Method and system for scheduling and route optimization of integrated home furnishing logistics with time windows

Abstract

The present invention relates to a method and system for scheduling and optimizing home furnishing logistics with a time window. The method comprises the following steps: S1: establishing an objective function and constraints based on a home furnishing logistics distribution model for the integrated home furnishing system: customer constraints, worker constraints, vehicle constraints, and skill and time window constraints; S2: solving the model using an improved ant colony algorithm, taking M optimal path plans as alternative path plans; substituting the customer constraints, worker constraints, vehicle constraints, and skill and time window constraints into the alternative path plans, performing logistics scheduling, taking the path plan that satisfies all constraints as the path plan, and obtaining the corresponding logistics schedule; if none of the M alternative path plans satisfies the constraints, adding workers or increasing working hours until a feasible path plan and logistics schedule are obtained. The method provided by the present invention uses an ant colony algorithm for minimum cost path planning and achieves optimal intelligent scheduling by establishing constraints.

Description

Technical Field

[0001] The present invention relates to the field of smart logistics technology, and specifically to a method and system for scheduling and route optimization of integrated home furnishing logistics with a time window. Background Art

[0002] With the development of smart logistics, smart logistics management has become a key competitive advantage for enterprises. To further enhance intelligent service levels, achieve refined operations, and reduce operating costs, integrated furniture service platforms must implement "last mile" smart logistics distribution. However, due to the unique characteristics of the furniture industry, it places higher demands on logistics distribution than existing logistics distribution in other industries. Intelligent distribution in the furniture industry requires consideration of numerous factors, including customer needs, product attributes, vehicle attributes, professional and technical attributes, personnel composition, vehicle scheduling, route planning, service time windows, and technician lunch break windows. However, current research and inventions primarily consider single and partial factors, and therefore fail to meet the intelligent management needs of integrated furniture service platforms. Summary of the Invention

[0003] In order to solve the above technical problems, the present invention provides a method and system for scheduling and route optimization of integrated home furnishing logistics with a time window.

[0004] The technical solution of the present invention is: a method for scheduling and optimizing the logistics routes of integrated home furnishings with a time window, comprising:

[0005] Step S1: Establish the objective function and constraints of the home furnishing logistics distribution model based on integrated assembly: customer constraints, worker constraints, vehicle constraints, and skill and time window constraints;

[0006] Step S2: Use an improved ant colony algorithm to solve the model, and use M optimal path plans as alternative path plans; substitute the customer constraints, worker constraints, vehicle constraints, and skill and time window constraints into the alternative path plans, perform logistics scheduling, and use the path plan that meets all the constraints as the path plan, and obtain the corresponding logistics schedule; if none of the M alternative path plans can meet the constraints, increase workers or increase working hours until a feasible path plan and logistics schedule are obtained.

[0007] Compared with the prior art, the present invention has the following advantages:

[0008] The present invention discloses a method for scheduling and optimizing routes for integrated home furnishing logistics with a time window. Aiming at the problem of intelligent scheduling and route planning for integrated home furnishing logistics distribution on a comprehensive furniture service platform, an ant colony algorithm is used for minimum cost route planning. Furthermore, by establishing constraints such as customer constraints, worker constraints, vehicle constraints, and skill and time window constraints, optimal intelligent scheduling is achieved, which can simultaneously meet the customer service time window requirements and the workers' lunch break requirements. The model of the present invention not only meets the requirements for scheduling and route planning for home furnishing logistics, but also further expands the multi-factor scheduling model. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 This is a flow chart of a method for scheduling and optimizing routes for integrated home furnishing logistics with time windows in an embodiment of the present invention;

[0010] Figure 2 Schematic diagram of the optimal delivery path solution in an embodiment of the present invention;

[0011] Figure 3 A diagram comparing resource usage between algorithmic and manual order scheduling in an embodiment of the present invention;

[0012] Figure 4 A comparison chart of the working hours of workers using algorithmic and manual order scheduling in an embodiment of the present invention;

[0013] Figure 5 This is a stability comparison chart of the algorithmic order scheduling and manual order scheduling in an embodiment of the present invention;

[0014] Figure 6 This is a comparison chart of the total costs of manual order scheduling and algorithmic order scheduling in an embodiment of the present invention;

[0015] Figure 7 This is a comparison chart of the total driving routes of manual scheduling and algorithm scheduling in an embodiment of the present invention. DETAILED DESCRIPTION

[0016] The present invention provides a method for scheduling and optimizing routes for integrated home furnishing logistics with a time window. The method adopts an ant colony algorithm for minimum cost route planning and realizes optimal intelligent scheduling by establishing constraints including customer constraints, worker constraints, vehicle constraints, and skill and time window constraints.

[0017] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below through specific implementation and in conjunction with the accompanying drawings.

[0018] The embodiments of the present invention are based on the following scenarios:

[0019] One distribution center, one assembly team, and several customer locations;

[0020] Assembly workers: different functional roles, including: pure driver, driver and installer, installer only;

[0021] Private car: refers to the assembler's own car, which can only be used by the assembler and his / her group. After the service is completed, the assembler usually returns to his / her home (i.e. the virtual destination);

[0022] Public bus: usually a fleet vehicle designated by the distribution center for reservation, which can be used by all assembly workers and returns to the distribution center after the service is completed;

[0023] Service unit: It consists of a vehicle and several assembly workers. Someone in the service unit must have the right to drive the vehicle and certain installation skills. Each service unit starts from the distribution center. After the service is completed, if it is a private car, it will return home; if it is a public car, it will return to the distribution center. The service unit can be dispatched a second time.

[0024] The rule is: Each customer request can only be served once by the same service unit. Service units have a limited working time, and each worker has fixed installation skills, with varying skill levels. The goal is to minimize costs while satisfying customer constraints, worker constraints, vehicle constraints, and skill and time window constraints.

[0025] The model assumptions are as follows:

[0026] (1) The vehicle has a passenger capacity limit of three people at most and is allowed to leave twice, regardless of special factors such as regional restrictions;

[0027] (2) Giving workers skill attributes, which are composed of driving skills and installation skills. In terms of installation skills, workers have different proficiency levels, in addition to the difference between those who know how to install and those who don’t.

[0028] (3) Batch delivery is not allowed. Each customer can only be served by one service unit and only once;

[0029] (4) Adopting a hard time window strategy, that is, the service unit must provide services within the customer's service time window and is not allowed to provide services after the timeout;

[0030] (5) Define the time when the service unit departs from the distribution center after loading as the starting working time, stipulate that the starting working time of all service units is the same, and there is a limit on the working time;

[0031] (6) The time from arrival at the customer's location to installation completion is defined as the customer's service time, which includes waiting, unloading, handling, and pure installation time, ignoring unloading time. Handling time is mainly related to environmental factors such as whether the community allows cars to enter and whether there is an elevator to the upper floor, as well as the weight, volume, and quantity of the goods. Handling time is simplified to be related only to quantity; pure installation time is related to the type and quantity of goods and the service capabilities of the installation team.

[0032] Example 1

[0033] like Figure 1 As shown, an embodiment of the present invention provides a method for scheduling and optimizing routes for integrated home furnishing logistics with a time window, comprising the following steps:

[0034] Step S1: Establish the objective function and constraints of the home furnishing logistics distribution model based on integrated assembly: customer constraints, worker constraints, vehicle constraints, and skill and time window constraints;

[0035] Step S2: Use the improved ant colony algorithm to solve the model, and take the M optimal path plans as alternative path plans; substitute customer constraints, worker constraints, vehicle constraints, and skill and time window constraints into the alternative path plans to perform logistics scheduling, and take the path plan that meets all the constraints as the path plan, and obtain the corresponding logistics schedule; if none of the M alternative path plans can meet the constraints, increase the number of workers or increase the working hours until a feasible path plan and logistics schedule are obtained.

[0036] In one embodiment, the above step S1: establishing the objective function and constraints of the home furnishing logistics distribution model based on integrated assembly: customer constraints, worker constraints, vehicle constraints, and skill and time window constraints, specifically includes:

[0037] Step S11: Establish the objective function of the path optimization based on the integrated home furnishing logistics distribution model as shown in formula (1):

[0038]

[0039] Where C is the set of all customer points i, C = {1, 2, ... n};

[0040] N is the set of all customer points C, distribution centers O, and virtual endpoints O', N = {0, 1, 2..., n + 1}, i = 0 represents the distribution center, i = n + 1 represents the virtual endpoint;

[0041] d ij is the distance from customer point i to customer point j; when i = 0, it represents the distance from the distribution center to customer point j; when j = n + 1, it represents the distance from customer point j to the virtual end point;

[0042] K is the set of vehicle k K = {1, 2, ... K};

[0043] x ijk Indicates that vehicle k travels from customer point i to customer point j, satisfying

[0044]

[0045] a is the fuel consumption cost per kilometer; b1 is the fixed cost per vehicle per day, including daily depreciation and insurance premiums;

[0046] Step S12: Establish the following constraints:

[0047] (1) Customer constraints

[0048] a) For each customer point, only one car can enter the point, as shown in formula (2):

[0049]

[0050] b) For each customer point, only one car can leave the point, as shown in formula (3):

[0051]

[0052] (2) Worker restraint

[0053] In the embodiment of the present invention, each worker is numbered and the order is arranged in the order of the worker numbers. The workers p1p2p3 on each vehicle are fixed.

[0054] a) A worker cannot appear on the same car repeatedly, as shown in formula (4):

[0055]

[0056] in, P is the set of workers p, P = {0, p1, p2, ... P}, where 0 represents a virtual person without any skills; p2 is the position number of worker 2 in the set P, and p2 is located after p1 and before p3;

[0057] b) A worker cannot appear on multiple vehicles, as shown in formula (5):

[0058]

[0059] c) Workers have fixed collocations, as shown in formula (6):

[0060] When p1, p2, and p3 contain only a portion of the members of a fixed collocation,

[0061] (3) Vehicle restraints

[0062] a) After completing the last customer's assembly task, the vehicle returns to the virtual address and does not depart from that location, as shown in formula (7):

[0063]

[0064] b) Path balance constraint, which ensures that the vehicle must leave the customer point after entering the customer point, as shown in formula (8)

[0065] As shown:

[0066]

[0067] in, Indicates that the vehicle returns to the distribution center;

[0068] Indicates that the vehicle does not return to the distribution center, but returns to the virtual address;

[0069] c) The clearance limit for each vehicle is as shown in formula (9):

[0070]

[0071] Where V is the clearance of each vehicle k; v j is the order volume of customer j;

[0072] d) There cannot be multiple combinations of people in a vehicle, as shown in formula (10):

[0073]

[0074] (4) Skill and time window constraints

[0075] a) Time balance constraint, driving skill and installation skill constraint, as shown in formula (11):

[0076]

[0077] Where M is the set of home furnishing category m; t i is the required handling time for customer i; T is the basic installation time for each category in M; The installation time for the worker combination p1p2p3 to install the mth category. If no worker in the combination will install the category, set a large integer E so that the constraint condition of formula (11) is not satisfied; q im The number of product categories that customer i needs to install; at ik is the time it takes for vehicle k to arrive at customer point i; u is the average speed of each vehicle;

[0078] b) Waiting time constraint, as shown in formula (12):

[0079]

[0080] Among them, wt ik is the waiting time of vehicle k at customer point i; e i It is the earliest delivery time booked by the customer;

[0081] c) Customer point time window constraint, as shown in formula (13):

[0082]

[0083] Among them, l i is the latest delivery time booked by the customer; i , l i ] is the time window for the customer to click i;

[0084] d) Specify the arrival time and number of installed products at the distribution center, as shown in formula (14):

[0085]

[0086] The embodiment of the present invention first adopts the ant colony path scheduling algorithm to solve formula (1), schedules the paths based on the expectation of vehicle capabilities, sorts them by cost, and obtains multiple paths with the lowest cost as alternative paths; then, according to the constraints, workers are filled into the alternative paths so that the alternative paths meet all the constraints and are feasible.

[0087] In one embodiment, step S2 above: using an improved ant colony algorithm to solve the model, taking M optimal path plans as alternative path plans; substituting customer constraints, worker constraints, vehicle constraints, and skill and time window constraints into the alternative path plans, performing logistics scheduling, taking the path plan that satisfies all constraints as the path plan, and obtaining the corresponding logistics schedule; if none of the M alternative path plans satisfies the constraints, adding workers or increasing working hours until a feasible path plan and logistics schedule are obtained, specifically including:

[0088] Step S21: Use the ant colony algorithm to solve formula (1). Preset m ants, start from the distribution center O, traverse all customer point sets C, and obtain path planning and its corresponding pheromone; the path with the minimum total cost among all ants is the optimal path of this iteration, and after iterating and incrementing the pheromone, the next iteration is performed; until the predetermined number of iterations is reached, the top M optimal path plans are output as alternative path plans;

[0089] The parameters of the ant colony algorithm in the embodiment of the present invention are as follows:

[0090] m: the number of ants performing optimization;

[0091] n: number of iterations;

[0092] TC: total cost of an ant;

[0093] BTC: optimal total cost in one iteration;

[0094] Stagnation: The number of stagnations, which is the number of iterations during which no solution exceeding the optimal solution has been found.

[0095] Ti ij : The amount of pheromone on the path. When initialized, the amount of pheromone on each path is a fixed value;

[0096] C ij The heuristic function, or visibility, is the expected degree of an ant's ability to move from client node i to client node j. It is generally related to the distance between the two client nodes. For this problem, it is related to the fuel cost, transfer angle, and waiting time from i to j.

[0097] P ij : The transfer rule of ants from node i to node j is the core step of the ant colony algorithm. Ants traverse the nodes in the network one by one according to this transfer rule and finally find the optimal solution to the problem.

[0098] D ij : represents the distance from customer point i to customer point j, which is usually a known distance matrix that can be obtained based on the coordinates of each customer point;

[0099] T ij : represents the time from customer point i to customer point j;

[0100] A ij : represents the angle between the route from the current customer point i to the next customer point j and the route from the previous customer point to i;

[0101] Q i : represents the waiting time penalty cost of reaching node i at this moment;

[0102] J i : represents the transfer angle penalty cost to reach node i at this moment;

[0103] α: Pheromone heuristic factor, which means the sensitivity of the ant to the amount of pheromone on the path or the degree to which the ant is affected by the amount of information on the path when choosing the next node;

[0104] β: expected heuristic factor, which represents the ant's ability to judge the distance to the next node or its ability to perceive the surrounding environment during the optimization process. Here, it is related to the movement cost;

[0105] γ: time window penalty cost heuristic factor, indicating the sensitivity of the ant to the time window penalty cost when selecting the next node;

[0106] δ: The heuristic factor of the transfer angle penalty cost, which indicates the sensitivity of the ant to the transfer angle penalty cost when selecting the next node;

[0107] ρ: represents the volatility coefficient of pheromone; RHO=1-ρ represents the residual coefficient of pheromone; where ρ∈(0,1);

[0108] TB: Taboo table, put the points that have been visited into the taboo table, and the ant will not go to this point next time;

[0109] AL: Permission form, complementary to TB;

[0110] c=D ij ×V (15)

[0111]

[0112]

[0113]

[0114]

[0115] Ti ij (t+1)=(1-ρ)Ti ij (t)+ΔTi ij (t,t+1) (20)

[0116]

[0117] Formula (15) calculates the cost of the next move using distance and variable cost.

[0118] Formula (16) normalizes the mobility cost. The selection coefficient of the optimal mobility cost is 1. The smaller the optimal mobility cost is, the lower the selection coefficient is, and the lowest is close to 0.

[0119] Formula (17) normalizes the penalty cost. The selection coefficient of the minimum penalty cost is 1. The higher the minimum penalty cost is, the lower the selection coefficient is, and the lowest is close to 0.

[0120] Formula (18) calculates the penalty cost of the rotation angle. If the rotation angle is too large (over 120 degrees) at the beginning of the calculation, it may cause a waste of round trips on the path. By limiting the rotation angle, a better initial solution can be found faster in the early stage.

[0121] Formula (19) is the probability of selecting the next point j at point i, which is related to pheromones, movement cost, waiting time penalty cost, rotation angle penalty cost and their coefficients.

[0122] Formula (20) and Formula (21) are the iterative rules of pheromone, including pheromone residue and pheromone increment on the optimal cost path.

[0123] For an ant traversing all service points, the ant starts at the distribution center and randomly selects the next customer point to visit based on a probability judgment function. Upon reaching the next customer point, it enters this point into the TB table. Meanwhile, the next selection only considers the permission table AL, which contains points it has not visited. This process continues until all points in the AL table are unreachable due to violating one of the time window, time, mileage, or volume constraints. At this point, the current vehicle K returns to the distribution center, and the ant departs from the distribution center, clearing its volume, mileage, and time. The ant continues traversing until all customer points have been visited and there are no more accessible points in the AL table. The next ant then begins traversing.

[0124] For an ant crawling route, if there are a total of 10 customer points, the ant's path can be recorded as a sequence, such as (0-6-4-2-9-11-0-1-3-5-12-0-7-8-10-13-0), and a total of 3 vehicles are dispatched.

[0125] Then the next ant is sent out until all ants have finished crawling. The ant with the smallest total cost among all ants is selected as the best cost of this iteration, and the pheromone is iterated and incremented. Then the next iteration is entered. After all iterations are completed, the top M best cost paths are output as alternative path planning.

[0126] Step S22: Filling the alternative path planning with workers to satisfy customer constraints, worker constraints, vehicle constraints, and skill and time window constraints;

[0127] ① Based on on-site research and data analysis, the combination of two installers on a vehicle is the most common and efficient combination in actual situations. Therefore, the algorithm focuses on two-person teams. A three-person team is only dispatched when the single-point order volume is too large and only a three-person team can complete the task. A one-person team is dispatched when only one person is left and the order is uncompleted.

[0128] ② For each vehicle, it can be either a private car with an owner or a public vehicle without an owner. If the owner belongs to a fixed group, the vehicle's skills are the skills of this fixed group. If the owner belongs to a non-fixed group, the vehicle's skills are the average skills of the owner and the non-fixed group without a vehicle. If it is a public vehicle without an owner, the vehicle's skills are the average skills of the non-fixed group.

[0129] ③ Regarding the order of vehicle dispatch, a vehicle sequence table will be prepared based on the situation on that day, and the order of vehicle dispatch will be carried out according to the vehicle sequence table. When dispatching a three-person team to a special point, the vehicle with the highest priority among the vehicles that meet the skill requirements will be dispatched.

[0130] ④ After determining the number of vehicle and passenger combinations, the ant colony algorithm is used to generate and iterate the paths, ultimately outputting the optimal solution. Furthermore, to address the shortcomings of the ant colony algorithm, the maximum-minimum pheromone method and dynamic parameter adjustment are used to prevent the ant colony from falling into local optima. Furthermore, the ant transition probability function is further designed to account for time window penalty costs and transition angles.

[0131] For the alternative path plan generated in step S21, personnel are filled in sequentially according to the order of the sub-paths. Different personnel combinations have different fitness for each sub-path. The personnel combination fitness table of each sub-path is iterated in a loop. If no solution is found, the previous sub-path is returned. Otherwise, the process continues until a feasible personnel filling solution is found.

[0132] Since a two-person team can meet more than 90% of the skill requirements, the staffing is very robust. Therefore, after the staffing of the algorithm is separated from the path scheduling, after multiple subsequent verifications of the algorithm, feasible solutions for staffing were successfully found in the results of 20 experimental runs, proving the feasibility of running the two parts of path scheduling and staffing separately.

[0133] Step S23: If none of the M alternative path plans can meet the constraints, then increase workers or increase working hours until a feasible path plan and logistics schedule are obtained.

[0134] If the optimal M solutions obtained in the ant colony algorithm cannot obtain a feasible solution for staff filling, then a plan for staff overtime or calling additional staff will be generated to obtain a feasible solution for the route scheduling plan.

[0135] Based on the above model, the following example is used for verification.

[0136] 1) Product database

[0137] 1. Product classification and basic installation time

[0138] The product database contains data such as the basic installation time of the product, the non-installation time associated with the product, the volume, and the sub-package relationship, as shown in Table 1.

[0139] Product categories are divided into three installation types: simple (delivery only, representative products such as home appliances), ordinary (simple installation, representative products such as complete furniture), and custom (complex installation). Taking cabinets as an example, all products can be divided into simple cabinets, ordinary cabinets, and custom cabinets.

[0140] Table 1 Installation duration database

[0141] Product Number Product Categories Basic installation time (min) volume Number of components 0 0 0.000 0 1 1 7 5.000 0.27 1 2 7 5.000 0.27 1 3 11 15.000 0.192 1 4 1 5.000 0.85 1 5 3 15.000 0.85 3 6 3 15.000 1.08 3 7 3 15.000 1.08 3 8 6 5.000 1.08 1 9 4 30.000 0.04 4 10 4 30.000 0.9804 4 11 4 30.000 0.21 4 12 4 30.000 0.18 4

[0142] 2. Non-installation time

[0143] Non-installation time is defined as the time it takes to load a product. This refers to the time it takes from the vehicle's arrival at the customer's location to the start of installation. This time is primarily influenced by the home environment and order attributes, including whether the vehicle can enter the residential complex, whether there are elevators in the complex, the number of items in the order, and the weight and volume of the order. Non-installation time is defined as a piecewise integer function. When the order number is 3 or fewer, the non-installation time is 15 minutes; when the order number is in the [4, 7] range, the non-installation time is 30 minutes; and when the order number exceeds 7, the non-installation time is 50 minutes.

[0144] 3. Product volume

[0145] The volume of the product and the vehicle's clearance determine the amount of cargo the vehicle can carry.

[0146] 4. Sub-components and kits

[0147] Due to the unique nature of furniture products, a single product may be packaged in multiple packages. Multiple components make up a kit. The volume of each component can be measured individually, while the installation time is determined by the total number of components in the kit. The installation time is evenly distributed across all components. For example, if the total installation time for a three-seater sofa is 15 minutes, then each component takes 5 minutes to install.

[0148] 2) Vehicle and Worker Database

[0149] 1. Vehicle data

[0150] The vehicle data includes the vehicle's headroom, daily depreciation cost, kilometer fuel consumption cost, and average driving speed. The vehicle calculation data for this experiment is shown in Table 2.

[0151] Table 2 Initial values ​​of vehicle attributes

[0152]

[0153] 2. Worker Data

[0154] Worker data includes bound workers, driving skills, installation skills, and various rate efficiencies. The rate efficiencies and basic installation time can be used to obtain the time taken by each (or group) worker to install a certain type of product. Rate efficiencies include personal main rate efficiencies, combined rate efficiencies, and personal auxiliary rate efficiencies. Personal main rate efficiencies refer to the efficiency increase of a worker who knows how to install a certain type of product relative to a person with a baseline efficiency of 1. A rate greater than 1 means faster than the baseline efficiency, and vice versa. The default worker's baseline efficiency does not differ in product categories, that is, the personal main rate efficiencies of workers installing Class A products and Class B products are the same. Combined rate efficiencies are defined as the improvement in personal efficiency of multiple people relative to one person, and auxiliary rate efficiencies refer to the improvement in overall efficiency of workers who do not know how to install by providing simple assistance. It can be considered as a substitute for personal main rate efficiencies when a certain skill is not known. In order to prevent the improvement in combined efficiency of people who do not know how to install from being higher than the improvement in combined efficiency of people who know how to install, this experiment sets the calculation formula for personal auxiliary efficiency as follows:

[0155] Two-person installation, C v =0.65*tan -1 (3x)

[0156] Three-person installation, C v =0.5*tan -1 (3x)

[0157] Where x is the average of the individual main rate efficiency gains of the workers who can install. This function ensures that the individual auxiliary efficiency is always lower than the lowest individual main rate efficiency gain in the combination. The final combination installation time can be obtained using the following formula.

[0158]

[0159] The worker information collection table is shown in Table 3:

[0160] Table 3 Worker information table

[0161]

[0162] The combined synergy is shown in Table 4:

[0163] Table 4 Combination synergy table

[0164] Combination Type Simple ordinary custom made Single 1 / / Double 1 1.3 1.2 Three people 1 1.2 1.1

[0165] 3. Relationship between people and vehicles

[0166] In the actual transportation and assembly process, there is a matching relationship between vehicles and workers, and between workers. That is, a vehicle can only be driven by a specific person, and some people may have combined vehicle attendance. These all restrict the available workers (or worker combinations) and vehicle attendance.

[0167] 3) Customer point database

[0168] Customer point data mainly includes the geographical location of the customer point, time window and product demand. The customer point calculation example of this experiment is shown in Table 5.

[0169] Table 5 Customer Demand Table

[0170]

[0171]

[0172] 4) Distance Matrix

[0173] The distance matrix represents the distances between distribution centers, customer points, and return destinations. Customer point data primarily includes the customer's geographic location, time window, and product demand. The system can obtain the latitude and longitude information for customer points, distribution centers, and vehicle return destinations. To ensure distance accuracy, this embodiment utilizes the AutoNavi Map API to return the actual driving distance between any two points using this latitude and longitude information.

[0174] 5) Other fixed parameters

[0175] 1. Distribution Center

[0176] The distribution center address is unique and is the warehouse address.

[0177] 2. Rest Time

[0178] Workers need a certain amount of rest time during the daily transportation and assembly process. It is stipulated that when the end time of a task falls after 12:30 for the first time, the workers will take a one-hour break and then start the subsequent work.

[0179] The embodiment of the present invention is to use matlab programming practice on a PC, and use ant colony algorithm to solve the model based on the above data to obtain the optimal solution. The specific results are shown in Table 6. The optimal delivery path solution is as follows: Figure 2 As shown, Figure 2 The horizontal and vertical coordinates in the middle represent longitude and latitude respectively. The five-pointed star represents distribution center 0, which shows the delivery path of 4 vehicles with a total cost of 520.62.

[0180] Table 6 Optimal intelligent dispatching solution

[0181] Solution Selection 1 The customer point visited by the first car is: 0 2 The first car crew is: 1 2 The customer point visited by the second car is: 0 8 12 13 7 The crew of the second car are: 3 8 The customer point visited by the third car is: 0 9 11 6 The crew of the third vehicle are: 5 7 The customer point visited by the fourth car is: 0 5 4 3 14 15 10 The crew of the fourth vehicle are: 4 6

[0182] Based on the data samples of 40 customer points in Table 5, the present invention found that the algorithm used in the present invention has the following advantages over manual order scheduling:

[0183] 1. Resource usage

[0184] Compare the number of workers and vehicles for manual and algorithmic scheduling. Manual scheduling requires 5 vehicles and 8 workers, while algorithmic scheduling requires 4 vehicles and 7 workers. Regardless of the number of vehicles or workers, algorithmic scheduling uses fewer resources to complete all customer service tasks. The resource usage comparison chart is as follows: Figure 3 shown.

[0185] 2. Workers’ working hours

[0186] Regarding the working hours of workers, the working hours of 8 people who manually arranged orders and the working hours of 7 people who arranged orders by algorithm are as follows: Figure 4 The average working hours for workers using manual scheduling were 7.73 hours, while the average working hours for workers using algorithmic scheduling were 9.32 hours. This is because the algorithmic scheduling strictly follows workers' off-duty times, and workers typically complete their work around 7 PM. However, due to scale constraints, the last order had insufficient customer points, so a single person could complete the work in just 6 hours.

[0187] In terms of stability, Figure 5 As shown in the figure, the algorithm scheduling is more stable, with a standard deviation of 1.19 hours. If only the first 6 workers are considered, the standard deviation is only 0.11 hours; manual scheduling has large fluctuations, with a standard deviation of 1.37 hours.

[0188] 3. Cost and path length

[0189] According to the cost calculation rules, each car needs a daily depreciation fee of 100 yuan, and the fuel consumption per kilometer is 1.2 yuan. The total cost comparison between manual scheduling and algorithm scheduling is as follows: Figure 6 As shown:

[0190] like Figure 7 As shown, the five vehicles manually scheduled traveled a total distance of 315.3 kilometers at a cost of 878.4 yuan. The four vehicles scheduled using the algorithm traveled a total distance of 254 kilometers at a cost of 704.8 yuan. Compared to manual scheduling, algorithmic scheduling reduced travel distance by 19.4% and reduced costs by 19.7%. Overall, algorithmic scheduling offers a significant improvement over manual scheduling by dispatchers.

[0191] The present invention discloses a method for scheduling and optimizing routes for integrated home furnishing logistics with a time window. Aiming at the problem of intelligent scheduling and route planning for integrated home furnishing logistics distribution on a comprehensive furniture service platform, an ant colony algorithm is used for minimum cost route planning. Furthermore, by establishing constraints such as customer constraints, worker constraints, vehicle constraints, and skill and time window constraints, optimal intelligent scheduling is achieved, which can simultaneously meet the customer service time window requirements and the workers' lunch break requirements. The model of the present invention not only meets the requirements for scheduling and route planning for home furnishing logistics, but also further expands the multi-factor scheduling model.

[0192] Example 2

[0193] The embodiment of the present invention provides a time-windowed integrated home furnishing logistics scheduling and route optimization system, including the following modules:

[0194] Model building module 31, used to establish the objective function and constraints of the home logistics distribution model based on integrated assembly: customer constraints, worker constraints, vehicle constraints, and skill and time window constraints;

[0195] The path planning and logistics scheduling module 32 is used to solve the model using an improved ant colony algorithm, and use M optimal path plans as alternative path plans; substitute the customer constraints, worker constraints, vehicle constraints, and skill and time window constraints into the alternative path plans, perform logistics scheduling, and use the path plan that meets all the constraints as the path plan, and obtain the corresponding logistics schedule; if none of the M alternative path plans can meet the constraints, then increase the number of workers or increase the working hours until a feasible path plan and logistics schedule are obtained.

[0196] The above embodiments are provided for the purpose of describing the present invention only and are not intended to limit the scope of the present invention. The scope of the present invention is defined by the appended claims. Various equivalent substitutions and modifications made without departing from the spirit and principles of the present invention are intended to be within the scope of the present invention.

Claims

1. A method for scheduling and optimizing the route of integrated home furnishing logistics with a time window, characterized in that: include: Step S1: Establish the objective function and constraints of the home furnishing logistics distribution model based on integrated assembly: customer constraints, worker constraints, vehicle constraints, and skill and time window constraints, specifically including: Step S11: Establish the objective function of the path optimization based on the integrated home logistics distribution model as shown in formula (1): （1） in, Points for all customers Collection of ; C, distribution center for all customer points and virtual endpoints A collection of , 0 represents the distribution center, represents a virtual endpoint; Points for customers To the customer point distance; For vehicles Collection ; Indicates that vehicle k travels from customer point i to customer point j, satisfying ； is the fuel consumption cost per kilometer; is the fixed cost per vehicle per day; Step S12: Establish the following constraints: (1) Customer constraints a) For each customer point, only one car can enter the point, as shown in formula (2): （2） b) For each customer point, only one car can leave the point, as shown in formula (3): （3） (2) Worker Constraints a) A worker cannot appear on the same car repeatedly, as shown in formula (4): （4） in, ; For workers A collection of , where 0 represents a virtual person without any skills; Indicates that worker 2 is in the collection The position number of Location Afterwards, in Before; b) A worker cannot appear on multiple vehicles, as shown in formula (5): （5） c) Workers have fixed collocations, as shown in formula (6): （6） (3) Vehicle restraints a) After completing the last customer assembly task, the vehicle returns to the virtual address but does not depart from that address, as shown in formula (7): （7） b) Path balance constraint, which ensures that a vehicle must leave a customer point after entering it, as shown in formula (8): （8） in, Indicates that the vehicle returns to the distribution center; Indicates that the vehicle does not return to the distribution center, but returns to the virtual address; c) The clearance limit for each vehicle is as shown in formula (9): （9） in, is the clearance of each vehicle k; For customers Order volume; d) There cannot be multiple combinations of people in a vehicle, as shown in formula (10): (10) (4) Skill and time window constraints a) Time balance constraint, driving skill and installation skill constraint, as shown in formula (11): （11） in, For home furnishings A collection of For customers The length of time required for the move; To correspond The basic installation time for each category in the For workers' union Install The installation time of a category. If the category is not installed in the combination, set a large integer E; For customers Need to install The number of species; It's a vehicle Arrival at customer point time; is the average speed of each vehicle; b) Waiting time constraint, as shown in formula (12): （12） in, It's a vehicle At the customer point Waiting time; It is the earliest delivery time booked by the customer; c) Customer point time window constraint, as shown in formula (13): （13） in, It is the latest delivery time booked by the customer; Points for customers time window; d) Specify the arrival time and number of installed products at the distribution center, as shown in formula (14): （14）； Step S2: Using an improved ant colony algorithm to solve the model, M optimal path plans are used as alternative path plans; the customer constraints, worker constraints, vehicle constraints, and skill and time window constraints are substituted into the alternative path plans to perform logistics scheduling. The path plan that satisfies all the constraints is used as the path plan, and the corresponding logistics schedule is obtained; if none of the M alternative path plans can satisfy the constraints, additional workers or increased working hours are added until a feasible path plan and logistics schedule are obtained, specifically including: Step S21: Use the ant colony algorithm to solve formula (1), and set An ant starts from the distribution center O and traverses all customer point sets C to obtain path plans and their corresponding pheromones. The path with the lowest total cost among all ants is the optimal path for this iteration. After iterating and incrementing the pheromones, the next iteration is performed. The first M optimal path plans are output as candidate path plans until the predetermined number of iterations is reached. Step S22: Filling the alternative path planning with workers so that it satisfies the customer constraints, worker constraints, vehicle constraints, and skill and time window constraints; Step S23: If none of the M alternative path plans can meet the constraints, workers are added or working hours are increased until a feasible path plan and logistics schedule are obtained.

2. A home furnishing logistics scheduling and route optimization system with time windows, characterized by: Includes the following modules: The model building module is used to establish the objective function and constraints of the home logistics distribution model based on integrated assembly: customer constraints, worker constraints, vehicle constraints, and skill and time window constraints, specifically including: Step S11: Establish the objective function of the path optimization based on the integrated home logistics distribution model as shown in formula (1): （1） in, Points for all customers Collection of ; C, distribution center for all customer points and virtual endpoints A collection of , 0 represents the distribution center, represents a virtual endpoint; Points for customers To the customer point distance; For vehicles Collection ; Indicates that vehicle k travels from customer point i to customer point j, satisfying ； is the fuel consumption cost per kilometer; is the fixed cost per vehicle per day; Step S12: Establish the following constraints: (1) Customer constraints a) For each customer point, only one car can enter the point, as shown in formula (2): （2） b) For each customer point, only one car can leave the point, as shown in formula (3): （3） (2) Worker Constraints a) A worker cannot appear on the same car repeatedly, as shown in formula (4): （4） in, ; For workers A collection of , where 0 represents a virtual person without any skills; Indicates that worker 2 is in the collection The position number of Location Afterwards, in Before; b) A worker cannot appear on multiple vehicles, as shown in formula (5): （5） c) Workers have fixed collocations, as shown in formula (6): （6） (3) Vehicle restraints a) After completing the last customer assembly task, the vehicle returns to the virtual address but does not depart from that address, as shown in formula (7): （7） b) Path balance constraint, which ensures that a vehicle must leave a customer point after entering it, as shown in formula (8): （8） in, Indicates that the vehicle returns to the distribution center; Indicates that the vehicle does not return to the distribution center, but returns to the virtual address; c) The clearance limit for each vehicle is as shown in formula (9): （9） in, is the clearance of each vehicle k; For customers Order volume; d) There cannot be multiple combinations of people in a vehicle, as shown in formula (10): (10) (4) Skill and time window constraints a) Time balance constraint, driving skill and installation skill constraint, as shown in formula (11): （11） in, For home furnishings A collection of For customers The required moving time; To correspond The basic installation time for each category in the For workers' union Install The installation time of a category. If the category is not installed in the combination, set a large integer E; For customers Need to install The number of species; It's a vehicle Arrival at customer point time; is the average speed of each vehicle; b) Waiting time constraint, as shown in formula (12): （12） in, It's a vehicle At the customer point Waiting time; It is the earliest delivery time booked by the customer; c) Customer point time window constraint, as shown in formula (13): （13） in, It is the latest delivery time booked by the customer; Points for customers time window; d) Specify the arrival time and number of installed products at the distribution center, as shown in formula (14): （14）； The route planning and logistics scheduling module is used to solve the model using an improved ant colony algorithm and select M optimal route plans as alternative route plans; substitute the customer constraints, worker constraints, vehicle constraints, and skill and time window constraints into the alternative route plans to perform logistics scheduling, select the route plan that meets all the constraints as the route plan, and obtain the corresponding logistics schedule; if none of the M alternative route plans meet the constraints, add workers or increase working hours until a feasible route plan and logistics schedule are obtained, specifically including: Step S21: Use the ant colony algorithm to solve formula (1), and set An ant starts from the distribution center O and traverses all customer point sets C to obtain path plans and their corresponding pheromones. The path with the lowest total cost among all ants is the optimal path for this iteration. After iterating and incrementing the pheromones, the next iteration is performed. The first M optimal path plans are output as candidate path plans until the predetermined number of iterations is reached. Step S22: Filling the alternative path planning with workers so that it satisfies the customer constraints, worker constraints, vehicle constraints, and skill and time window constraints; Step S23: If none of the M alternative path plans can meet the constraints, then workers are added or working hours are increased until a feasible path plan and logistics schedule are obtained.

Images