Supply chain scheduling optimization method based on transport outsourcing and time window constraint

Abstract

The invention provides a supply chain scheduling optimization method based on transport outsourcing and time window constraint, and relates to the technical field of supply chain scheduling. On the basis of a strict time window constraint framework and under the condition that self transport capacity cannot completely meet order transport requirements, collaborative scheduling is carried out on order production and transport processes, the earnings of order autonomous transport and outsourcing transport are balanced, a production scheduling scheme and a transport scheduling scheme under the outsourcing condition of part of transport tasks are comprehensively considered, and the scheduling efficiency is improved. According to the method, the approximate optimal solution of the problem is dynamically and accurately solved, and collaborative optimization and configuration can be powerfully carried out on resources in two stages of production and transportation of the supply chain, so that the overall cost of the supply chain is reduced to the greatest extent, and the resource collaboration efficiency and management capability of the enterprise supply chain are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of supply chain scheduling, and in particular to a supply chain scheduling optimization method based on transportation outsourcing and time window constraints. Background Art

[0002] In modern supply chain management, enterprises are increasingly adopting a business model of production-to-order. In actual operation, there is little or no finished product inventory, which makes production and transportation plans closely linked to eliminate the cost increase and product circulation time caused by high inventory; enterprises must also ensure that products are delivered on time within a strict time window to reduce the loss of corporate revenue and reputation caused by delayed delivery. In order to meet the time window constraints, enterprises often rely on transportation outsourcing to provide flexible transportation support. Transportation outsourcing has become an important strategy for enterprises to improve efficiency and reduce costs. By outsourcing part of the transportation tasks to professional logistics service providers, enterprises can optimize resource allocation and reduce the risk of delayed delivery. However, the introduction of transportation outsourcing makes the trade-off between cost, benefit and other factors in the supply chain collaborative scheduling plan particularly complicated, and it is necessary to find the optimal balance between independent transportation and transportation outsourcing.

[0003] However, few existing studies have considered the collaborative scheduling of supply chain production and transportation based on transportation outsourcing scenarios. From the perspective of actual supply chain management, current research on collaborative scheduling of production and transportation is difficult to help companies cope with the challenges brought by strict time window constraints and the trade-off between transportation outsourcing benefits, achieve the expected on-time delivery performance at the lowest total cost, and ensure the efficiency and reliability of the overall operation. Therefore, it is of great significance to comprehensively consider the production and transportation coordination issues under the framework of strict time window constraints in the case of outsourcing of some transportation tasks to achieve efficient coordination of the supply chain. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a supply chain scheduling optimization method based on transportation outsourcing and time window constraints, which solves the problem of supply chain production and transportation collaborative scheduling under product transportation outsourcing and strict time window constraints.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0006] A supply chain scheduling optimization method based on transportation outsourcing and time window constraints, the scheduling optimization method comprises the following steps:

[0007] S1. According to the data related to customer orders, production process and transportation process, set the input parameters of the simulated annealing algorithm based on neighborhood search;

[0008] S2. Obtain a global optimal solution based on the input parameters and a simulated annealing algorithm based on neighborhood search, and obtain the minimum total cost of the supply chain, the optimal single-machine scheduling plan for the production process, and the optimal order transportation plan based on the global optimal solution.

[0009] Preferably, in S1, the input parameters of the simulated annealing algorithm based on neighborhood search specifically include:

[0010] Order set I (i, h = 0, 1, 2, ..., l); customer number, order number, distribution point number i, h, where i, h = 0, 1, 2, ..., l, i = 0 is the production center; order production sequence number on the production line j = 1, 2, ..., l; transport vehicle number k = 0, 1, 2, ..., m, k = 0 is transported by outsourced vehicles; transport batch b = 1, 2, ..., n during vehicle patrol distribution;

[0011] The number of products produced per unit time on each production line q;

[0012] The production cost per unit time of each production line is c0;

[0013] The quantity of product N required for order i i ;

[0014] Order i arrival deadline d i ;

[0015] The transport capacity Q of a single autonomous transport vehicle;

[0016] The transportation speed v of autonomous and outsourced transportation vehicles;

[0017] The distance D between the factory and customer i 0i , correspondingly, the distance D between order i and order h ih ;

[0018] The cost coefficient c1 of the autonomous transport vehicle per unit of travel time;

[0019] Unit distance cost coefficient c2 for outsourced transportation per unit of product;

[0020] The penalty factor c3 for each unit of late time;

[0021] Transportation time T between factory J0 and order i 0i , correspondingly, the shipping time between order i and order h is T ih ;

[0022] The complete set A of scheduling plans for a single machine in the production process, a certain feasible scheduling plan π∈A, denoted as π={π(1),π(2),…,π(l)}, where π(i) means that order i is produced at the π(i)th place in the production sequence set π, and π(i)=j;

[0023] Production start time s for order i i ;

[0024] The production end time f of order i i ;

[0025] The complete set of multi-vehicle scheduling solutions for the transportation process B, a certain feasible transportation solution μ∈B, is expressed as μ={μ(1),μ(2),…,μ(l)}, where μ(i) is the transportation vehicle number of the corresponding order i, when μ(i)=0, the transportation of the order is outsourced, and λ(i) is the transportation batch number of the corresponding order i in the transportation vehicle;

[0026] The shipping departure time ST for order i i ;

[0027] Arrival time CT of order i i ;

[0028] The transport order set I corresponding to the b-th transport of vehicle k kb ;

[0029] The departure time of vehicle k for the bth transport from the factory is ST kb ;

[0030] The time CT when vehicle k returns to the factory for the bth transport kb ;

[0031] Production cost SP for order i i ;

[0032] The road transportation cost SA of order i i ;

[0033] Arrival delay cost SL for order i i ;

[0034] Order i outsourced shipping costs SO i ;

[0035] Total supply chain cost ST.

[0036] Preferably, S2 specifically includes:

[0037] S2.1, set the operating parameters of the simulated annealing algorithm and the initial solution θ0, the operating parameters include: initial temperature t ini , termination temperature t final , temperature attenuation function coefficient e;

[0038] S2.2. Obtain and record the new solution θ through perturbation α ;

[0039] S2.3, calculating the energy increment based on the initial solution and the new solution;

[0040] S2.4, if the energy increment is less than 0, the new solution can be used as the current solution, and the current solution is recorded as θ; if the energy increment is not less than 0, the acceptance probability is calculated, and the new solution is received according to the acceptance probability; if the new solution is accepted, the initial temperature is updated based on the given temperature decay function coefficient; if the new solution is not accepted, return to S2.2;

[0041] S2.5. Select the neighborhood structure based on the roulette wheel method, and select the neighborhood structure NS according to the selected neighborhood structure. k Generate a neighborhood solution θ' of the current solution θ;

[0042] S2.6. Using the selected neighborhood structure NS k Perform local search on the neighborhood solution θ' to obtain the local optimal solution θ", and compare the local optimal solution θ" with the current solution θ obtained by the simulated annealing algorithm. If θ" is better than θ, set θ = θ", and record the neighborhood structure NS k The number of successful searches for the local optimal solution; if θ” is not better than θ, the neighborhood structure NS is recorded. k The number of failures to search for the local optimal solution and update the neighborhood structure NS k probability;

[0043] S2.7, determine whether the current temperature is lower than the final temperature, if so, output the global optimal solution θ best =θ; if it is not lower than, take the current temperature as the initial temperature, the current solution θ as the initial solution, and return to S2.2.

[0044] Preferably, the S2.1 specifically includes:

[0045] Based on the simulated annealing and variable neighborhood search hybrid algorithm, the customer orders, production process, and transportation process related data in the supply chain production and transportation collaborative scheduling process are encoded, the production and transportation model is initialized, and any feasible solution that can meet the normal production and transportation of all orders is randomly selected as the initial solution; the initial solution is composed of two sequences, SC and PC-SX, where the SC code represents the order of order scheduling on a single machine, and the PC-SX code represents the transportation plan of autonomous transportation orders on multiple vehicles. The initial solution at the initial temperature is decoded to obtain θ best .

[0046] Preferably, in S2.1, the encoding process includes:

[0047] S2.1.1. Based on the characteristics of the two key stages of the supply chain production process and transportation process, the solution X consists of the solution X1 of the production stage and the solution X2 of the transportation stage. The solution of the production stage is represented by a one-dimensional integer array, and the solution of the transportation stage is represented by two one-dimensional integer arrays. The order production sequence SC contained in the solution X1 has a sequence length of The numbers in SC represent the order numbers, and the number of times the numbers appear represents the number of products required for the order. Solution X2 consists of two parts: the order transportation sequence PC and the vehicle number list SX corresponding to the order. The length of the sequence or list is PC and SX are composed of autonomous transport orders and vehicle numbers, respectively. Each column in the array represents that the order with the corresponding number is transported by the vehicle with the corresponding number;

[0048] The decoding process includes:

[0049] S2.1.2. Generate a set of processing orders assigned to the production line for processing from solution X1, that is, a certain feasible scheduling scheme π∈A, expressed as π={π(1),π(2),…,π(l)}, where π(i) means that order i is the π(i)th production in the production order set π, π(i)=j, where j=1,2,…,l; the production end time of order l is An example is as follows: Let π(4) = 2, that is, order 4 is the second one to be produced on the production line, that is, the production end time of order 4 is And the production cost of order i is

[0050]

[0051] S2.1.3. Generate a set of orders assigned to different vehicles from solution X2, that is, a feasible transportation plan μ∈B, expressed as μ={μ(1),μ(2),…,μ(l)}, where μ(i) is the transportation vehicle number corresponding to order i. When μ(i)=0, the transportation of the order is outsourced. Then we can get the order set I transported by vehicle k. k = {i|μ(i) = k, i∈I}, when μ(i) = 0, the order set I0 is the outsourced transportation order set; due to the capacity limitation of the vehicle itself, the order set I k It will be divided into multiple batches for roving distribution; the rules satisfied are: Order set I k The internal orders are sorted according to the known order production completion time. After sorting, the transportation batches λ(i) of the vehicles are divided according to the capacity constraints, and the transportation order set I corresponding to the b-th transportation of vehicle k is obtained. kb ={i|μ(i)=k,λ(i)=b,i∈I}, and the set I kb The total quantity of products required for internal orders shall not exceed the capacity limit of the vehicle itself;

[0052] S2.1.4. When the order is transported autonomously, order i∈I kb The transportation start time is ST i =ST kb =max{f i |i∈I kb}; The transportation end time of order i is CT i , when order i is in order set I kb When the first one is transported, that is, i = min{π(i)|i∈I kb}, the transportation end time Ct i =St i +T oi ; When order i is in order set I kb If it is not the first one to be transported, the transport end time CT i =CT h +T hi , order h in this transport batch is the previous transport location of order i, that is, π(i)-π(h)=min{π(i)-π(w),i,w∈I kb ,π(w)<π(i)}, and the condition π(h)<π(i) is satisfied; when the order is outsourced for transportation, the transportation start time of order i∈I0 is ST i =f i ; The transportation end time of order i is

[0053] S2.1.5. At this time, the transportation cost of order i is divided into the total cost of independent transportation or the total cost of outsourced transportation. When μ(i)≠0, the total transportation cost of order i is the sum of the transportation cost of independent transportation and the arrival delay cost, where the transportation cost of independent transportation SA i =T hi c1, arrival delay cost SL i =(CT i -d i )c3; When μ(i) = 0, the total outsourcing transportation cost of order i is the sum of the outsourcing transportation cost and the arrival delay cost, where the outsourcing transportation cost of order i is SO i =D 0i N i c2, arrival delay cost SL i =(CT i -d i )c3;

[0054] S2.1.6. For vehicles transported in a single batch, the vehicle departure time is ST kb =ST i =max{f i |i∈Ikb}; Vehicle return factory time CT kb = max{CT i | i ∈ I kb}+ T ho , where π(h) = max{π(i)| i ∈ I kb};

[0055] S2.1.7. Therefore, the total supply chain cost

[0056] Preferably, the S2.3 specifically includes:

[0057] S2.3.1. According to the initial solution θ0 and the new solution θ α , decode to obtain the total supply chain costs of the initial solution and the new solution, and record the values as f0 and f α ;

[0058] S2.3.2. Calculate the difference between f0 and f α , that is, f α - f0, and use it as the energy increment.

[0059] Preferably, the S2.4 specifically includes:

[0060] When the energy increment is not less than 0, calculate the acceptance probability of the new solution. The acceptance probability can be expressed as Take a random number r, 0 < r < 1. When r < P, update the new solution to the current solution; otherwise, do not update. The new solution is not accepted, and return to S2.2.

[0061] Preferably, the S2.5 specifically includes:

[0062] Perform operations on the current solution θ within the selected neighborhood structure NS k to enable the initial solution at this stage to jump to another solution in the feasible region, thereby avoiding repeated iteration under the same solution and preventing the algorithm from falling into a local optimum.

[0063] Preferably, the S2.6 specifically includes:

[0064] S2.6.1. The neighborhood structure NS of the variable neighborhood search algorithm k (k = 1, 2,..., k mak ). During the stage of setting the algorithm parameters, set the initial selection probability SP k of each neighborhood structure to be the same, where

[0065] S2.6.2. During the local search process of the variable neighborhood search algorithm, record the number of times S k that each neighborhood structure successfully searches to the local optimum and the number of failures Lk , update the selection probability of each neighborhood according to the following formula: Among them SUC k Represents the neighborhood structure NS k The local search success rate is calculated as Δ is a sufficiently small positive number;

[0066] S2.6.3. Define a variable R′, generate a random number in the interval (0,1), assign the random number to the variable R′, and determine the interval to which the variable R′ belongs. If SUC k-1 <R′<SUC k , then select the neighborhood structure NS k The type of neighborhood search.

[0067] Preferably, the supply chain scheduling optimization system includes: an input parameter acquisition module and a solution module;

[0068] The input parameter acquisition module is used to set the input parameters of the simulated annealing algorithm based on neighborhood search according to customer orders, production process, order data in the transportation process, production line data, transportation vehicle data, and transportation outsourcing cost data;

[0069] The solution module is used to obtain the global optimal solution based on the input parameters and the simulated annealing algorithm based on neighborhood search, and based on the global optimal solution, obtain the minimum total cost of the supply chain, the optimal single-machine scheduling plan for the production process, and the optimal order transportation plan.

[0070] The present invention provides a supply chain scheduling optimization method based on transportation outsourcing and time window constraints. Compared with the prior art, it has the following beneficial effects:

[0071] In the present invention, based on the framework of strict time window constraints and in the case where self-transportation capacity cannot fully meet the order transportation needs, the order production and transportation processes are coordinated and scheduled, the benefits of independent order transportation and outsourced transportation are weighed, and the production scheduling plan and transportation scheduling plan under the situation of outsourcing part of the transportation tasks are comprehensively considered. The approximate optimal solution to the problem is obtained dynamically and accurately, which can effectively coordinate and optimize and allocate resources in the two stages of supply chain production and transportation, thereby minimizing the overall cost of the supply chain and improving the resource collaboration efficiency and management capabilities of the enterprise supply chain. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] 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 some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0073] Figure 1 A schematic diagram of a flow chart of a supply chain scheduling optimization method based on transportation outsourcing and time window constraints provided by an embodiment of the present invention;

[0074] Figure 2 A flowchart of a simulated annealing algorithm based on neighborhood search provided by an embodiment of the present invention;

[0075] Figure 3 A decoding schematic diagram provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0076] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0077] The embodiment of the present application solves the problem of product transportation outsourcing and supply chain production and transportation collaborative scheduling under strict time window constraints by providing a supply chain scheduling optimization method based on transportation outsourcing and time window constraints.

[0078] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0079] Example:

[0080] like Figure 1 As shown, the present invention provides a supply chain scheduling optimization method based on transportation outsourcing and time window constraints, and the scheduling optimization method includes the following steps:

[0081] S1. According to the data related to customer orders, production process and transportation process, set the input parameters of the simulated annealing algorithm based on neighborhood search;

[0082] S2. Obtain a global optimal solution based on the input parameters and a simulated annealing algorithm based on neighborhood search, and obtain the minimum total cost of the supply chain, the optimal single-machine scheduling plan for the production process, and the optimal order transportation plan based on the global optimal solution.

[0083] like Figure 2 As shown, in S1, according to the customer order, production process, and transportation process related data, the input parameters of the simulated annealing algorithm based on neighborhood search are set to respectively construct the total cost model of the supply chain based on partial product transportation outsourcing and strict time window constraints. The specific implementation process is as follows:

[0084] Obtain data related to customer orders, production processes, and transportation processes through manual entry, and set the input parameters of the simulated annealing algorithm based on neighborhood search.

[0085] The input parameters of the simulated annealing algorithm based on neighborhood search include:

[0086] Order set I (i, h = 0, 1, 2, ..., l); customer number, order number, distribution point number i, h, where i, h = 0, 1, 2, ..., l, i = 0 is the production center; order production sequence number on the production line j = 1, 2, ..., l; transport vehicle number k = 0, 1, 2, ..., m, k = 0 is transported by outsourced vehicles; transport batch b = 1, 2, ..., n during vehicle patrol distribution;

[0087] The number of products produced per unit time on each production line q;

[0088] The production cost per unit time of each production line is c0;

[0089] The quantity of product N required for order i i ;

[0090] Order i arrival deadline d i ;

[0091] The transport capacity Q of a single autonomous transport vehicle;

[0092] The transportation speed v of autonomous and outsourced transportation vehicles;

[0093] The distance D between the factory and customer i 0i , correspondingly, the distance D between order i and order h ih ;

[0094] The cost coefficient c1 of the autonomous transport vehicle per unit of travel time;

[0095] Unit distance cost coefficient c2 for outsourced transportation per unit of product;

[0096] The penalty factor c3 for each unit of lateness;

[0097] Transportation time T between factory J0 and order i 0i , correspondingly, the shipping time between order i and order h is T ih ;

[0098] The complete set A of scheduling plans for a single machine in the production process, a certain feasible scheduling plan π∈A, denoted as π={π(1),π(2),…,π(l)}, where π(i) means that order i is produced at the π(i)th place in the production sequence set π, and π(i)=j;

[0099] Production start time s for order i i ;

[0100] The production end time f of order i i ;

[0101] The complete set of multi-vehicle scheduling solutions for the transportation process B, a certain feasible transportation solution μ∈B, is expressed as μ={μ(1),μ(2),…,μ(l)}, where μ(i) is the transportation vehicle number of the corresponding order i, when μ(i)=0, the transportation of the order is outsourced, and λ(i) is the transportation batch number of the corresponding order i in the transportation vehicle;

[0102] The shipping departure time ST for order i i ;

[0103] Arrival time CT of order i i ;

[0104] The transport order set I corresponding to the b-th transport of vehicle k kb ;

[0105] The departure time of vehicle k for the bth transport from the factory is ST kb ;

[0106] The time CT when vehicle k returns to the factory for the bth transport kb ;

[0107] Production cost SP for order i i ;

[0108] The road transportation cost SA of order i i ;

[0109] Arrival delay cost SL for order i i ;

[0110] Order i outsourced shipping costs SO i ;

[0111] Total supply chain cost ST;

[0112] In S2, the global optimal solution is obtained according to the input parameters and the simulated annealing algorithm based on neighborhood search, and the minimum total cost of the supply chain, the optimal production process single machine scheduling plan, and the optimal order transportation plan are obtained according to the global optimal solution. The specific implementation process is as follows:

[0113] S2.1, set the operating parameters of the simulated annealing algorithm and the initial solution θ0, the operating parameters include: setting the operating parameters, including the initial temperature t ini , termination temperature t final , temperature attenuation function coefficient e; specifically including:

[0114] Based on the simulated annealing and variable neighborhood search hybrid algorithm, the customer orders, production process, and transportation process related data in the supply chain production and transportation collaborative scheduling process are encoded, the production and transportation model is initialized, and any feasible solution that can meet the normal production and transportation of all orders is randomly selected as the initial solution. The initial solution is composed of two sequences, SC and PC-SX, where the SC code represents the order of order scheduling on a single machine, and the PC-SX code represents the transportation plan for autonomous transportation orders on multiple vehicles. The initial solution at the initial temperature is decoded to obtain θ best The specific implementation process is as follows:

[0115] like Figure 3 As shown, the encoding process is as follows:

[0116] S2.1.1. Based on the characteristics of the two key stages of the supply chain production process and transportation process, the solution X consists of the solution X1 of the production stage and the solution X2 of the transportation stage. The solution of the production stage is represented by a one-dimensional integer array, and the solution of the transportation stage is represented by two one-dimensional integer arrays. The order production sequence SC contained in the solution X1 has a sequence length of The numbers in SC represent the order numbers, and the number of times the numbers appear represents the number of products required for the order. Solution X2 consists of two parts: the order transportation sequence PC and the vehicle number list SX corresponding to the order. The length of the sequence or list is PC and SX are composed of autonomous transport orders and vehicle numbers, respectively. Each column in the array represents the order with the corresponding number being transported by the vehicle with the corresponding number. In particular, the order in SC and PC-SX is determined by the following rules:

[0117] The order arrangement order in PC will be arranged according to the order production order in SC; the autonomous transport vehicle and outsourcing number in SX, from front to back, represents the transportation order of the order in the vehicle with that number.

[0118] It should be understood that the vehicles that can be selected during the transportation phase are mainly to ensure the normal transportation, and the actual transportation demand for the completed order products is often greater than the vehicle capacity, which requires the vehicle to be put into the next round of delivery immediately after completing the current round of delivery. In the case of outsourced transportation, factors such as the batch and capacity restrictions of outsourced transportation vehicles are not considered. The end time of order production is the start time of the outsourced transportation of the order. In addition, in the case of delayed arrival, the total transportation cost of the order needs to be calculated, including the transportation cost of the independent transportation route or the outsourced transportation cost, as well as the arrival delay cost caused by delayed arrival.

[0119] The decoding process is to calculate the total cost of the entire supply chain based on the feasible solution obtained according to the actual scenario. It mainly includes the following operations:

[0120] S2.1.2. Generate a set of processing orders assigned to the production line for processing from solution X1, that is, a certain feasible scheduling scheme π∈A, expressed as π={π(1),π(2),…,π(l)}, where π(i) means that order i is the π(i)th production in the production order set π, π(i)=j, where j=1,2,…,l. The production end time of order l is An example is as follows: Let π(4) = 2, that is, order 4 is the second one to be produced on the production line, that is, the production end time of order 4 is And the production cost of order i is

[0121]

[0122] S2.1.3. Generate a set of orders assigned to different vehicles from solution X2, that is, a certain feasible transportation plan μ∈B, expressed as μ={μ(1),μ(2),…,μ(l)}, where μ(i) is the transportation vehicle number corresponding to order i, and when μ(i)=0, the transportation of the order is outsourced. The order set I transported by vehicle k can be obtained. k = {i|μ(i) = k, i∈I}, when μ(i) = 0, the order set I0 is the outsourced transportation order set. Due to the capacity limitation of the vehicle itself, the order set I k It will be divided into multiple batches for roving distribution. The rules satisfied are: Order set I k The internal orders are sorted according to the known order production completion time. After sorting, the transportation batches λ(i) of the vehicles are divided according to the capacity constraints, and the transportation order set I corresponding to the b-th transportation of vehicle k is obtained. kb ={i|μ(i)=k,λ(i)=b,i∈I}, and the set I kb The total quantity of products required for internal orders shall not exceed the capacity limit of the vehicle itself;

[0123] S2.1.4. When the order is transported autonomously, order i∈I kb The transportation start time is ST i =ST kb =max{f i |i∈I kb}. The shipping end time of order i is CT i , when order i is in order set I kb When the first one is transported, that is, i = min{π(i)|i∈I kb}, the transportation end time Ct i =St i +T oi ; When order i is in order set I kb If it is not the first one to be transported, the transport end time CTi =CT h +T hi , order h in this transport batch is the previous transport location of order i, that is, π(i)-π(h)=min{π(i)-π(w),i,w∈I kb ,π(w)<π(i)}, and the condition π(h)<π(i) is satisfied. When the order is outsourced for transportation, the transportation start time of order i∈I0 is ST i =f i The shipping end time for order i is

[0124] S2.1.5. At this time, the transportation cost of order i is divided into the total cost of independent transportation or the total cost of outsourced transportation. When μ(i)≠0, the total transportation cost of order i is the sum of the transportation cost of independent transportation and the arrival delay cost, where the transportation cost of independent transportation SA i =T hi c1, arrival delay cost SL i =(CT i -d i )c3; When μ(i) = 0, the total outsourcing transportation cost of order i is the sum of the outsourcing transportation cost and the arrival delay cost, where the outsourcing transportation cost of order i is SO i =D 0i N i c2, arrival delay cost SL i =(CT i -d i )c3;

[0125] S2.1.6. For vehicles transported in a single batch, the vehicle departure time is ST kb =ST i =max{f i |i∈I kb}; Vehicle return time CT kb =max{CT i |i∈I kb}+T ho , where π(h)=max{π(i)|i∈I kb};

[0126] S2.1.7, therefore the total supply chain cost

[0127] S2.2. Obtain and record the new solution θ through perturbation α The specific implementation process is as follows:

[0128] S2.2.1. The perturbation operation is to randomly modify the initial solution θ0 within a certain range to explore the solution space and randomly select a new candidate solution θ in the neighborhood of the initial solution θ0. α The random modification methods include:

[0129] 1. Exchange method: In the production sequence SC, two orders are randomly selected, such as order i and order h, and the corresponding production positions are exchanged; in the order vehicle numbering scheme PC-SX, two orders are randomly selected in PC, such as order i and order h, and the corresponding vehicle numbers in SX are exchanged;

[0130] 2. Reversal method: In the production sequence SC, randomly select two orders, such as order i and order h, and swap the production positions of the group of orders from order i to order h (including order i to order h), so that the order with the last production order is placed at the first production position, and the order with the first production order is placed at the end; in the order vehicle numbering scheme PC-SX, randomly select two orders in PC, such as order i and order h, and swap the vehicle numbers in the corresponding SX for the group of orders from order i to order h (including order i to order h);

[0131] 3. Insertion method: In the production sequence SC, randomly select two orders, such as order i and order h, change the production position of order i, and place the production position of order i next to order h; in the order vehicle numbering scheme PC-SX, randomly select two orders in PC, such as order i and order h, change the vehicle number in SX corresponding to order i to the vehicle number corresponding to order h, and change the vehicle numbers in SX corresponding to the orders produced after order i to the last order produced in PC to the vehicle numbers in SX corresponding to the previous production order.

[0132] It is worth noting that in order to achieve effective global optimization in the solution space, the perturbation can be large in the high temperature stage to promote extensive exploration of the solution space; as the temperature decreases, the perturbation should gradually decrease to enhance the convergence of the algorithm.

[0133] S2.3, calculating the energy increment based on the initial solution and the new solution;

[0134] S2.3.1. Based on the initial solution θ0 and the new solution θ α , decode to obtain the total supply chain cost of the initial solution and the new solution, and record the values ​​as f0 and f α ,The decoding process is shown in S2.1;

[0135] S2.3.2. Calculate f0 and f α The difference, i.e. f α -f0, and as energy increment;

[0136] S2.4. If the energy increment is less than 0, i.e., f α - f0 < 0, then the new solution can be used as the current solution, and the current solution is denoted as θ; if the energy increment is not less than 0, then calculate the acceptance probability and accept the new solution according to the acceptance probability. If the new solution is accepted, update the initial temperature based on the given temperature decay function coefficient; if the new solution is not accepted, return to S2.2. The specific implementation process is as follows:

[0137] When the energy increment is not less than 0, calculate the new solution acceptance probability, and the acceptance probability can be expressed as Take a random number r, 0 < r < 1. When r < P, update the new solution to the current solution; otherwise, do not update, the new solution is not accepted, and return to S2.2.

[0138] S2.5. Select the neighborhood structure based on the roulette wheel method, and generate the neighborhood solution θ' of the current solution θ according to the selected neighborhood structure NS k The specific implementation process is as follows:

[0139] Perform operations on the current solution θ within the selected neighborhood structure NS k so that the initial solution at this stage can jump to another solution in the feasible region, thus avoiding repeated iteration under the same solution and causing the algorithm to fall into local optimum. Considering the complexity of the supply chain production and transportation collaborative scheduling and the requirements of the constraint relationship, in order to ensure that each solution generated by the operation change is a feasible solution, the following 4 neighborhood structures are specifically designed:

[0140] Neighborhood structure 1: Define variables a and b. Arbitrarily select two orders i and h from SC, assign the order positions to variables a and b, compare the corresponding order arrival deadlines d i and d h , if d i < d h and a > b, place the code at position a in the sequence SC before the code at position b; define variables c and d, randomly select two orders in PC, such as order i and order h, assign the order positions to variables c and d, and randomly select two numbers from the interval [0, m] to assign to the code at position c in the sequence SX and the code placed at position d;

[0141] Neighborhood structure 2: Define variables a and b. Arbitrarily select two orders i and h from SC, assign the order positions to variables a and b, compare the corresponding order arrival deadlines d i and d h , if d i < d hAnd a > b, swap the encodings at positions a and b in sequence SC; define variables c and d, randomly select two orders in PC, such as order i and order h, assign the order positions to variables c and d, and swap the encodings at positions c and d in sequence SX;

[0142] Neighborhood structure 3: Define variables a and b, arbitrarily select two orders i and h from SC, assign the order positions to variables a and b, and compare the required product quantities N i and N h , if N i < N h and a < b, swap the encodings at positions a and b in sequence SC; define variables c and d, randomly select two orders in PC, such as order i and order h, assign the order positions to variables c and d, and place the encoding at position c in sequence SX before the encoding at position d;

[0143] Neighborhood structure 4: Define variables a and b, arbitrarily select two orders i and h from SC, assign the order positions to variables a and b, and compare the distance D between the corresponding factory and customer i 0i and D 0h , if D 0i < D 0h and a < b, swap the encodings at positions a and b in sequence SC; define variables c and d, randomly select two orders in PC, such as order i and order h, assign the order positions to variables c and d, and swap the encodings at positions c and d in sequence SX;

[0144] S2.6. Use the selected neighborhood structure NS k Perform local search on the neighborhood solution θ' to obtain the local optimal solution θ", and compare the local optimal solution θ" with the current solution θ obtained by the simulated annealing algorithm. If θ" is better than θ, then let θ = θ", and record the neighborhood structure NS k The successful number of times of searching for the local optimal solution; if θ" is not better than θ, then record the neighborhood structure NS k The failure number of times of searching for the local optimal solution, and update the probability of selecting this neighborhood structure NS k . The specific implementation process is as follows:

[0145] S2.6.1. The neighborhood structure NS of the variable neighborhood search algorithm k (k = 1, 2,..., k max ), in the stage of setting algorithm parameters, set the initial selection probability SP of each neighborhood structure k to be the same, where

[0146] S2.6.2. During the local search process of the variable neighborhood search algorithm, record the number of times each neighborhood structure successfully searches for the local optimal solution S k and the number of failures L k , update the selection probability of each neighborhood according to the following formula: Among them SUC k Represents the neighborhood structure NS k The local search success rate is calculated as Δ is a sufficiently small positive number;

[0147] S2.6.3. Define a variable R′, generate a random number in the interval (0,1), assign the random number to the variable R′, and determine the interval to which the variable R′ belongs. If SUC k-1 <R′<SUC k , then select the neighborhood structure NS k The type of neighborhood search.

[0148] S2.7, determine whether the current temperature is lower than the final temperature, if so, output the global optimal solution θ best =θ; if it is not lower than, take the current temperature as the initial temperature, the current solution θ as the initial solution, and return to S2.2. Specifically include:

[0149] According to the global optimal solution θ output by the algorithm best , the SC sequence and PC-SX sequence in the solution are analyzed according to the actual scenario of the enterprise supply chain, the order processing order and order transportation plan are obtained, and the minimum total cost of the supply chain is obtained.

[0150] The embodiment of the present invention further provides a supply chain scheduling optimization system based on transportation outsourcing and time window constraints, the supply chain scheduling optimization system comprising: an input parameter acquisition module and a solution module;

[0151] The input parameter acquisition module is used to set the input parameters of the simulated annealing algorithm based on neighborhood search according to customer orders, production process, order data in the transportation process, production line data, transportation vehicle data, and transportation outsourcing cost data;

[0152] The solution module is used to obtain the global optimal solution based on the input parameters and the simulated annealing algorithm based on neighborhood search, and based on the global optimal solution, obtain the minimum total cost of the supply chain, the optimal single-machine scheduling plan for the production process, and the optimal order transportation plan.

[0153] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0154] 1. In the embodiment of the present invention, based on the framework of strict time window constraints and when self-transportation capacity cannot fully meet the order transportation needs, the order production and transportation processes are coordinated and scheduled, the benefits of independent order transportation and outsourced transportation are weighed, and the production scheduling plan and transportation scheduling plan under the situation of outsourcing part of the transportation tasks are comprehensively considered. The approximate optimal solution of the problem is obtained dynamically and accurately, which can effectively coordinate and optimize and allocate resources in the two stages of supply chain production and transportation, thereby minimizing the overall cost of the supply chain and improving the resource collaboration efficiency and management capabilities of the enterprise supply chain.

[0155] 2. In the embodiment of the present invention, a simulated annealing algorithm based on neighborhood search is designed, which combines the global optimization capability of the simulated annealing algorithm and the local search capability of the variable neighborhood search algorithm; at the same time, a corresponding adaptive selection mechanism is designed in the simulated annealing algorithm and the variable neighborhood search algorithm, and a special crossover operator and a local search neighborhood structure are designed, which effectively improves the algorithm's solution efficiency and the quality of the solution.

[0156] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0157] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A supply chain scheduling optimization method based on transportation outsourcing and time window constraints, characterized in that: The scheduling optimization method comprises the following steps: S1. According to the data related to customer orders, production process and transportation process, set the input parameters of the simulated annealing algorithm based on neighborhood search; S2. Obtain a global optimal solution based on the input parameters and a simulated annealing algorithm based on neighborhood search, and obtain the minimum total cost of the supply chain, the optimal single-machine scheduling plan for the production process, and the optimal order transportation plan based on the global optimal solution.

2. The supply chain scheduling optimization method based on transportation outsourcing and time window constraints as claimed in claim 1 is characterized in that: In S1, the input parameters of the simulated annealing algorithm based on neighborhood search specifically include: Order set I (i, h = 0, 1, 2, ..., l); customer number, order number, distribution point number i, h, where i, h = 0, 1, 2, ..., l, i = 0 is the production center; order production sequence number on the production line j = 1, 2, ..., l; transport vehicle number k = 0, 1, 2, ..., m, k = 0 is transported by outsourced vehicles; transport batch b = 1, 2, ..., n during vehicle patrol distribution; The number of products produced per unit time on each production line q; The production cost per unit time of each production line is c0; The quantity of product N required for order i i ; Order i arrival deadline d i ; The transport capacity Q of a single autonomous transport vehicle; The transportation speed v of autonomous and outsourced transportation vehicles; The distance D between the factory and customer i 0i , correspondingly, the distance D between order i and order h ih ; The cost coefficient c1 of the autonomous transport vehicle per unit of travel time; Unit distance cost coefficient c2 for outsourced transportation per unit of product; The penalty factor c3 for each unit of late time; Transportation time T between factory J0 and order i 0i , correspondingly, the shipping time between order i and order h is T ih ; The complete set A of scheduling plans for a single machine in the production process, a certain feasible scheduling plan π∈A, denoted as π={π(1),π(2),…,π(l)}, where π(i) means that order i is produced at the π(i)th place in the production sequence set π, and π(i)=j; Production start time s for order i i ; The production end time f of order i i ; The complete set of multi-vehicle scheduling solutions for the transportation process B, a certain feasible transportation solution μ∈B, is expressed as μ={μ(1),μ(2),…,μ(l)}, where μ(i) is the transportation vehicle number of the corresponding order i, when μ(i)=0, the transportation of the order is outsourced, and λ(i) is the transportation batch number of the corresponding order i in the transportation vehicle; The shipping departure time of order i is ST i ; Arrival time CT of order i i ; The transport order set I corresponding to the b-th transport of vehicle k kb ; The departure time of vehicle k for the bth transport from the factory is ST kb ; The time CT when vehicle k returns to the factory for the bth transport kb ; Production cost SP for order i i ; The road transportation cost SA of order i i ; Arrival delay cost SL for order i i ; Order i outsourced shipping costs SO i ; Total supply chain cost ST.

3. The supply chain scheduling optimization method based on transportation outsourcing and time window constraints as claimed in claim 1 is characterized in that: The S2 specifically includes: S2.1, set the operating parameters of the simulated annealing algorithm and the initial solution θ0, the operating parameters include: initial temperature t ini , termination temperature t final , temperature attenuation function coefficient e; S2.

2. Obtain and record the new solution θ through perturbation α ; S2.3, calculating the energy increment based on the initial solution and the new solution; S2.4, if the energy increment is less than 0, the new solution can be used as the current solution, and the current solution is recorded as θ; if the energy increment is not less than 0, the acceptance probability is calculated, and the new solution is received according to the acceptance probability; if the new solution is accepted, the initial temperature is updated based on the given temperature decay function coefficient; if the new solution is not accepted, return to S2.2; S2.

5. Select the neighborhood structure based on the roulette method, and select the neighborhood structure NS according to the selected neighborhood structure. k Generate a neighborhood solution θ' of the current solution θ; S2.

6. Using the selected neighborhood structure NS k Perform local search on the neighborhood solution θ' to obtain the local optimal solution θ", and compare the local optimal solution θ" with the current solution θ obtained by the simulated annealing algorithm. If θ" is better than θ, set θ = θ", and record the neighborhood structure NS k The number of successful searches for the local optimal solution; if θ” is not better than θ, the neighborhood structure NS is recorded. k The number of failures to search for the local optimal solution and update the neighborhood structure NS k probability; S2.7, determine whether the current temperature is lower than the final temperature, if so, output the global optimal solution θ best =θ; if it is not lower than, take the current temperature as the initial temperature, the current solution θ as the initial solution, and return to S2.

2.

4. The supply chain scheduling optimization method based on transportation outsourcing and time window constraints as claimed in claim 3 is characterized in that: The S2.1 specifically includes: Based on the simulated annealing and variable neighborhood search hybrid algorithm, the customer orders, production process, and transportation process related data in the supply chain production and transportation collaborative scheduling process are encoded, the production and transportation model is initialized, and any feasible solution that can meet the normal production and transportation of all orders is randomly selected as the initial solution; the initial solution is composed of two sequences, SC and PC-SX, where the SC code represents the order of order scheduling on a single machine, and the PC-SX code represents the transportation plan of autonomous transportation orders on multiple vehicles. The initial solution at the initial temperature is decoded to obtain θ best .

5. The supply chain scheduling optimization method based on transportation outsourcing and time window constraints as claimed in claim 4 is characterized in that: In S2.1, the encoding process includes: S2.1.

1. Based on the characteristics of the two key stages of the supply chain production process and transportation process, the solution X consists of the solution X1 of the production stage and the solution X2 of the transportation stage. The solution of the production stage is represented by a one-dimensional integer array, and the solution of the transportation stage is represented by two one-dimensional integer arrays. The order production sequence SC contained in the solution X1 has a sequence length of ∑ l i＝1 N i ; The numbers in SC represent the order numbers, and the number of times the numbers appear represents the number of products required for the order; Solution X2 consists of two parts, the order transportation sequence PC and the vehicle number list SX corresponding to the order. The length of the sequence or list is PC and SX are composed of autonomous transport orders and vehicle numbers, respectively. Each column in the array represents that the order with the corresponding number is transported by the vehicle with the corresponding number; The decoding process includes: S2.1.

2. Generate a set of processing orders assigned to the production line for processing from solution X1, that is, a certain feasible scheduling scheme π∈A, expressed as π={π(1),π(2),…,π(l)}, where π(i) means that order i is the π(i)th production in the production order set π, π(i)=j, where j=1,2,…,l; the production end time of order l is An example is as follows: Let π(4) = 2, that is, order 4 is the second one to be produced on the production line, that is, the production end time of order 4 is And the production cost of order i is S2.1.

3. Generate a set of orders assigned to different vehicles from solution X2, that is, a feasible transportation plan μ∈B, expressed as μ={μ(1),μ(2),…,μ(l)}, where μ(i) is the transportation vehicle number corresponding to order i. When μ(i)=0, the transportation of the order is outsourced. Then we can get the order set I transported by vehicle k. k = {i|μ(i) = k, i∈I}, when μ(i) = 0, the order set I0 is the outsourced transportation order set; due to the capacity limitation of the vehicle itself, the order set I k It will be divided into multiple batches for roving distribution; the rules satisfied are: Order set I k The internal orders are sorted according to the known order production completion time. After sorting, the transportation batches λ(i) of the vehicles are divided according to the capacity constraints, and the transportation order set I corresponding to the b-th transportation of vehicle k is obtained. kb ={i|μ(i)=k,λ(i)=b,i∈I}, and the set I kb The total quantity of products required for internal orders shall not exceed the capacity limit of the vehicle itself; S2.1.

4. When the order is transported autonomously, order i∈I kb The transportation start time is ST i =ST kb =max{f i |i∈I kb }; The transportation end time of order i is CT i , when order i is in order set I kb When the first one is transported, that is, i = min{π(i)|i∈I kb }, the transportation end time Ct i =St i +T oi ; When order i is in order set I kb If it is not the first one to be transported, the transport end time CT i =CT h +T hi , order h in this transport batch is the previous transport location of order i, that is, π(i)-π(h)=min{π(i)-π(w),i,w∈I kb ,π(w)<π(i)}, and the condition π(h)<π(i) is satisfied; when the order is outsourced for transportation, the transportation start time of order i∈I0 is ST i =f i ; The transportation end time of order i is S2.1.

5. At this time, the transportation cost of order i is divided into the total cost of independent transportation or the total cost of outsourced transportation. When μ(i)≠0, the total transportation cost of order i is the sum of the transportation cost of independent transportation and the arrival delay cost, where the transportation cost of independent transportation SA i =T hi c1, arrival delay cost SL i =(CT i -d i )c3; When μ(i) = 0, the total outsourcing transportation cost of order i is the sum of the outsourcing transportation cost and the arrival delay cost, where the outsourcing transportation cost of order i is SO i =D 0i N i c2, arrival delay cost SL i =(CT i -d i )c3; S2.1.

6. For vehicles transported in a single batch, the vehicle departure time is ST kb =ST i =max{f i |i∈I kb }; Vehicle return time CT kb =max{CT i |i∈I kb }+T ho , where π(h)=max{π(i)|i∈I kb }; S2.1.7, therefore the total supply chain cost 6. The supply chain scheduling optimization method based on transportation outsourcing and time window constraints as claimed in claim 3 is characterized in that: The S2.3 specifically includes: S2.3.

1. Based on the initial solution θ0 and the new solution θ α , decode to obtain the total supply chain cost of the initial solution and the new solution, and record the values ​​as f0 and f α ; S2.3.

2. Calculate f0 and f α The difference, i.e. f α -f0, and as energy increment.

7. The supply chain scheduling optimization method based on transportation outsourcing and time window constraints as claimed in claim 6 is characterized in that: The S2.4 specifically includes: When the energy increment is not less than 0, calculate the acceptance probability of the new solution, and the acceptance probability can be expressed as Generate a random number r, where 0 < r < 1. When r < P, update the new solution to the current solution; otherwise, do not update, the new solution is not accepted, and return to S2.

2.

8. The supply chain scheduling optimization method based on transportation outsourcing and time window constraints as claimed in claim 3 is characterized in that: The S2.5 specifically includes: For the current solution θ in the selected neighborhood structure NS k The operation is performed within the solution so that the initial solution of this stage can jump to another solution in the feasible domain, thereby avoiding repeated iterations under the same solution and causing the algorithm to fall into the local optimum.

9. The supply chain scheduling optimization method based on transportation outsourcing and time window constraints as claimed in claim 3 is characterized in that: The S2.6 specifically includes: S2.6.

1. Neighborhood structure NS of variable neighborhood search algorithm k (k=1,2,...,k max ), in the algorithm parameter setting stage, set the initial selection probability SP of each neighborhood structure k The same, among which S2.6.

2. During the local search process of the variable neighborhood search algorithm, record the number of times each neighborhood structure successfully searches for the local optimal solution S k and the number of failures L k , update the selection probability of each neighborhood according to the following formula: Among them SUC k Represents the neighborhood structure NS k The local search success rate is calculated as Δ is a sufficiently small positive number; S2.6.

3. Define a variable R′, generate a random number in the interval (0,1), assign the random number to the variable R′, and determine the interval to which the variable R′ belongs. If SUC k-1 <R′<SUC k , then select the neighborhood structure NS k The type of neighborhood search.

10. A supply chain scheduling optimization system based on transportation outsourcing and time window constraints, characterized in that: The supply chain scheduling optimization system includes: an input parameter acquisition module and a solution module; The input parameter acquisition module is used to set the input parameters of the simulated annealing algorithm based on neighborhood search according to customer orders, production process, order data in the transportation process, production line data, transportation vehicle data, and transportation outsourcing cost data; The solution module is used to obtain the global optimal solution based on the input parameters and the simulated annealing algorithm based on neighborhood search, and based on the global optimal solution, obtain the minimum total cost of the supply chain, the optimal single-machine scheduling plan for the production process, and the optimal order transportation plan.

Images