CN114912976A - Collaborative fulfillment optimization method for full-channel drug retail orders - Google Patents
Collaborative fulfillment optimization method for full-channel drug retail orders Download PDFInfo
- Publication number
- CN114912976A CN114912976A CN202210503610.8A CN202210503610A CN114912976A CN 114912976 A CN114912976 A CN 114912976A CN 202210503610 A CN202210503610 A CN 202210503610A CN 114912976 A CN114912976 A CN 114912976A
- Authority
- CN
- China
- Prior art keywords
- order
- decision
- model
- vehicle
- distribution
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0633—Lists, e.g. purchase orders, compilation or processing
- G06Q30/0635—Processing of requisition or of purchase orders
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
- G06Q10/083—Shipping
- G06Q10/0835—Relationships between shipper or supplier and carriers
- G06Q10/08355—Routing methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
- G06Q10/087—Inventory or stock management, e.g. order filling, procurement or balancing against orders
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0605—Supply or demand aggregation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Development Economics (AREA)
- Physics & Mathematics (AREA)
- Finance (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Marketing (AREA)
- Human Resources & Organizations (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Game Theory and Decision Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the field of all-channel operation management, and provides a collaborative fulfillment optimization method for all-channel drug retail orders. Aiming at the joint decision problem of order distribution, cooperative multi-commodity network flow and vehicle scheduling of the full-channel medicine retail order fulfillment decision, a decomposition-iteration-search algorithm is designed and realized based on the idea of decomposition and iteration, the association subproblems in the complex joint decision problem are decoupled and decomposed, and a decision scheme is obtained through parameter transfer and iterative solution. The invention has the advantages that the medicine full-channel order fulfillment problem of a plurality of inventory positions with order complete delivery requirements is effectively solved by adopting a decomposition idea, the information among decision sub-problems is fully utilized, the order processing process and the vehicle path are optimized, the order fulfillment cost is reduced, the scientificity of order fulfillment scheme decision and the efficiency of order processing are improved, and the decision support is provided for the full-channel medicine order fulfillment operation practice.
Description
Technical Field
The invention belongs to the field of all-channel operation management, and relates to a collaborative fulfillment optimization method for all-channel drug retail orders.
Background
The whole channel medicine retail mode can combine the advantages of price advantage, class advantage, service interaction and the like of an online pharmacy with the convenience of an offline physical pharmacy, butt various customer demands and dispersed medicine stocks through the form of 'ordering by online pharmacy, performing by offline pharmacy', breaks the space-time limitation of medicine supply and demand, provides additional medicine services such as insurance, pharmacist consultation and on-time delivery by offline pharmacy, and improves the service experience and satisfaction of consumers.
However, while the retail scale of medicine in a full channel continues to expand, its daily operation is also greatly challenged, with the first challenge being order fulfillment decisions, i.e., from which pharmacies the medicine is shipped from and when and via which routes it is delivered by which vehicle. Due to the particularity of the commodities such as medicines, the demand of the whole-channel retail of the medicines on the timeliness and the punctuality of distribution is high, and meanwhile, the contradiction between dispersed inventory and coordinated distribution is faced, and the fulfillment decision is faced with the following difficulties: in the aspect of inventory, at present, China can retail up to 20 ten thousand medicines on line, the storage and transportation conditions of the medicines are strict and diversified, and the types and the quantity of the medicines which can be stored in physical drugstores are limited, so that the medicine full-channel operation has the characteristics of multiple inventory places and overlapped inventory; in distribution, the treatment of a disease often requires the combination and matching of multiple drugs, the drug compatibility is also a special attribute of the drugs, the drug order often has an attribute of 'one single product and multiple products', the order requirement has high randomness, and the drug distribution needs to ensure that the order is completely delivered in consideration of the specificity of drug commodities. In conclusion, the order fulfillment optimization of the medicine full-channel retail is a collaborative decision for order distribution, collaborative multi-commodity network flow and vehicle scheduling problems, is a new collaborative decision problem, has huge decision space and higher requirements on solving speed and quality, needs to construct a new collaborative decision model and a corresponding efficient and rapid optimization algorithm, and ensures efficient and on-time medicine full-channel order distribution.
Aiming at the characteristics of various drug industries and dispersed inventory and the use characteristics of drug compatibility and time requirements, the fulfillment optimization of the drug full-channel retail order comprehensively considers the characteristics of multiple inventory places, order time limitation, vehicle distribution of multiple parking lots and the like, and the NP difficult problems of mutual coupling of an order distribution problem, a multi-commodity network flow problem and a vehicle path problem are included. The problem is the coupling integration of two scientific problems of an unpaired goods taking and delivery vehicle path problem (UPDVRP) and a multi-time goods taking and delivery vehicle path problem (MPD-VRP-TW) with a time window, and is a new emerging scientific problem along with the popularization of medicine full-channel retail. The invention combines the two problems for joint optimization, the problem complexity is higher, and at present, in the field of all-channel operation optimization, no patent is available about the integration problem of medicine all-channel collaborative order fulfillment with time limitation and a related optimization method. In the aspect of academic research, although the achievements of optimization methods for solving two scientific problems of UPDVRP and MPD-VRP-TW exist, for a new order fulfillment problem which is considered by the patent and is coupled with UPDVRP and MPD-VRP-TW, the new problem has the characteristics of large decision space, complex and irregular decision variable relation, and no related research achievements exist at present. The precise algorithm has long solving time, so the application practice is difficult to meet; the existing heuristic algorithm solves the problem of irregular relation of the decision variables and has the defects of low search efficiency, poor scheme effect and the like, so the existing optimization method is difficult to solve the new problem.
Under the background of medicine full-channel operation, aiming at a batch of multi-product online orders received at a certain moment in a city, the medicine demand in the orders is combined, the inventory distribution of physical drugstores, the positions of vehicles distributed in a city road network are considered, the medicines in the orders are distributed to which drugstores, and how to arrange the vehicles for order coordinated distribution, wherein the decision content comprises the contents of decision order distribution, vehicle dispatching and delivery coordinated distribution paths and the like, so that the order fulfillment operation cost is reduced. The inventory of the full channel mode has the characteristics of position dispersion and logic sharing, the requirement of inventory cooperation exists, and the difficulty of full channel order fulfillment decision is further increased. In the invention, due to the inventory holding limitation, multiple physical pharmacy stores may be required to complete the stock of the same order in a coordinated manner, that is, different drug requirements in the same order may be allocated to different physical pharmacy stores for stock, and then the order complete delivery requirement is realized through sequential access of the same vehicle, that is, the order requirement is completed through one-time delivery. In enterprise practice, greedy rules are often adopted for decision making when facing such problems, but because the problems comprise a plurality of decision sub-problems and coupling relations exist among the sub-problems, the decision variables of the problems are complex and irregular in structure, and the decision making method based on the greedy rules is poor in effect.
Disclosure of Invention
In a full channel operation mode, according to the inventory characteristics of multiple inventory positions and distributed inventory of full channel medicine retail and the order characteristics of 'one single multiple product' and complete order delivery, the invention comprehensively considers inventory limitation and time requirements, and carries out decisions such as order distribution, coordinated multi-product network flow, vehicle scheduling and the like among a plurality of fulfillment subjects, thereby reducing the fulfillment cost of full channel retail orders. The invention provides a collaborative fulfillment decision optimization method of a full-channel medicine retail order by decomposition-iteration-search, which is used for jointly deciding the problems of order distribution, multi-commodity network flow, vehicle path and the like on the premise of considering the coupling relation of a plurality of stock positions. The invention designs and realizes a decomposition iteration search algorithm based on the idea of decomposition iteration, decouples and decomposes the associated subproblems in the complex joint decision problem, and iteratively realizes joint solution through iterative parameter transfer obtained by subproblem solution so as to obtain the solution of the original problem, thereby scientifically and effectively solving the problem of the plurality of coupled decision subproblems. The invention can provide decision support for the distribution scheduling of the whole channel operation or the logistics park with the same requirement, and improve the scientificity, effectiveness and high efficiency of the operation scheme.
The basic principle of the invention is as follows: the problem of collaborative fulfillment of full-channel retail drug orders is described as an operational problem for a region of full-channel drug retailers to satisfy the needs of drug orders for residents of the region using a network of drug stores comprised of a plurality of physical drug stores and an own fleet of distribution vehicles. The retailer runs multiple varieties of pharmaceuticals, and the inventory varieties between drug stores are partially overlapping, providing the customer with an order delivery time commitment. Because strict storage and transportation conditions of the medicines require temperature control during distribution, the distribution process has the medicine temperature control transportation cost related to the transportation time, and the cost is related to the medicine types. It is assumed that the order quantity of the same drug in a single order is only one at most, because multiple requested orders are processed in substantially the same way as a single requested order. Meanwhile, the vehicles of the distribution fleet are dispersed in a road network, and all the vehicles have transportation volume limitation. Under the condition, the problem is divided into two sub-problems by adopting a research idea of decomposition-iteration-search based on inventory information and order data, then an improved Hungarian-Dijkstra algorithm (MHDA) and a Greedy Search Algorithm (GSA) are respectively designed to solve the two sub-problems, the iteration process of the algorithm is adjusted through commodity flow variables until the commodity flow variables obtained by the two sub-problems can be matched, and finally the iteration solution is further optimized by utilizing the whole variable neighborhood search algorithm to further obtain a problem result.
The technical scheme of the invention is as follows:
a collaborative fulfillment optimization method for a full-channel medicine retail order adopts a decision support system which comprises an initialization module, an information input module, a calculation module 1, a calculation module 2, a calculation module 3, an iteration coordination module, a calculation module 4 and a decision output module. The functions of the various modules are as follows: the initialization module is used for acquiring entity network data and fleet data as parameters of a decision model and resource constraint information; the information input module is used for inputting order demand data of a certain batch; the calculation module 1 is used for modeling the collaborative fulfillment problem of the full channel medicine retail order, and decomposing the original problem model according to the entity network data and the order demand information to obtain two sub models (a model A and a model B) and related parameters; the calculation module 2 adopts an improved Hungarian-Dijkstra algorithm (MHDA) to solve the model A of the calculation module 1, and obtains a distribution scheme of the medicines in the order and distribution information of the order and the vehicle as input and iteration parameters of the calculation module 3; the calculation module 3 solves the model B of the calculation module 1 by adopting a search algorithm based on a plurality of neighborhoods based on the distribution scheme obtained by the calculation module 2 to obtain a complete order fulfillment scheme and obtain new distribution information of orders and vehicles as iteration parameters; the iterative coordination module judges whether the orders of the calculation module 2 and the calculation module 3 are matched with the distribution information of the vehicles or not to control an iterative program; the calculation module 4 performs search optimization on the solution after the iteration is completed through neighborhood search; the decision output module is used for expressing the decision result obtained by the calculation module 4 in a readable form and providing the decision result to a decision maker as decision support.
Based on the above thought and system module description, the technical scheme of the invention comprises the following specific steps:
(1) an initialization module: problem parameter setting
Step 1: the method comprises the steps of representing a pharmacy network of a certain area as a network consisting of | P | physical pharmacies dispersed in the area, providing | M | medicines for the area, wherein the medicine inventory of each physical pharmacy is partially overlapped but not completely identical, and each medicine has corresponding temperature-controlled transportation cost due to strict requirements of medicine transportation conditions. In addition, the vehicles with the same quality are distributed at different positions in the road network of the region, a volume limit V exists for the vehicles, and the vehicles have time parameters and cost parameters in the road network. These network settings constitute the decision problem parameters for the problem and serve as resource constraint information for subsequent models.
(2) Information input module
Step 2: data of a batch of | O | order requirements is input, which mainly comprises order positions, order times and medicine requirements in the orders.
(3) The calculation module 1: problem modeling and problem decomposition
Step 3.1: problem modeling
The method comprises the steps of describing a collaborative fulfillment problem of a full-channel medicine retail order based on an initialization module and an information input module, and establishing a mixed integer programming mathematical model TD-UMPD-VRPT, wherein main decisions comprise a vehicle dispatching decision U, a medicine distribution decision Y in the order, a commodity flow decision Z and a vehicle flow decision X.
Step 3.2: problem resolution
Decomposing the original problem model according to entity network data such as pharmacy distribution, pharmacy inventory data and the like and order demand information to obtain two decomposed submodels (a model A and a model B), and adding effective constraints to the submodels through corresponding coupling constraints and parameters to facilitate subsequent iteration;
(4) the calculation module 2: solving model A and corresponding iteration parameters
Step 4.1: solution of model A
The improved Hungarian-Dijkstra algorithm (MHDA) is obtained by embedding a Dijkstra algorithm in the Hungarian algorithm to solve the model A, namely, the Dijkstra algorithm is used for solving the problem of the shortest path of multiple vehicles to obtain a cost matrix corresponding to the multiple-medicine commodity flow, and the Hungarian algorithm is used for solving the medicine distribution problem in the order according to the cost matrix to obtain the order distribution result comprising the multiple-commodity network flow.
Step 4.2: obtaining an iteration parameter
Obtaining a distribution scheme of the medicines in the order and distribution information of the order and the vehicle according to the solving result of the model A, and taking the distribution scheme and the distribution information as input and iteration parameters of the next step;
(5) the calculation module 3: solving model B and corresponding iteration parameters
Step 5.1: solution of model B
And (3) based on the distribution scheme obtained in the step (4.2), obtaining an initial path scheme by adopting a Greedy Search Algorithm (GSA), improving the initial path scheme through neighborhood search, further obtaining a solving result of the model B, and obtaining a complete order fulfillment scheme.
And step 5.2: solution of model B
Obtaining new order and vehicle distribution information as iteration parameters according to the solving result of the model B;
(6) iterative coordination module
Step 6: judging whether the orders obtained in the step 4.2 and the step 5.2 are matched with the distribution information of the vehicles, if so, obtaining a final iterative order fulfillment scheme, and carrying out the next step; otherwise, feeding back the order and vehicle distribution information of the step 5.2 to the step 4, and re-executing the iterative program;
(7) the calculation module 4: search improvement
And 7: and (4) further searching and optimizing the final iteration order fulfillment scheme obtained in the step (6) by adopting a neighborhood structural design variable neighborhood searching algorithm.
(8) Decision output module
And 8: and (4) converting the results of the vehicle dispatching decision U, the traditional Chinese medicine distribution decision Y, the commodity flow decision Z and the vehicle flow decision X obtained in the step (7) into a readable form for representation, and providing the readable form as decision support for a decision maker.
The invention has the advantages that:
1. aiming at the whole-channel medicine retail operation, the invention provides a new order fulfillment problem caused by the whole-channel characteristics and the medicine retail characteristics, and formally describes the problem through modeling;
2. the invention considers the characteristics of stock dispersion, medicine compatibility, time requirement and the like, adopts a decomposed research idea to effectively solve the fulfillment problem of the medicine full-channel retail orders of a plurality of stock positions with the complete delivery requirement of the orders, and can fully utilize the information among the decision sub-problems and optimize the order processing process and the vehicle path;
3. the invention provides decision support for the distribution practice developed by the full-channel drug retailer, can improve the scientificity of order fulfillment scheme decision and the high efficiency of order processing, and further improves the efficiency of enterprise operation and resource utilization.
Drawings
Fig. 1 is a schematic diagram of the structural relationship between the system modules of the decision support system in the collaborative fulfillment method of the full channel retail order of drugs of the present invention.
FIG. 2 is a block diagram of a method for collaborative fulfillment optimization of a full channel retail order for pharmaceuticals of the present invention.
Detailed Description
In order to make the contents and advantages of the present invention clearer, the following detailed description of the embodiments of the present invention will be fully described in conjunction with the accompanying drawings.
Fig. 1 is a schematic diagram of a structural relationship between system modules of a decision support system for collaborative fulfillment of a full-channel drug retail order of the present invention, and it can be seen that the decision support system for collaborative fulfillment of a full-channel drug retail order of the present invention is composed of an initialization module, an information input module, a calculation module 1, a calculation module 2, a calculation module 3, an iterative coordination module, a calculation module 4, and a decision output module, wherein the initialization module is information of enterprise operation resources, the information input module receives full-channel order information, and then, for the order information, the allocation scheme and the distribution path scheme of the order are calculated by the calculation module 1, the calculation module 2, the calculation module 3, the iterative coordination module, and the calculation module 4 under the existing operation resource conditions, so as to complete a decision for collaborative order fulfillment. Referring to fig. 1, the collaborative fulfillment optimization method for a full channel retail drug order provided by the present invention includes the following steps:
step 1: the initialization module is used for inputting the operation resources into the system and determining the decision problem parameters and the corresponding resource constraints; the related parameters comprise the position of the physical pharmacy, the inventory condition of the pharmacy, the current position of the vehicle, the time parameter and the cost parameter of the vehicle driving, and the temperature-controlled transportation cost parameter of the medicine.
Step 2: the information input module inputs data of order demands of a certain batch, and the data mainly comprises order positions, order times and medicine demands in the orders.
And step 3: the calculation module 1 describes the fulfillment problem of the whole channel medicine retail order collaboration based on the initialization module and the information input module, and establishes a mixed integer programming mathematical model, wherein the main decisions comprise a vehicle dispatching decision U, a medicine distribution decision Y in the order, a commodity flow decision Z and a vehicle flow decision X. Decomposing the original problem model according to entity network data such as pharmacy distribution, pharmacy inventory data and the like and order demand information to obtain two decomposed submodels (a model A and a model B), and adding effective constraints to the submodels through corresponding coupling constraints and parameters to facilitate subsequent iteration;
and 4, step 4: the calculation module 2 adopts an improved Hungarian-Dijkstra algorithm (MHDA) to solve the model A, obtains an order distribution result comprising a multi-commodity network flow, further obtains a distribution scheme of medicines in the order and distribution information of the order and vehicles, and takes the distribution scheme and the distribution information as input and iteration parameters of the next step;
and 5: and (3) obtaining an initial path scheme by adopting a Greedy Search Algorithm (GSA) based on the distribution scheme obtained in the step (4), improving the initial path scheme through neighborhood search, further obtaining a solving result of the model B, and obtaining a complete order fulfillment scheme. Further obtaining new order and vehicle distribution information as iteration parameters;
step 6: the iteration coordination module is used for judging whether the orders obtained in the step 4 and the step 5 are matched with the distribution information of the vehicles or not, if so, a final iteration order fulfillment scheme is obtained, and the next step is carried out; otherwise, feeding back the order and vehicle distribution information of the step 5 to the step 4, and re-executing the iterative program;
and 7: and (4) adopting a proper neighborhood structure design variable neighborhood searching algorithm to further search and optimize the final iteration order fulfillment scheme obtained in the step (6).
And 8: and the decision output module is used for converting the vehicle dispatch decision U, the medicine distribution decision Y in the order, the commodity flow decision Z and the vehicle flow decision X obtained in the step 7 into readable form representation to serve as decision support to be provided for a decision maker.
In the method provided by the invention, a time-dependent unpaired and multipoint goods taking and delivering vehicle path problem TD-UMPD-VRPT (time-dependent unpaired-multiple pick-up and delivery with delivery time contract) model can be established by operating the resource information and the order demand information, a decomposition-iteration-search algorithm is designed, the problem is decomposed into two sub-problems to be solved respectively, then the iteration process of the algorithm is adjusted by commodity flow variables until the iteration program is ended, a proper neighborhood structure is designed, the iteration result is further optimized by variable neighborhood search, and finally a collaborative order fulfillment scheme is obtained. According to the method provided by the invention, under the full-channel medicine retail mode, the order fulfillment characteristics and the medicine demand characteristics of the full channel are fully considered, and an optimization scheme is provided for the physical pharmacy stock and vehicle distribution in the full-channel operation.
With reference to fig. 2, a framework diagram of a collaborative fulfillment optimization method for a full-channel retail drug order of the present invention is shown, taking a full-channel pharmacy network composed of | P | physical pharmacies, | K | vehicles, and | M | drug stocks, and taking a collaborative order fulfillment scheme for determining | O | order requirements as an example, the following steps are respectively implemented as follows:
step 1: the operation resource allocation condition of a full-channel retailer is determined firstly, wherein M is a medicine set, K is a vehicle set, P is a physical pharmacy set, and N is P Representing a set of pharmacy locations. In terms of setting parameters, stock conditions s pm Indicating the stock of the drugs m held by the pharmacy p, and belongs to N P ,m∈M；v m Represents the volume of the drug m, V is the volume of the vehicle; α is the fixed use cost of the vehicle, β ij Is the cost of the vehicle's path from point i to point j, θ ij Is the vehicle's path time from point i to point j; gamma ray m Is the temperature-controlled transportation cost of the medicine m in unit time.
Step 2: determining received order demand information, O being an order set, N o Indicating a set of order locations, N being the set of all points in the area, N ═ N P ∪N O Where i ∈ N uses plane coordinates (i ∈ N) a ,i b ) Represents; d o Is the delivery deadline for order O, O ∈ O.
And step 3: the calculation module 1 describes the collaborative fulfillment problem of the full-channel medicine retail order based on the operation resource configuration of the initialization module and the order demand information of the information input module, constructs a TD-UMPD-VRPT model, and acquires an order fulfillment scheme comprising commodity flow and vehicle flow according to information such as pharmacy distribution, pharmacy inventory data, order demand and the like. The objective function of the TD-UMPD-VRPT model is expressed by formula (1):
wherein M is a drug set, M o Representing a drug set in an order O, K being a vehicle set, P being a physical pharmacy set, N P Representing a set of pharmacy locations, O being a set of orders, N O Indicating a set of order locations, N being the set of all points in the area, N ═ N P ∪N O Where i ∈ N uses plane coordinates (i ∈ N) a ,i b ) Represents; α is the fixed use cost of the vehicle, β ij Is the cost of the vehicle's path from point i to point j, θ ij Is the vehicle's path time from point i to point j; gamma ray m Is the temperature-controlled transportation cost of the medicine m in unit time; e.g. of the type k Is the current position of the vehicle k, e ′ Is a virtual end point, and the distances from all points to the virtual end point are zero; min is a minimum function, decision variable u of type 0-1 k Means for indicating whether to use decision variable x of type k, 0-1 for vehicle kij Decision variable of type 0-1, representing whether vehicle k has traveled a path from point i to point j
Indicating whether the vehicle k is carrying the drug m in order o, is traversing the path from point i to point j.
The constraints of the TD-UMPD-VRPT model are expressed by equations (2) - (20):
(a) order splitting and allocation decision-related constraints
(b) Coupled constraints for order distribution and multi-commodity flow decision
(c) Multi-commodity flow decision-related constraints
(d) Coupling constraints for multiple commodity flow and vehicle path decisions
(e) Taking and delivering vehicle path decision related constraints
Wherein, decision variables of type 0-1
Decision variable b of type 0-1, indicating that the drug m in order o is in charge of stock by pharmacy p kij Is an auxiliary variable indicating whether node i is before node j (not necessarily between neighbors) in the path of vehicle k. L is a real number not less than 10^6 and is a continuous decision variable t ki And
respectively representing the time at which the vehicle k arrives and departs from point i. Stock status s pm Indicating the stock of the drugs m held by the pharmacy p, and belongs to N P ， m∈M；v m Represents the volume of the drug m, V is the volume of the vehicle; d is a radical of o Is an order formThe delivery deadline of O, O ∈ O.
By removing equation (10), the model can be decomposed into two sub-problems, model a and model B.
Model A:
constraint conditions are as follows: including formula (2), formula (3), formula (8) and
wherein, K is O Is an order-vehicle matching information set, k o Refers to the vehicle responsible for order o. By gathering K order-vehicle matching information O For incorporation into model A, partial constraints in the original model are updated to new constraints by aggregating K the order-vehicle matching information O Incorporated into model a, the partial constraints in the original model are updated to new constraint equations (22) - (26). For the first iteration, default allocation of all orders to a certain vehicle may be set, e.g., for the first iteration, K is set 1 O The vehicle 1 is responsible for all orders, which does not affect the first iterationSolving effect of the time model A.
Model B:
constraint conditions are as follows: formulae (11) to (20) and
wherein N is + Is the set of physical pharmacies that get the assignment, and formula (28) represents that the physical pharmacy that gets the assignment is visited at least once, this formula being the solution to model a as input to model B. Vehicle flow decision x for model B kij Corresponding to corresponding commodity flow variables
Therefore, solving the model B can obtain new order-vehicle matching pair information K O This information is fed back into model a for a second iteration.
The newly added constraints of the model A and the model B are provided, so that the transmission and the connection of the solving processes of the two sub-problems are realized, and the solution of the problem is continuously corrected and optimized along with the iteration process. And respectively solving the two subproblems, controlling an iteration process through iteration parameters obtained by the subproblems until the iteration parameters are matched to complete joint solution, and obtaining a solution of the original problem.
And 4, step 4: the calculation module 2 adopts an improved Hungarian-Dijkstra algorithm (MHDA) to solve the model a, the model a mainly comprises two decisions, namely an order distribution decision and a multi-commodity network flow decision, the objective function is to minimize the total commodity flow cost, and the total commodity flow cost in the model a is the sum of the flow direction cost of each medicine m of each order o, namely, the assignment problems of each medicine are mutually independent, so that the corresponding assignment problem solution can be performed on each medicine, namely, a cost matrix is given, and the medicine distribution decision can be performed by solving a standard distribution problem (AP). Also, given a commodity allocation plan, minimizing commodity flow costs can be viewed as a fixed multi-vehicle shortest path problem with one start and end node, so model a can be represented as a combination of APs and MVSPs. Therefore, the model A can be converted into a combination of standard distribution problems (AP) of various medicines m and a multi-vehicle shortest path problem (MVSP), a Dijkstra algorithm is embedded into a Hungarian algorithm, a polynomial time optimization algorithm (improved Hungarian-Dijkstra algorithm, MHDA) is constructed for solving the model, namely, the MVSP problem is solved through the Dijkstra algorithm to obtain a cost matrix formed by commodity flow cost of each medicine as parameters of the AP problem, and the AP problem is solved through the Hungarian algorithm to obtain a distribution scheme of the medicines in the order.
The demand of the medicines in the order provided by each pharmacy holding each medicine is determined according to the inventory condition, and then the commodity flow cost distributed to all pharmacies holding the inventory of the medicines for each medicine is calculated. More specifically, element d in the cost matrix of drug m ab Means that a pharmacy a holding a drug m is responsible for the commodity flow cost of an order drug pair b, wherein the order drug pair is an order o having the demand for the drug m m An information pair with the medicine m, which is task index information for each column in the cost matrix, and execution subject index information for each row in the cost matrix is the pharmacy a holding the medicine m. Cost parameter d ab From pharmacy a to order o m The minimum commodity flow cost of the position node comprises commodity flow direction information and responsible vehicle information. Solution Y by model A 1 The distribution scheme M of the medicines in the order can be obtained P And order and vehicle allocation information
As input and iteration parameters for the next step.
The following illustrates how the iteration parameters are derived from the solution of model a.
For example, the solution for model A is
Wherein, the first isThe column is the order number, the second column is the drug number, and the third column is the pharmacy number. From Y 1 It can be seen that order 1 is assigned to pharmacy 1 and pharmacy 2 for stock, order 2 is assigned to pharmacy 3 for stock, and order 3 is assigned to pharmacy 3 and pharmacy 4 for responsibility. In addition, pharmacy distribution result M of the medicines in the order P Responsible for pharmacy 1 for the drugs 5 in order 1; pharmacy 2 is responsible for order 1 medication 2; pharmacy 3 is responsible for drugs 1 and 4 in order 2, and drug 2 in order 3; pharmacy 3 is responsible for drugs 1 and 3 in order 3; these results will be used as input for model B in the next step.
And 5: calculating module 3, distribution scheme M of medicine in order obtained in step 4 P A Greedy Search Algorithm (GSA) is designed for inputting the model B, and an effective neighborhood structure is designed for the multiple goods taking characteristics of the model B to improve the search efficiency of the algorithm. GSA includes two operations: path generation and local search. The former generates a path scheme by inserting nodes in a path having a minimum insertion cost, and the latter generates a path scheme by pairing decision variables u k And
to improve the path scheme.
Taking vehicle k as an example, first, the distribution information of the order and the vehicle is obtained based on model A
A feasible vehicle path is generated for the vehicle having the ordered mission. Based on the access sequence, vehicle load and delivery time constraints, pluggable candidate nodes for the current path of vehicle k are found. And then, selecting the candidate node position with the lowest insertion cost to be inserted into the node to be accessed. Sequentially inserting nodes to be accessed greedily, and updating the path content of the vehicle k until all order demands are met, so that a feasible path scheme is obtained; if some order tasks cannot be distributed by the corresponding vehicles, the order is distributed to the unused vehicles to be responsible. After the initial routing plan is generated, a local search is applied to optimize the path plan. Finally, obtaining the result of the solution model B, namely obtaining the complete orderAnd (4) fulfilling the scheme, wherein new order and vehicle distribution information can be obtained based on the path scheme, and is used as an iterative parameter to be fed back to the model.
The following illustrates how the iteration parameters are derived from the solution of model B.
For example, the solution for model B is
Where the first column is the order number, the second column is the drug number, the third and fourth columns are the arcs in the path, and the fifth column is the vehicle. From Z, new order and vehicle distribution information can be obtained
Order 1 is charged by vehicle 1 and both order 2 and order 3 are charged by vehicle 2. These results will be fed back as iteration parameters into the model a of the previous step.
Step 6: the iterative coordination module obtains the optimal solution distribution scheme Y by solving the model A 0 And commodity flow information Z 0 Further, the distribution scheme of each medicine in each order, the picking task of each pharmacy and the vehicle information matched with each order can be obtained
They are passed into model B as input information for model B. Then, a vehicle path scheme is obtained by solving the model B, and further, new order-vehicle distribution information can be obtained
Judgment of
And
whether the order is matched or not is judged, if so, the iteration process is completed to obtain a final iteration order fulfillment scheme, and the next search improvement is carried out; if it isIf not, the order-vehicle distribution information obtained by the result of the model B is used
Input to model a as feedback information and re-execute the iterative procedure.
And 7: and the calculation module 4 is used for further searching and optimizing the final iteration order fulfillment scheme by adopting a proper neighborhood structure design variable neighborhood searching algorithm in order to improve the order fulfillment scheme obtained after the iteration is finished. The main idea is to obtain an order requirement set to be met by finally iterating an order fulfillment scheme through different neighborhood structure destruction, and then reinserting the order requirement to be met into the scheme by using a local search program based on minimum insertion cost to meet all order requirements, so as to obtain an order fulfillment scheme of information. According to the characteristics of problems, the designed neighborhood structure mainly comprises three types of all goods taking and delivering tasks for exchanging two vehicles, a partial distribution result adjustment part and a partial path result adjustment part.
And 8: a decision output module for sending the final vehicle dispatch decision u k Order form for medication dispensing decision
Commodity flow decision making
Vehicle flow decision x kij The result is converted into a readable form representation and provided to the decision maker as decision support.
Claims (2)
1. A cooperative fulfillment optimization method for a full-channel drug retail order is characterized in that an adopted decision support system comprises an initialization module, an information input module, a calculation module 1, a calculation module 2, a calculation module 3, an iteration coordination module, a calculation module 4 and a decision output module; the functions of the various modules are as follows: the initialization module is used for acquiring entity network data and fleet data as parameters of a decision model and resource constraint information; the information input module is used for inputting order demand data of a certain batch; the calculation module 1 is used for modeling the collaborative fulfillment problem of the full channel medicine retail order, and decomposing the original problem model according to the entity network data and the order demand information to obtain two sub models and related parameters; the calculation module 2 adopts an improved Hungarian-Dijkstra algorithm MHDA to solve the model A of the calculation module 1, and obtains a distribution scheme of medicines in the order and distribution information of the order and the vehicle as input and iteration parameters of the calculation module 3; the calculation module 3 adopts a search algorithm based on a plurality of neighborhoods to solve the model B of the calculation module 1 based on the distribution scheme obtained by the calculation module 2 to obtain a complete order fulfillment scheme and obtain new distribution information of orders and vehicles as iteration parameters; the iterative coordination module judges whether the orders of the calculation module 2 and the calculation module 3 are matched with the distribution information of the vehicles or not to control an iterative program; the calculation module 4 performs search optimization on the solution after the iteration is completed through neighborhood search; the decision output module is used for carrying out readable form representation on the decision result obtained by the calculation module 4 and providing the decision result as decision support for a decision maker; the method comprises the following specific steps:
(1) an initialization module: problem parameter setting
The method comprises the steps that a pharmacy network of a certain area is represented as a network consisting of | P | physical pharmacies dispersed in the area, | M | medicines are provided for the area, medicine inventory of each physical pharmacy is partially overlapped but not completely identical, and each medicine has corresponding temperature control transportation cost due to strict requirements of medicine transportation conditions; in addition, the vehicles with the same quality are distributed at different positions in a road network of the region, the vehicles have a volume limit V, and the vehicles have time parameters and cost parameters in the road network;
(2) information input module
Inputting data of | O | order demands of a certain batch, wherein the data mainly comprises order positions, order times and medicine demands in orders;
(3) the calculation module 1: problem modeling and problem decomposition
Step 3.1: problem modeling
Describing the collaborative fulfillment problem of a full-channel medicine retail order based on an initialization module and an information input module, and establishing a mixed integer programming mathematical model TD-UMPD-VRPT, wherein main decisions comprise a vehicle dispatching decision U, a medicine distribution decision Y in the order, a commodity flow decision Z and a vehicle flow decision X;
step 3.2: problem resolution
Decomposing the original problem model according to entity network data such as pharmacy distribution, pharmacy inventory data and the like and order demand information to obtain two decomposed submodels, namely a model A and a model B, and adding effective constraints to the submodels through corresponding coupling constraints and parameters to facilitate subsequent iteration;
(4) the calculation module 2: solving model A and corresponding iteration parameters
Step 4.1: solution of model A
Embedding Dijkstra algorithm in the Hungarian algorithm to obtain an improved Hugarian-Dijkstra algorithm MHDA solving model A, namely, solving the shortest path problem of multiple vehicles by utilizing the Dijkstra algorithm to obtain a cost matrix corresponding to multiple commodity flows of the medicines, solving the medicine distribution problem in an order by utilizing the Hungarian algorithm according to the cost matrix to obtain an order distribution result comprising the network flows of the multiple commodities;
step 4.2: obtaining an iteration parameter
Obtaining a distribution scheme of the medicines in the order and distribution information of the order and the vehicle according to the solving result of the model A, and taking the distribution scheme and the distribution information as input and iteration parameters of the next step;
(5) the calculation module 3: solving model B and corresponding iteration parameters
Step 5.1: solution of model B
Based on the distribution scheme obtained in the step 4.2, obtaining an initial path scheme by adopting a greedy search algorithm GSA, improving the initial path scheme through neighborhood search, further obtaining a solving result of the model B, and obtaining a complete order fulfillment scheme;
step 5.2: solution of model B
Obtaining new order and vehicle distribution information as iteration parameters according to the solving result of the model B;
(6) iterative coordination module
Judging whether the order obtained in the step 4.2 and the step 5.2 is matched with the distribution information of the vehicle, if so, obtaining a final iteration order fulfillment scheme, and carrying out the next step; otherwise, feeding back the order and vehicle distribution information of the step 5.2 to the step 4, and re-executing the iterative program;
(7) the calculation module 4: search improvement
Performing further search optimization on the final iteration order fulfillment scheme obtained in the step 6 by adopting a neighborhood structural design variable neighborhood search algorithm;
(8) decision output module
And (4) converting the results of the vehicle dispatch decision U, the traditional Chinese medicine distribution decision Y, the commodity flow decision Z and the vehicle flow decision X obtained in the step (7) into a readable form for representation, and providing the readable form as decision support for a decision maker.
2. The method of claim 1, wherein the objective function of the TD-UMPD-VRPT model is expressed by formula (1):
wherein M is a drug set, M o Representing a drug set in order o, K being a vehicle set, P being a physical pharmacy set, N P Representing a set of pharmacy locations, O being a set of orders, N O Indicating a set of order locations, N being the set of all points in the area, N ═ N P ∪N O Where i ∈ N uses plane coordinates (i ∈ N) a ，i b ) Represents; α is the fixed use cost of the vehicle, β ij Is the cost of the vehicle's path from point i to point j, θ ij Is the vehicle's path time from point i to point j; gamma ray m Is the temperature-controlled transportation cost of the medicine m in unit time; e.g. of the type k Is the current position of the vehicle k, e' is the virtual end point, and the distances from all points to the virtual end point are zero; min is a minimum function, decision variable u of type 0-1 k Means for indicating whether to use decision variable x of type k, 0-1 for vehicle kij Decision variables of type 0-1, representing whether vehicle k has traversed the path from point i to point j
Indicating whether the vehicle k carries the medicine m in the order o through a path from the point i to a point j;
the constraints of the TD-UMPD-VRPT model are expressed by equations (2) - (20):
(a) order splitting and allocation decision-related constraints
(b) Coupled constraints for order distribution and multi-commodity flow decision
(c) Multi-commodity flow decision-related constraints
(d) Coupling constraints for multiple commodity flow and vehicle path decisions
(e) Taking and delivering vehicle path decision related constraints
Wherein, decision variables of type 0-1
Indicating that the drug m in the order o is stocked by the pharmacy p, decision variable b type 0-1 kij Is an auxiliary variable indicating whether node i is before node j in the path of vehicle k; l is a real number not less than 10^ 6; continuous decision variable t ki And
respectively representing the time of arrival and departure of the vehicle k at the point i; stock status s pm Indicating the stock of the drugs m held by the pharmacy p, and belongs to N P ，m∈M；v m Represents the volume of the drug m, V is the volume of the vehicle; d o Is the delivery deadline for order O, O ∈ O;
decomposing the model into two subproblems, model A and model B, by removing equation (10);
model A:
constraint conditions are as follows: including formula (2), formula (3), formula (8) and
wherein, K is O Is an order-vehicle matching information set, k o Refers to the vehicle responsible for order o; by gathering K order-vehicle matching information O For incorporation into model A, partial constraints in the original model are updated to new constraints by aggregating K the order-vehicle matching information O Including in model A, the partial constraints in the original model are updated to new constraint equations (22) - (26); during initial iteration, setting default that all orders are distributed to a certain vehicle;
model B:
constraint conditions are as follows: formulae (11) to (20) and
wherein N is + Is the set of physical pharmacies for which assignment is made, and formula (28) indicates that the physical pharmacy for which assignment is made is accessed at least once, and that formula is to beThe solving result of the model A is used as the input of the model B; vehicle flow decision x for model B kij Corresponding to corresponding commodity flow variables
Therefore, solving the model B can obtain new order-vehicle matching pair information K O This information is fed back into model a for a second iteration.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210503610.8A CN114912976A (en) | 2022-05-09 | 2022-05-09 | Collaborative fulfillment optimization method for full-channel drug retail orders |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210503610.8A CN114912976A (en) | 2022-05-09 | 2022-05-09 | Collaborative fulfillment optimization method for full-channel drug retail orders |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114912976A true CN114912976A (en) | 2022-08-16 |
Family
ID=82766247
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210503610.8A Pending CN114912976A (en) | 2022-05-09 | 2022-05-09 | Collaborative fulfillment optimization method for full-channel drug retail orders |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114912976A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116362844A (en) * | 2023-04-03 | 2023-06-30 | 大连理工大学 | "first merging list-then dispatching list" takeout order distribution scheme generation method |
-
2022
- 2022-05-09 CN CN202210503610.8A patent/CN114912976A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116362844A (en) * | 2023-04-03 | 2023-06-30 | 大连理工大学 | "first merging list-then dispatching list" takeout order distribution scheme generation method |
| CN116362844B (en) * | 2023-04-03 | 2023-11-03 | 大连理工大学 | "first merging list-then dispatching list" takeout order distribution scheme generation method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Leung et al. | A B2C e-commerce intelligent system for re-engineering the e-order fulfilment process | |
| US11580860B2 (en) | Optimizing task assignments in a delivery system | |
| Masoud et al. | A decomposition algorithm to solve the multi-hop peer-to-peer ride-matching problem | |
| Masoud et al. | Autonomous or driver-less vehicles: Implementation strategies and operational concerns | |
| Torabi et al. | Fulfillment source allocation, inventory transshipment, and customer order transfer in e-tailing | |
| Abdirad et al. | A two-stage metaheuristic algorithm for the dynamic vehicle routing problem in Industry 4.0 approach | |
| Zangwill | A backlogging model and a multi-echelon model of a dynamic economic lot size production system—a network approach | |
| Peng et al. | Stable ride-sharing matching for the commuters with payment design | |
| Ma et al. | Real-time city-scale taxi ridesharing | |
| Federgruen et al. | A simple forward algorithm to solve general dynamic lot sizing models with n periods in 0 (n log n) or 0 (n) time | |
| CN1628285A (en) | Agent, method and computer system for negotiating in a virtual environnement | |
| CN109902990A (en) | A kind of supply chain logistics distribution routing planning system | |
| Paul et al. | Optimizing omni-channel fulfillment with store transfers | |
| Lee et al. | Development of an algorithm for optimal demand responsive relocatable feeder transit networks serving multiple trains and stations | |
| Li et al. | A Benders decomposition algorithm for the order fulfilment problem of an e-tailer with a self-owned logistics system | |
| Hezarkhani et al. | Collaborative replenishment in the presence of intermediaries | |
| Jiang et al. | Spatial and temporal optimization for smart warehouses with fast turnover | |
| CN114912976A (en) | Collaborative fulfillment optimization method for full-channel drug retail orders | |
| Cota et al. | Integrating vehicle scheduling and open routing decisions in a cross-docking center with multiple docks | |
| Gómez-Lagos et al. | On a pickup to delivery drone routing problem: Models and algorithms | |
| Wang et al. | Joint optimization of parcel allocation and crowd routing for crowdsourced last-mile delivery | |
| Farvaresh et al. | A coalition structure algorithm for large-scale collaborative pickup and delivery problem | |
| Salehi Sarbijan et al. | Emerging research fields in vehicle routing problem: a short review | |
| US20040093101A1 (en) | Production control system, production control method and production control program | |
| Akyüz et al. | Minimum cost delivery of multi-item orders in e-commerce logistics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |