CN108537491A - A kind of fresh agricultural products Distribution path optimization method, storage medium - Google Patents

Abstract

A kind of fresh agricultural products Distribution path optimization method of present invention offer and storage medium, the information such as the position coordinates of oneself, the car loading of required dispatching and required distribution time range are submitted to system by each demand point, to minimize distribution cost and maximize customer satisfaction as optimization aim, fresh agricultural products Distribution path Model for Multi-Objective Optimization is established.Simultaneously using based on adaptive genetic algorithm solving model, Distribution path prioritization scheme is obtained.The present invention establishes the speed characteristic model under different weather situation, different periods；Establish the time window punishment cost function of fresh agricultural products；Construct the Model for Multi-Objective Optimization of fresh agricultural products Distribution path；And condition of road surface when can be according to dispatching formulates distribution route, to reach distribution cost minimum, the maximum target of customer satisfaction.

Description

A distribution route optimization method and storage medium for fresh agricultural products

technical field

The invention belongs to the field of logistics distribution, and in particular relates to a method for optimizing distribution routes of fresh agricultural products and a storage device.

Background technique

With the rapid development of social economy and the continuous improvement of residents' living standards, people have higher and higher requirements for the quality of fresh agricultural products. Fresh agricultural products mainly include fruits and vegetables, aquatic products, meat, flowers, etc., which are perishable and fragile, and have a short shelf life, which puts forward higher control requirements for the circulation of fresh agricultural products. Therefore, how to scientifically and rationally arrange distribution routes to ensure the freshness of fresh agricultural products, improve distribution efficiency, and reduce distribution costs is one of the important issues faced in the logistics and distribution of fresh agricultural products.

At present, domestic and foreign scholars have carried out a lot of research work on the optimization of logistics distribution path of fresh agricultural products. Some methods have established the vehicle routing problem of fresh food logistics distribution with a time window on the basis of comprehensive consideration of various factors such as distribution distance, vehicle fixed cost, and fresh food loss; some methods use fuzzy membership function to express customer satisfaction at distribution points degree, and established a multi-objective distribution route optimization model that minimizes distribution costs and maximizes customer satisfaction; the above studies all assume that the travel time and transportation costs of distribution vehicles are only related to the distribution distance, ignoring the impact of different road conditions on vehicle speed. and delivery costs. For the cold chain logistics distribution route optimization model under time-varying conditions, there are some methods that consider the dynamic driving speed and improve the tabu search algorithm to find the balance point between the distribution service quality and the distribution cost. There are also some methods that build an optimization model of cold chain logistics distribution according to the traffic conditions of each delivery section in different time periods, and design a hybrid genetic algorithm to solve the model. There are also some methods combined with real-time traffic information to study the path optimization of intra-city cold chain logistics distribution. Although the above studies considered the time-varying speed of delivery vehicles in the optimization process, they did not establish the connection between different road conditions and the delivery optimization model.

Contents of the invention

The invention provides a method for optimizing distribution routes of fresh agricultural products in consideration of road conditions, so as to solve the problems existing in the prior art.

The present invention adopts following technical scheme:

A method for optimizing distribution routes of fresh agricultural products, comprising:

(1) With the goal of minimizing transportation costs and maximizing customer satisfaction with logistics services, the path optimization target model for fresh agricultural product distribution is established:

s.t.

Among them, _Z1 represents the delivery cost, Z2 represents customer satisfaction; N represents the number of distribution points, k represents the kth vehicle, i represents the i-th distribution point, j represents the j-th distribution point, t _ij represents the k-th vehicle from the The time from delivery point i to delivery point j, Indicates the demand of delivery point i, Indicates the price of a unit of fresh agricultural products, express, Indicates the sensitivity of fresh agricultural products to time, C ₃ indicates the time window penalty cost, and t _i indicates the time when the vehicle arrives at delivery point i;

Indicates the operating cost of the delivery vehicle;

v _k is a 0-1 variable, when the kth car is used, v _k is 1, otherwise v _k is 0;

x _ijk is a 0-1 variable, when the kth vehicle travels from delivery point i to delivery point j, x _ijk is 1, otherwise,

x _ijk is 0;

(2) Use the genetic algorithm to solve the model to obtain the distribution plan of fresh agricultural products.

The genetic algorithm for solving the model includes the following steps:

S1: Suppose the population size is L, that is, there are L feasible solutions, and the lth individual is expressed as Among them, r _n (0<r _n <1) is the genetic information of the individual, which is generated by a random function;

in is the number of genes, which is equal to the sum of the number of delivery points and the maximum number of delivery vehicles required:

S2: Build fitness function:

W(g)＝w ₀ ·a _w ^g g＝1,2,…,g _max

Among them, F ₁ represents the fitness of the l-th individual, l=1, 2,...L; Z ₁ and Z ₂ are the delivery cost and customer satisfaction corresponding to the l-th individual respectively; W(g) represents the annealing temperature function , is related to the population algebra g; w ₀ = g _max represents the initial annealing temperature, and g _max represents the maximum evolutionary generation of the population; a _w is the annealing temperature coefficient;

S3: Individual selection: Use the roulette wheel rule to select excellent individuals from the population for crossover and mutation operations:

The crossover probability p _c is expressed as

Among them, F _max represents the maximum fitness in the current population; Represents the average fitness of the population; F' represents the higher fitness value between paired individuals; k' ₁ and k' ₂ represent the crossover parameters;

The mutation probability p _l,m of the lth individual is defined as

Among them, F ₁ represents the fitness of the lth individual; k' ₃ and k' ₄ represent the variation parameters. The algorithm uses Gaussian mutation to realize the mutation operation;

S4: Decoding: During the chromosome decoding process, the algorithm compares each gene in the chromosome, and obtains its position value in the array in ascending order. When the value is greater than the number of delivery points, the value becomes 0; when When there are 0 values adjacent to each other, the deduplication operation is performed to obtain the delivery route plan.

The vehicle transportation cost C1 is expressed as:

Among them, N is the number of distribution points; Indicates the vehicle driving cost per unit time; d _ij is the distance from delivery point i to delivery point j; is the average driving speed of the delivery vehicle; ζ _con is the influence rate of the vehicle speed under different weather conditions and at different time periods, sun, rain, snow, and fog represent sunny days, rainy days, snowy days, and foggy days respectively:

The loss cost of fresh goods is expressed as follows:

in, Indicates the demand of distribution point i; Indicates the price of a unit of fresh agricultural products; Indicates the sensitivity of the quality of fresh agricultural products to time, The larger the value, the lower the sensitivity of product quality to time; t _ik represents the time when the kth vehicle arrives at delivery point i.

The time window penalty cost is expressed as:

Among them, M is a positive number; α _c , β _c (α _c , β _c <0) are weight parameters;

and is the delivery time range that the customer rejects;

is the customer's ideal delivery time frame;

and is the acceptable shipping time frame for the customer.

The customer satisfaction cost of delivery point i is expressed as:

Among them, the parameter is the weight parameter.

A storage medium stores a computer program inside, and when the computer program is read by a processor, the above-mentioned method is executed.

Beneficial effects of the present invention:

(1) The vehicle speed characteristic model under different weather conditions and different time periods is established;

(2) The time window penalty cost function of fresh agricultural products is established;

(3) Constructed a multi-objective optimization model for the delivery route of fresh agricultural products;

(4) The system can formulate delivery routes according to the road conditions during delivery, so as to achieve the goal of minimum delivery cost and maximum customer satisfaction.

Description of drawings

Fig. 1 is a flowchart of the present invention.

Figure 2 is a schematic diagram of the decoding process.

Detailed ways

The present invention will be described in further detail below in conjunction with the accompanying drawings and specific embodiments.

In the actual delivery process, factors such as weather conditions, morning and evening peak hours, and special festivals directly affect the speed of delivery vehicles, which in turn leads to changes in delivery costs and customer satisfaction. Therefore, according to different weather conditions and vehicle speed characteristics at different time periods, the present invention proposes a distribution plan for fresh agricultural products under comprehensive consideration of factors such as road conditions, time windows, and fresh food loss. Maximum satisfaction.

Specific steps are as follows:

1) First, each demand point submits information such as its location coordinates, the quantity of goods to be delivered, and the required delivery time range to the system.

2) To minimize the distribution cost and maximize customer satisfaction as the optimization goal, a multi-objective optimization model for the distribution route of fresh agricultural products was established.

The details are as follows:

①Vehicle transportation cost

The vehicle transportation cost C ₁ includes the operating cost and driving cost during the use of the vehicle, expressed as

Among them, k represents the kth vehicle, and N is the number of distribution points; Indicates the operating cost of the delivery vehicle, mainly including the fixed loss cost of the vehicle and the driver's salary cost; v _k is a 0-1 variable, when the kth vehicle is used, v _k is 1, otherwise, v _k is 0; Indicates the vehicle driving cost per unit time; x _ijk is a 0-1 variable, when the kth vehicle travels from delivery point i to delivery point j, x _ijk is 1, otherwise, x _ijk is 0; d _ij is delivery point i The distance to delivery point j; is the average driving speed of the delivery vehicle; ζ _con (con=sun, rain, snow, fog) is the influence rate of different weather conditions (sunny, rainy, snowy, foggy) and different time periods on the vehicle speed, which are respectively expressed by the formula (2)～Equation (5) expresses,

Among them, t represents a certain time in 24 hours when the vehicle travels.

②Fresh goods loss cost

Fresh agricultural products are easily affected by factors such as temperature and time, and damage costs will be incurred during the distribution process. Considering that fresh agricultural products are transported by cold chain logistics and the temperature is relatively stable, the cargo damage cost C2 can be expressed as

in, Indicates the demand of distribution point i; Indicates the price of a unit of fresh agricultural products; Indicates the sensitivity of the quality of fresh agricultural products to time, The larger the value, the lower the sensitivity of product quality to time; t _ik represents the time when the kth vehicle arrives at delivery point i.

③Time window penalty cost

The establishment of the time window is to require the delivery vehicle to arrive within the time range specified by the customer. According to whether customers have strict requirements on service time, the time window can be divided into hard time window and soft time window. Different from the hard time window, the soft time window allows the delivery vehicle to arrive outside the time range specified by the customer, but a certain penalty fee needs to be paid, and the farther the vehicle arrives at the customer from the specified time range, the higher the penalty fee paid . In order to fit the actual delivery situation, the present invention uses a soft time window to construct a time window penalty cost function for the delivery of fresh agricultural products, as shown in formula (7).

Among them, M is a large positive number; α _c , β _c (α _c , β _c <0) are weight parameters, and the value depends on the customer's demand for delivery time. Equation (7) expresses, and It is the delivery time range that the customer rejects. When the delivery vehicle arrives within this time period, a large penalty fee needs to be paid; It is the customer's ideal delivery time range, and no penalty fee will be paid when the delivery vehicle arrives within this time period; and It is the delivery time range acceptable to the customer. When the delivery vehicle arrives within this time period, a certain penalty fee will be paid. According to the perishable and fragile characteristics of fresh agricultural products, distribution vehicles When the range arrives (earlier than the ideal delivery time), it will not have a great impact on product quality and sales, and the penalty fee changes linearly with time; however, the delivery vehicle is in When the range arrives (later than the ideal delivery time), the remaining shelf life of the product will be shortened, which will have a greater impact on the quality and sales of the product, and the penalty fee will change exponentially with time.

④Customer Satisfaction

Customer satisfaction refers to the degree of customer satisfaction with logistics services, reflecting the logistics service level of fresh agricultural products. In the link of logistics distribution, the main factor affecting customer satisfaction is the time when the delivery vehicle arrives at the delivery point. Therefore, the customer satisfaction function of delivery point i is expressed as,

Among them, the weight parameter The value of depends on the customer's demand for delivery time.

⑤ Model establishment

The present invention takes the minimization of distribution costs and the maximization of customer satisfaction as optimization goals, and establishes a multi-objective optimization model for distribution routes of fresh agricultural products:

In the above optimization problem, constraint condition formula (11) ~ formula (12) means that the distribution vehicle departs from the distribution center and returns to the distribution center after serving the distribution points; constraint condition formula (13) ~ formula (14) means that each The delivery point can only be served by the delivery vehicle once; the constraint condition (15) constrains the loading capacity of the delivery vehicle, where, is the maximum loading capacity of the distribution vehicle, y _ik is a 0-1 variable, when the distribution point i is delivered by the kth vehicle, y _ik is 1, otherwise, y _ik is 0. t _i represents the time when the vehicle arrives at delivery point i.

When the model of the present invention is solved, combined with the idea of simulated annealing, an adaptive genetic algorithm is used to solve the problem, and an optimal distribution route scheme is obtained.

The detailed method is as follows:

① Coding and initializing the population

A random array is used to encode feasible solutions: assuming that the population size is L, that is, L feasible solutions, the lth individual is expressed as Among them, r _n (0<r _n <1) is the genetic information of the individual, which is generated by a random function; is the number of genes, which is equal to the sum of the number of distribution points and the theoretically required maximum number of distribution vehicles.

For example, there are 4 distribution points in the distribution network, and the theoretically required maximum number of distribution vehicles is 4, and the encoded individual is expressed as: [0.25, 0.32, 0.41, 0.40, 0.36, 0.68, 0.16, 0.88].

② Calculation of fitness

The present invention combines simulated annealing to construct the fitness function F _l

W(g)＝w ₀ ·a _w ^g g＝1,2,...,g _max (17)

Among them, F ₁ represents the fitness of the l-th individual, l=1, 2,...L; Z ₁ and Z ₂ are the delivery cost and customer satisfaction corresponding to the l-th individual respectively; W(g) represents the annealing temperature function , is related to the population algebra g; in formula (17), w ₀ =g _max represents the initial annealing temperature, and g _max represents the maximum evolutionary generation of the population; a _w is the annealing temperature coefficient.

③Individual choice

The invention adopts the roulette wheel rule to select excellent individuals from the population for crossover and mutation.

④Cross operation

The present invention defines the crossover probability p _c as

Among them, F _max represents the maximum fitness in the current population; Represents the average fitness of the population; F' represents the higher fitness value among paired individuals; k' ₁ and k' ₂ represent the crossover parameters. The method uses arithmetic crossover to realize the crossover operation.

⑤ mutation operation

The present invention defines the mutation probability p _l,m of the lth individual as

Among them, F ₁ represents the fitness of the lth individual; k' ₃ and k' ₄ represent the variation parameters. The algorithm uses Gaussian mutation to realize the mutation operation.

⑥ decoding

During the chromosome decoding process, the algorithm compares each gene in the chromosome, and obtains its position value in the array in ascending order. When the value is greater than the number of delivery points, the value becomes 0; When adjacent, de-duplicate operations are performed to obtain the delivery route plan (the optimal delivery route refers to the ordering of demand points with the lowest delivery cost and the highest customer satisfaction), where 0 represents the distribution center, and other numbers represent different distribution points. For example, the individual [0.25, 0.32, 0.41, 0.40, 0.36, 0.68, 0.16, 0.88] is decoded and expressed as [2, 3, 0, 4, 0, 1, 0], thus: the delivery paths are 0 -2-3-0; 0-4-0; 0-1-0. The decoding process is shown in Figure 2 below.

The present invention also provides a storage medium, which stores a computer program inside. When the computer program is read by a processor, the path optimization method of the present invention is executed. The storage medium is an existing storage medium, and the storage medium may be connected to the processor or belong to a memory in the processor.

What has been described above is only the preferred embodiment of the present invention. It should be pointed out that for those skilled in the art, some changes and improvements can be made without departing from the overall concept of the present invention, and these should also be regarded as the present invention. scope of protection.

Claims (7)

1. A fresh agricultural product delivery path optimization method is characterized by comprising the following steps:

(1) establishing a path optimization target model of fresh agricultural product delivery by taking the goals of minimizing the transportation cost and maximizing the satisfaction degree of customers to logistics service as targets:

s.t.

wherein Z is₁Indicating delivery cost, Z2 indicating customer satisfaction; n denotes the number of delivery points, k denotes the kth vehicle, i denotes the ith delivery point, j denotes the jth delivery point, t_ijIndicating the time from the ith delivery point to the jth delivery point for the kth vehicle,indicating the amount of demand at the delivery point i,the price of the unit fresh agricultural product is expressed,it is shown that,representing the sensitivity of fresh agricultural products to time, C₃Represents a time window penalty cost, t_iIndicating the time at which the vehicle reaches delivery point i；

Represents the operating cost of the delivery vehicle;

v_kis a variable of 0-1, v is when the k-th vehicle is used_kIs 1, otherwise v_kIs 0;

x_ijkx is a variable from 0 to 1 when the k-th vehicle travels from delivery point i to delivery point j_ijkIs 1, otherwise, x_ijkIs 0;

(2) and solving the model by using a genetic algorithm to obtain a distribution scheme of the fresh agricultural products.

2. The fresh agricultural product delivery path optimization method of claim 1, wherein:

the genetic algorithm for solving the model comprises the following steps:

s1: assuming a population size of L, i.e., L feasible solutions, the ith individual is represented asWherein r is_n(0＜r_n< 1) is the genetic information of the individual, generated by a random function;

whereinIs the number of genes, equal to the sum of the number of delivery points and the maximum number of vehicles required to be delivered:

s2: constructing a fitness function:

W(g)＝w₀·a_w ^gg＝1,2,…,g_max

wherein, F_lDenotes the fitness of the ith individual, L ═ 1,2, L; z₁And Z₂Distribution cost and customer fullness corresponding to the first individualDegree of intention; w (g) represents an annealing temperature function and is related to a population generation number g; w is a₀＝g_maxDenotes initial annealing temperature, g_maxRepresenting the maximum evolutionary algebra of the population; a is_wIs the annealing temperature coefficient;

s3: individual selection: and (3) selecting excellent individuals from the population by adopting a roulette rule to perform cross operation and mutation operation:

cross probability p_cIs shown as

Wherein, F_maxRepresenting the maximum fitness in the current population;representing the average fitness of the population; f' represents a higher fitness value between the paired individuals; k'₁And k'₂Represents a crossover parameter;

mutation probability p of the l-th individual_l,mIs defined as

Wherein, F_lRepresenting the fitness of the ith individual; k'₃And k'₄Representing a variation parameter. The algorithm adopts Gaussian variation to realize variation operation;

s4: and (3) decoding: in the chromosome decoding process, comparing each gene in the chromosome by an algorithm, and acquiring the position values of the genes in the array from small to large, wherein when the value is larger than the number of distribution points, the value is changed into 0; and when 0 values are adjacent, carrying out duplicate removal operation to obtain a distribution path scheme.

3. The fresh agricultural product delivery path optimization method of claim 1, wherein:

the vehicle transportation cost C₁Expressed as:

wherein N is the number of distribution points;represents a vehicle running cost per unit time; d_ijIs the distance from delivery point i to delivery point j;is the average travel speed of the delivery vehicle; zeta_conThe influence rate of the vehicle speed under different weather conditions and different time periods is, sun, rain, snow and fog respectively represent sunny days, rainy days, snowy days and foggy days:

4. the fresh agricultural product delivery path optimization method of claim 1, wherein:

the raw fresh loss cost is expressed as follows:

wherein,representing the demand of the delivery point i;expressing the price of unit fresh agricultural product;the sensitivity of the quality of the fresh agricultural products to the time is shown,the larger the value is, the lower the sensitivity of the product quality to time is; t is t_ikIndicating the time at which the kth vehicle reaches delivery point i.

5. The fresh agricultural product delivery path optimization method of claim 1, wherein:

the time window penalty cost is expressed as:

wherein M is a positive number, α_c，β_c(α_c，β_c< 0) is a weight parameter;

andis the delivery time range rejected by the customer;

is the ideal delivery time range for the customer;

andis the range of delivery times acceptable to the customer.

6. The fresh agricultural product delivery path optimization method of claim 1, wherein:

the customer satisfaction cost for distribution point i is expressed as:

wherein the parametersIs a weight parameter.

7. A storage medium storing a computer program therein, characterized in that: the computer program, when read by a processor, performs the method of any of claims 1 to 6.