and lower bounds ub and nb of the search space. If so, correct it to the boundary to obtain the updated spider bee population A of the Tth generation (T) ,

Images