Multi-objective optimization method and system for emergency planning of medical supplies based on demand forecasting

Abstract

The present invention provides a multi-objective optimization method and system for emergency plan of medical supplies based on demand forecasting, and relates to the technical field of task optimization. The present invention performs multi-objective optimization on the medical supplies demand plan predicted based on the gray model, taking into account the inconsistent capabilities of suppliers, comprehensively considering the minimum and maximum response time and minimum cost of suppliers, solving the problem through a hybrid intelligent algorithm and obtaining an approximate optimal solution, which can effectively coordinate and optimize and design the order quantity of multiple cycles, thereby maximizing the use of supplier resources and reducing the overall cost of the emergency supply chain, and improving the management capability and operation efficiency of the emergency medical supplies supply chain.

Description

Technical Field

[0001] The present invention relates to the technical field of task optimization, and in particular to a multi-objective optimization method and system for a medical supplies emergency plan based on demand forecasting. Background Art

[0002] Emergency medical supply planning is a crucial tool for responding to emergencies. Especially when faced with public health emergencies, countries' emergency medical supply management systems face severe challenges. Emergency medical supply planning typically uses emergency response time as its objective function, taking into account constraints such as cargo type, quantity, and vehicle transport capacity. It utilizes various medical supply stocking strategies and routing optimization theory to ensure timely and adequate delivery of emergency medical supplies to different demand points.

[0003] The existing optimization of emergency medical supplies plans pays more attention to advance stocking strategies based on emergency demand forecasts and medical supplies transportation plans based on ordering strategies. The emphasis is on the classified stocking of medical supplies, the reasonable allocation of existing medical supplies, and the combined transportation of medical supplies. There is rarely a combination purchase of medical supplies from the supplier's perspective, and there is also less attention paid to the process based on the combination of medical supplies demand forecasts and supplier selection. From the perspective of emergency supply chain management, the current research on the optimization of emergency medical supplies plans cannot well guide the actual emergency medical supplies management, which in turn leads to insufficient resource utilization. Summary of the Invention

[0004] (1) Technical problems solved

[0005] In response to the shortcomings of the existing technology, the present invention provides a multi-objective optimization method and system for medical supplies emergency plan based on demand forecasting, which solves the technical problem of optimizing medical supplies emergency plan without considering inconsistent supplier capabilities.

[0006] (2) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0008] A multi-objective optimization method for emergency planning of medical supplies based on demand forecasting, including:

[0009] S1. Based on the historical data of medical supply demand over the same period, predict the demand for medical supplies in each future period;

[0010] S2. Setting input parameters for a hybrid non-dominated sorting genetic algorithm based on neighborhood search based on the demand for medical supplies, combined with supplier supply capacity data, supplier supply type data, and demand point warehousing data;

[0011] S3. According to the input parameters, based on the hybrid non-dominated sorting genetic algorithm, a global optimal solution is obtained, and the set and quantity of medical supplies ordered from suppliers in each future period, as well as the outsourced amount of medical supplies, are decoded and obtained.

[0012] Preferably, the medical supplies forecasting model based on grey theory in S1 predicts the demand for medical supplies in each future period.

[0013] Preferably, the input parameters in S2 include:

[0014] Define the period index i = 1, 2, ..., l; supplier index j = 1, 2, ..., m; medical supplies index k = 1, 2, ..., n;

[0015] The type of medical supplies supplied by the supplier {Q 11 ,…,Q jk ,…,Q mn}, where Q jk indicates that supplier j supplies medical supplies k;

[0016] Initial medical supplies storage at the demand point {SQ 11 ,…,SQ ik ,…,SQ ln}, where SQ ik represents the inventory of medical supplies k in period i;

[0017] Supplier production capacity set {C 11 ,C 12 ,…,C 1k ,…,C jk ,…,C mn}, where C jk represents the ability of supplier j to supply medical supplies k;

[0018] Supplier medical supplies purchase price collection {[c i11 ,…,c i1k ,…,c i1n ],……,[c ij1 ,…,c ijk ,…,c ijn ],……,[c im1 ,…,c imk ,…,c imn ]}, where c ijk represents the purchase price of medical supplies k supplied by supplier j in period i;

[0019] Medical supplies outsourcing unit price collection in represents the unit price of medical supplies outsourcing for medical supplies k in period i;

[0020] Medical supplies storage unit price collection in represents the storage unit price of medical supplies k in period i;

[0021] The demand for medical supplies in each period is {[n 11 ,…,n 1k ,…,n 1n ],……,[n i1 ,…,n ik ,…,n in ],……,[n l1 ,…,n 1k ,…,n ln ]}, where n ik represents the demand for medical supplies k in period i.

[0022] Preferably, the S3 includes:

[0023] S31. Set the execution parameters of the non-dominated sorting hybrid genetic algorithm based on neighborhood search, including: the maximum number of iterations Iter1, set iter1 = 1; the population size N of the algorithm; the mutation probability P v ; The maximum number of iterations of local search Iter2; The initial cycle selection probability PT i =1 / T; mutation probability P v ; Neighborhood set Each neighborhood structure NS l The initial selection probability P l =1 / l max ;

[0024] S32. Generate an initial population based on the input parameters. The individuals in the population are represented by two solutions, SN-MT-OQ and MT-EC. Decode the solutions X1 and X2 to obtain the non-dominated order and crowding value of the individuals. Record the current optimal solution as the global optimal solution τ. best ;

[0025] The numbers in SN in the solution X1 represent the supplier number, the numbers in MT represent the medical supplies category number, and OQ represents the order quantity of medical supplies. The length is Solution: In X2, MT represents the medical material category number and EC represents the outsourced medical material quantity, with a length of n*l.

[0026] S33, performing a crossover operation on individuals in the current population based on a crossover operator to generate a population of the same size, and merging the population with the current population to obtain a new population;

[0027] S34, based on the non-dominated sorting and crowding values ​​of each individual in the merged new population, a selection operation is performed using the minimum dominance and maximum crowding strategies to obtain a new population with the same population size;

[0028] S35. Decode and sort the individuals in the current population, calculate the crowding value of each individual, mark the individual with the largest crowding value in the same non-dominated level except the first and last as τ, and select a neighborhood structure NS based on the roulette wheel method. l , generate a neighborhood solution τ′ according to the neighborhood structure;

[0029] S36, using neighborhood structure NS l Perform a local search on the solution τ′ to obtain the local optimal solution τ″. Compare the congestion of τ and τ″. If τ″ is better than τ, set τ = τ″. At the same time, record the number of successful searches for the local optimal solution under this neighborhood structure. Otherwise, record the number of failed searches for the local optimal solution under this neighborhood structure and update the selection probability of the neighborhood structure.

[0030] S37, compare the optimal solution τ under the current neighborhood with the worst individual in the previous generation population. If the optimal solution τ is better than the worst solution of the previous generation population, then use the optimal solution τ to replace the worst solution, and compare the optimal solution τ under the current neighborhood with the global optimal solution τ best Compare, if τ is better than τ best , then let τ best =τ;

[0031] S38, determine whether iter1≤Iter1 is established. If so, set iter1=iter1+1, and use the updated population as the input solution for the next iteration and return to step S32. Otherwise, output the global optimal solution τ best , and decode to obtain the set and quantity of medical supplies ordered from suppliers in each future period, as well as the amount of medical supplies outsourced.

[0032] Preferably, the decoding process in S32 is as follows:

[0033] S3202a, generate the quantity OQ of medical supplies k ordered from supplier j in each period based on solution X1. ijk ; Use the following formula to calculate the cost of ordering medical supplies: Procurement cost Of which: OQ ijk Indicates the quantity of medical supplies ordered from suppliers during the period;

[0034] S3202b, calculate the medical supplies storage quantity SQ of the demand point in each cycle based on the solution X1 and the initial medical supplies storage quantity ik , where SQ i+1,k =SQ ik +OQ i+1,k-n ik ; Use the following formula to calculate the cost of storing medical supplies:

[0035] S3202c, generate the outsourcing quantity EC of medical supplies k in each period from solution X2. ik ; Use the following formula to calculate the cost of outsourcing medical supplies:

[0036] S3202d. Calculate the total cost and the maximum response time within the cycle based on solutions X1 and X2 using the following formula: in Indicates the time of outsourcing of medical supplies within the cycle.

[0037] Preferably, the S33 includes:

[0038] S3301, decoding to obtain the minimum and maximum response time and total cost of each individual in the population, respectively recorded as f1(x)=C max and f2(x)=TC;

[0039] S3302: Based on the crossover operator of the hybrid optimization algorithm, a new population X1 is generated by operating the current population X. The number of the new population X1 is N. The initial population X and the new population X1 are merged. The number of the merged population is 2N. The main operations include:

[0040] S3302a, select individual X from the current population s and individual X t , randomly select a cycle in the interval (1, T) according to the roulette strategy, perform the crossover operation in the cycle, and record the number of selections CT in the cycle i ;

[0041] S3302b, in this period, in the interval Randomly generate two integers Cut1 and Cut2 as intersection points, where New individual X′ s Copy individual X t The SN-MT-OQ is located between Cut1 and Cut2, and the remaining codes are reserved for individual X. s The encoding in; the new individual X′ t Copy individual X s The SN-MT-OQ is located between Cut1 and Cut2, and the remaining codes are reserved for individual X. t The encoding in;

[0042] S3302c, compare individual X s and X′ s 、X t and X′t , if X′ s Better than X s or X′ t Better than X t If there is at least one, then record the number of successful ST cycles i =ST i +1, and calculate the success rate of this cycle according to the formula PS i , formula such as: PS i =ST i / CT i ;

[0043] S3302d, determine whether the crossover number d is greater than popsize / 2, if so, terminate the crossover operation and update the cycle selection probability Otherwise, go to S3302a.

[0044] Preferably, the S34 includes:

[0045] S3401. Define the non-dominated sorting rule: for individual x s and individual x t , for any s=1,2,…,N, f1(x s )≤f1(x t ), and there exists at least one f1(x s ) <f1(x t ) holds true, then the individual x s Dominant individual x t , and set s=1;

[0046] S3402. For all t=1, 2, ..., N and t≠s, compare the individual x according to the definition. s and individual x t The non-dominance relationship between them; if there is no individual x t Better than individual x s , then label the individual x s is a non-dominated individual and is added to the set F1;

[0047] S3403, let s = s + 1, go to S3402, until all non-dominated individuals are found, go to S3404;

[0048] S3404. Calculate the difference between population X and set F1 and use it as the new set X′. Go to S3402 and repeat the non-dominated selection on the individuals in set X′ until all individuals in the population are classified according to the non-dominated relationship. Go to S3405.

[0049] S3405. Calculate the crowding degree C for individuals in the same non-dominated level. r , the formula includes:

[0050]

[0051]

[0052] Among them, m represents the objective function index;

[0053] C r [s] m represents the crowding degree of individual x in the m-th objective function s ;

[0054] x max represents the maximum value of individual x under the m function; s

[0055] x min represents the minimum value of individual x under the m function; s

[0056] S3406. Select levels in ascending order of non-dominated levels, and select individuals in descending order of crowding degree within the same level as individuals of the next generation population until N population individuals are selected.

[0057] Preferably, the non-dominated sorting hybrid genetic algorithm designs 6 neighborhood structures:

[0058] Neighborhood structure 1: Randomly select a period, define variables a and b, randomly generate two integers in the interval (1, m) and assign them to a and b, where a < b, and exchange the codes between positions a and b;

[0059] Neighborhood structure 2: Randomly select a period, define variables a and b, randomly generate two integers in the interval (1, m) and assign them to a and b, where a < b, and reverse the codes between positions a and b in the sequence SN - OQ;

[0060] Neighborhood structure 3: Randomly select a period, define variables a and b, randomly generate two integers in the interval (1, m) and assign them to a and b, where a < b, and exchange the codes to the left of position a and to the right of position b in the sequence SN - OQ;

[0061] Neighborhood structure 4: Randomly select a supplier in the sequence SN - OQ, define variables a and b, randomly generate two integers in the interval (1, T) and assign them to a and b, where a < b, and exchange the codes between positions a and b;

[0062] ​​Neighborhood structure 5: Randomly select a supplier in the sequence SN-OQ, define variables a and b, randomly generate two integers within the interval (1, T) and assign them to a and b, where a < b, and reverse the encoding between positions a and b;

[0063] Neighborhood structure 6: Randomly select a supplier in the sequence SN-OQ, define variables a and b, randomly generate two integers within the interval (1, T) and assign them to a and b, where a < b, and swap the encoding to the left of position a and to the right of position b.

[0064] Preferably, the S36 includes:

[0065] S3601. For the neighborhood structure NS l , during the operation of performing variable neighborhood local search, record the successful search times S l and the failure times F l of the local optimal solutions within each neighborhood structure, and perform an update operation on the neighborhood selection probability according to the successful search times. The formula is as follows: where SR l represents the local search success rate of the neighborhood structure NS l , SR l = S l / (S l + F l ) + ε, where ε is a sufficiently small positive number;

[0066] S3602. Define a variable V as the neighborhood selection variable, where V is a number randomly generated within the interval (0, 1). Judge the range of V. If it satisfies SR l-1 < V < SR l , then select the neighborhood structure NS l to perform the neighborhood search operation.

[0067] A multi-objective optimization system for medical material emergency plans based on demand forecasting includes:

[0068] A prediction module for predicting the demand for medical materials in each future period according to the historical data of the same period of medical material demand;

[0069] A setting module for setting the input parameters of the hybrid non-dominated sorting genetic algorithm based on neighborhood search according to the demand for medical materials, combined with the supplier supply capacity data, the data of the types of medical materials supplied by the supplier, and the demand point storage data;

[0070] A solving module for obtaining the global optimal solution based on the hybrid non-dominated sorting genetic algorithm according to the input parameters, and decoding to obtain the set of medical materials ordered from suppliers, the quantity, and the amount of medical materials outsourced in each future period

[0071] (3) Beneficial effects

[0072] The present invention provides a multi-objective optimization method and system for emergency medical supplies planning based on demand forecasting. Compared with existing technologies, it has the following advantages:

[0073] The present invention recognizes that there are differences in the medical supplies supply capabilities and costs of suppliers. That is, taking into account the inconsistency of supplier capabilities, the predicted medical supplies demand plan is optimized with multiple objectives, and the capability indicators of different suppliers are comprehensively considered. The problem is solved and an approximate optimal solution is obtained through a hybrid non-dominated sorting genetic algorithm based on neighborhood search. This can effectively coordinate the optimization and design of order quantities for multiple cycles, thereby maximizing the use of supplier resources and reducing the overall cost of the emergency supply chain, thereby improving the management capability and operational efficiency of the emergency medical supplies supply chain. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0075] Figure 1 A flowchart of a multi-objective optimization method for emergency medical supplies planning based on demand forecasting provided by an embodiment of the present invention;

[0076] Figure 2 A flowchart of a hybrid non-dominated sorting genetic algorithm based on neighborhood search provided by an embodiment of the present invention;

[0077] Figure 3 A coding schematic diagram provided by an embodiment of the present invention;

[0078] Figure 4 A decoding schematic diagram provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0079] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0080] The embodiments of the present application solve the technical problem of optimizing medical supplies emergency plans without considering inconsistent supplier capabilities by providing a multi-objective optimization method and system for medical supplies emergency plans based on demand forecasting.

[0081] The technical solution in the embodiments of the present application is to solve the above technical problems, and the overall idea is as follows:

[0082] As mentioned in the background technology, the existing emergency medical supplies planning optimization focuses on the classification and stocking of medical supplies, the rational allocation of existing medical supplies, and the combined transportation of medical supplies. There is rarely any combination purchase of medical supplies from the supplier's perspective, and there is also little attention paid to the process based on the combination of medical supplies demand forecasting and supplier selection.

[0083] The embodiment of the present invention performs multi-objective optimization on the medical supplies demand plan predicted based on the gray model while taking into account the inconsistency of supplier capabilities, comprehensively considers the minimum and maximum response time and minimum cost of suppliers, solves the problem through a hybrid intelligent algorithm and obtains an approximate optimal solution, which can effectively coordinate and optimize and design the order quantity of multiple cycles, thereby maximizing the use of supplier resources and reducing the overall cost of the emergency supply chain, thereby improving the management capability and operational efficiency of the emergency medical supplies supply chain.

[0084] In order to better understand the above technical solution, the above technical solution will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0085] Example:

[0086] like Figure 1 As shown, an embodiment of the present invention provides a multi-objective optimization method for medical supplies emergency plan based on demand forecasting, including:

[0087] S1. Based on the historical data of medical supply demand over the same period, predict the demand for medical supplies in each future period;

[0088] S2. Setting input parameters for a hybrid non-dominated sorting genetic algorithm based on neighborhood search based on the demand for medical supplies, combined with supplier supply capacity data, supplier supply type data, and demand point warehousing data;

[0089] S3. According to the input parameters, based on the hybrid non-dominated sorting genetic algorithm, a global optimal solution is obtained, and the set and quantity of medical supplies ordered from suppliers in each future period, as well as the outsourced amount of medical supplies, are decoded and obtained.

[0090] The embodiment of the present invention performs multi-objective optimization on the predicted medical supply demand plan while taking into account the inconsistent capabilities of suppliers. It comprehensively considers the capability indicators of different suppliers, solves the problem through a hybrid intelligent algorithm and obtains an approximate optimal solution. It can effectively coordinate and optimize the order quantity of multiple cycles, thereby maximizing the use of supplier resources and reducing the overall cost of the emergency supply chain, thereby improving the management capability and operational efficiency of the emergency medical supply supply chain.

[0091] The following sections describe the various steps of the above solution in detail:

[0092] In step S1, the demand for medical supplies in each future period is predicted based on the historical data of medical supply demand in the same period.

[0093] This step uses the medical supplies forecasting model based on grey theory to predict the demand for medical supplies in each future period, as follows:

[0094] Obtain historical data on medical supplies for the same period in It represents the actual demand for medical supplies k in the same historical period i.

[0095] Add the data in sequence to form a new list Where d represents the accumulated quantity index, N 0 Represents the number of cycles of historical demand, using the least squares method Among them, a and u are the parameters of the least squares method. Further solving the above formula yields in represents the demand quantity of emergency medical supplies k in the future period i+1, which corresponds to the demand quantity of medical supplies n in the subsequent step S2 1k .

[0096] In step S2, according to the demand for medical supplies, combined with supplier supply capacity data, supplier supply medical supplies type data and demand point storage data, the input parameters of the hybrid non-dominated sorting genetic algorithm based on neighborhood search are set.

[0097] For example, historical data analysis, manual data entry, and other methods can be used to obtain supplier supply capacity data, supplier supply type data, and the number of warehouses at demand points in different future periods. Based on the obtained data, input parameters of the variable neighborhood search algorithm and the non-dominated sorting genetic algorithm can be set; specifically, the following can be included:

[0098] Define the period index i = 1, 2, ..., l; supplier index j = 1, 2, ..., m; medical supplies index k = 1, 2, ..., n;

[0099] The types of medical supplies supplied by suppliers {Q 11,…,Q jk ,…,Q mn}, where Q jk indicates that supplier j supplies medical supplies k;

[0100] Initial medical supplies storage at the demand point {SQ 11 ,…,SQ ik ,…,SQ ln}, where SQ ik represents the inventory of medical supplies k in period i;

[0101] Supplier production capacity set {C 11 ,C 12 ,…,C 1k ,…,C jk ,…,C mn}, where C jk represents the ability of supplier j to supply medical supplies k;

[0102] Supplier medical supplies purchase price collection {[c i11 ,…,c i1k ,…,c i1n ],……,[c ij1 ,…,c ijk ,…,c ijn ],……,[c im1 ,…,c imk ,…,c imn ]}, where c ijk represents the purchase price of medical supplies k supplied by supplier j in period i;

[0103] Medical supplies outsourcing unit price collection in represents the unit price of medical supplies outsourcing for medical supplies k in period i;

[0104] Medical supplies storage unit price collection in represents the storage unit price of medical supplies k in period i;

[0105] The demand for medical supplies in each period is {[n 11 ,…,n 1k ,…,n 1n ],……,[n i1 ,…,n ik ,…,n in ],……,[n l1 ,…,n 1k ,…,n ln ]}, where n ikrepresents the demand for medical supplies k in period i.

[0106] In step S3, according to the input parameters, based on the hybrid non-dominated sorting genetic algorithm, a global optimal solution is obtained, and the set and quantity of medical supplies ordered from suppliers in each future period and the amount of medical supplies outsourced are decoded and obtained.

[0107] like Figure 2 As shown, this step specifically includes:

[0108] S31, setting execution parameters of the non-dominated sorting hybrid genetic algorithm based on neighborhood search, including:

[0109] Maximum number of iterations Iter1;

[0110] The population size N of the algorithm;

[0111] Mutation probability P v ;

[0112] The maximum number of iterations of local search Iter2;

[0113] Initial cycle selection probability PT i =1 / T;

[0114] Mutation probability P v ;

[0115] Neighborhood set {NS1,…,NS l ,…,NS lmax Each neighborhood structure NS l The initial selection probability P l =1 / l max , illustrative example of the present invention designs 6 neighborhood structures, see below for details.

[0116] S32. Generate an initial population based on the input parameters. The individuals in the population are represented by two solutions, SN-MT-OQ and MT-EC. Decode the solutions X1 and X2 to obtain the non-dominated order and crowding value of the individuals. Record the current optimal solution as the global optimal solution τ. best .

[0117] Based on the hybrid algorithm of non-dominated genetic algorithm and variable neighborhood search, the order quantity in the emergency medical supplies reserve plan is encoded to obtain an initial population X0 consisting of N initial solutions. Each solution in X0 is composed of two sequences, SN-MT-OQ and MT-EC. SN-MT-OQ represents the quantity of medical supplies ordered from suppliers within the period, and MT-EC represents the quantity of medical supplies outsourced within the period. The initial population is decoded and the minimum maximum response time C is used to select the optimal solution. maxPerform non-dominated sorting on the individuals in the population for the main goal and calculate the crowding value C of the individuals in the population r , which is calculated from a function value of adjacent individuals and is used to sort individuals at the same non-dominated level; the best individual among them is taken as the global optimal solution τ under the current conditions best , set the initial iteration number of the hybrid intelligent optimization algorithm iter1 = 1, and set the initial iteration number of the variable neighborhood search iter2 = 1.

[0118] like Figure 3 As shown, the encoding process is as follows:

[0119] Based on the deterministic demand for emergency medical supplies within a period, the solution X consists of the solution X1 of the order quantity within each period and the solution X2 of the medical supplies outsourcing quantity within the period.

[0120] Solution X1 is represented by a three-dimensional integer array. Solution X1 consists of three parts: supplier list SN, medical material category MT, and order quantity OQ within the period. The length of the sequence or list is The numbers in SN represent the supplier's number, the numbers in MT represent the number of the medical supplies category, and OQ represents the order quantity of medical supplies; the list SN-MT represents the categories of medical supplies that the supplier can supply, and the list SN-MT-OQ represents the quantity of medical supplies ordered from the supplier during the period.

[0121] Solution X2 is represented by a two-dimensional integer array consisting of the medical material category code MT and the medical material outsourcing quantity EC. The length of the sequence or list is n*T, and each column in the array represents the outsourcing quantity of each medical material in the period. In particular, the quantity in EC is calculated according to the following formula: Based on the medical material demand and the local storage quantity of medical materials in each period, combined with the order quantity in the period, calculate the outsourcing quantity EC of medical material k in period i. ik =n ik -SQ ik -OQ ik , OQ ik represents the order quantity of medical supplies k in period i.

[0122] like Figure 4 As shown, the decoding process is as follows:

[0123] The decoding process is to calculate the total cost and minimum and maximum response time of medical supplies emergency response in all cycles according to the actual scenario based on the obtained feasible solution. It mainly includes the following operations:

[0124] S3202a, generate the quantity OQ of medical supplies k ordered from supplier j in each period based on solution X1. ijk ; Use the following formula to calculate the cost of ordering medical supplies: Procurement cost Of which: OQijk Indicates the quantity of medical supplies ordered from suppliers during the period;

[0125] S3202b, calculate the medical supplies storage quantity SQ of the demand point in each cycle based on the solution X1 and the initial medical supplies storage quantity ik , where SQ i+1,k =SQ ik +OQ i+1,k -n ik ; Use the following formula to calculate the cost of storing medical supplies:

[0126] S3202c, generate the outsourcing quantity EC of medical supplies k in each period from solution X2. ik ; Use the following formula to calculate the cost of outsourcing medical supplies:

[0127] S3202d. Calculate the total cost and the maximum response time within the cycle based on solutions X1 and X2 using the following formula: in Indicates the time of outsourcing of medical supplies within the cycle.

[0128] S33, performing a crossover operation on individuals in the current population based on a crossover operator to generate a population of the same size, and merging the population with the current population to obtain a new population; including:

[0129] S3301, decoding to obtain the minimum and maximum response time and total cost of each individual in the population, respectively recorded as f1(x)=C max and f2(x)=TC;

[0130] S3302: Based on the crossover operator of the hybrid optimization algorithm, a new population X1 is generated by operating the current population X. The number of the new population X1 is N. The initial population X and the new population X1 are merged. The number of the merged population is 2N. The main operations include:

[0131] S3302a, select individual X from the current population s and individual X t , randomly select a cycle in the interval (1, T) according to the roulette strategy, perform the crossover operation in the cycle, and record the number of selections CT in the cycle i ;

[0132] S3302b, in this period, in the interval Randomly generate two integers Cut1 and Cut2 as intersection points, where New individual X′ s Copy individual X tThe SN-MT-OQ is located between Cut1 and Cut2, and the remaining codes are reserved for individual X. s The encoding in; the new individual X′ t Copy individual X s The SN-MT-OQ is located between Cut1 and Cut2, and the remaining codes are reserved for individual X. t The encoding in;

[0133] S3302c, compare individual X s and X′ s 、X t and X′ t , if X′ s Better than X s or X′ t Better than X t If there is at least one, then record the number of successful ST cycles i =ST i +1, and calculate the success rate of this cycle according to the formula PS i , formula such as: PS i =ST i / CT i ;

[0134] S3302d, determine whether the crossover number d is greater than popsize / 2, if so, terminate the crossover operation and update the cycle selection probability Otherwise, go to S3302a.

[0135] S34. Based on the non-dominated sorting and crowding values ​​of each individual in the merged new population, a selection operation is performed using the minimum dominance and maximum crowding strategies to obtain a new population with the same population size; including:

[0136] S3401. Define the non-dominated sorting rule: for individual x s and individual x t , for any s=1,2,…,N, f1(x s )≤f1(x t ), and there exists at least one f1(x s ) <f1(x t ) holds true, then the individual x s Dominant individual x t , and set s=1;

[0137] S3402. For all t=1, 2, ..., N and t≠s, compare the individual x according to the definition. s and individual x t The non-dominance relationship between them; if there is no individual x t Better than individual x s , then label the individual xs is a non-dominated individual and is added to the set F1;

[0138] S3403, let s = s + 1, go to S3402, until all non-dominated individuals are found, go to S3404;

[0139] S3404. Calculate the difference between population X and set F1 and use it as the new set X′. Go to S3402 and repeat the non-dominated selection on the individuals in set X′ until all individuals in the population are classified according to the non-dominated relationship. Go to S3405.

[0140] S3405. Calculate the crowding degree C for individuals in the same non-dominated level. r , the formula includes:

[0141]

[0142]

[0143] Where m represents the objective function index;

[0144] C r [s] m Represents the individual x in the mth objective function s the degree of congestion;

[0145] x max Represents individual x s The maximum value under the m function;

[0146] x min Represents individual x s The minimum value under the m function;

[0147] S3406. Select levels in descending order of non-dominated levels, and select individuals in the same level in descending order of crowding as individuals of the next generation population, guiding the selection of N population individuals.

[0148] S35. Decode and sort the individuals in the current population, calculate the crowding value of each individual, mark the individual with the largest crowding value in the same non-dominated level except the first and last as τ, and select a neighborhood structure NS based on the roulette wheel method. l , a neighborhood solution τ′ is generated according to the neighborhood structure; including:

[0149] S3501. Decode the individuals in the current population to obtain the total cost TC and maximum response time C of each individual. max The best individual is denoted as τ. For the specific decoding process, refer to S3202;

[0150] S3502. Randomly select a supplier, extract the medical supplies quantity of the supplier in different periods from the list SN-MT-OQ according to the medical supply category, and form a sequence SN-OQ. For example, Figure 4 as shown, perform the neighborhood selection operation;

[0151] S3503. Select a neighborhood structure NS l based on the roulette wheel method, and generate a neighborhood solution τ′ of τ according to this neighborhood structure NS l . Specifically, it includes:

[0152] Perform operations on the feasible solution τ within the selected neighborhood structure NS l so that the initial solution at this stage can jump to another solution in the feasible region and avoid the algorithm falling into local optimality. Considering the complexity and constraint requirements of inconsistent emergency medical supply suppliers, to ensure that the solutions generated by each operation change are all feasible solutions, specifically design the following 6 neighborhood structures

[0153] Neighborhood structure 1: Randomly select a period, define variables a and b, randomly generate two integers in the interval (1, m) and assign them to a and b, where a < b, and exchange the codes between positions a and b;

[0154] Neighborhood structure 2: Randomly select a period, define variables a and b, randomly generate two integers in the interval (1, m) and assign them to a and b, where a < b, and reverse the order of the codes between positions a and b in the sequence SN-OQ;

[0155] Neighborhood structure 3: Randomly select a period, define variables a and b, randomly generate two integers in the interval (1, m) and assign them to a and b, where a < b, and exchange the codes on the left side of position a and on the right side of position b in the sequence SN-OQ;

[0156] Neighborhood structure 4: Randomly select a supplier in the sequence SN-OQ, define variables a and b, randomly generate two integers in the interval (1, T) and assign them to a and b, where a < b, and exchange the codes between positions a and b;

[0157] Neighborhood structure 5: Randomly select a supplier in the sequence SN-OQ, define variables a and b, randomly generate two integers in the interval (1, T) and assign them to a and b, where a < b, and reverse the order of the codes between positions a and b;

[0158] Neighborhood structure 6: Randomly select a supplier in the sequence SN-OQ, define variables a and b, randomly generate two integers in the interval (1, T) and assign them to a and b, where a < b, and exchange the codes on the left side of position a and on the right side of position b.

[0159] S36, using neighborhood structure NS l Perform a local search on the solution τ′ to obtain the local optimal solution τ″, compare the congestion of τ and τ″, if τ″ is better than τ, set τ = τ″, and record the number of successful searches for the local optimal solution under the neighborhood structure. Otherwise, record the number of failed searches for the local optimal solution under the neighborhood structure and update the selection probability of the neighborhood structure; including:

[0160] S3601, for the neighborhood structure NS l , in the operation of performing variable neighborhood local search, record the number of successful searches for the local optimal solution in each neighborhood structure S l and the number of failures F l , the neighborhood selection probability is updated according to the number of successful searches to achieve neighborhood selection adaptation, which is beneficial for the rapid convergence of the algorithm results. The formula is as follows: Among them SR l Represents the neighborhood structure NS l The local search success rate, SR l =S l / (S l +F l )+ε, ε is a sufficiently small positive number;

[0161] S3602, define variable V as neighborhood selection variable, where V is a randomly generated number in the interval (0,1), and determine the range of V. If SR l-1 <V<SR l , then select the neighborhood structure NS l Performs a neighborhood search operation.

[0162] S37, compare the optimal solution τ under the current neighborhood with the worst individual in the previous generation population. If the optimal solution τ is better than the worst solution of the previous generation population, then use the optimal solution τ to replace the worst solution, and compare the optimal solution τ under the current neighborhood with the global optimal solution τ best Compare, if τ is better than τ best , then let τ best =τ.

[0163] S38, determine whether iter1≤Iter1 is established. If so, set iter1=iter1+1, and use the updated population as the input solution for the next iteration and return to step S32. Otherwise, output the global optimal solution τ best , and decode to obtain the set and quantity of medical supplies ordered from suppliers in each future period, as well as the amount of medical supplies outsourced.

[0164] This embodiment of the present invention designs a hybrid non-dominated sorting genetic algorithm based on neighborhood search, combining the global optimization capabilities of the non-dominated sorting genetic algorithm with the local search capabilities of the variable neighborhood search algorithm. A corresponding adaptive selection mechanism is designed in the non-dominated sorting genetic algorithm and the variable neighborhood search algorithm, along with a special crossover operator and local search neighborhood structure, effectively improving the algorithm's solution efficiency and solution quality.

[0165] The embodiment of the present invention further provides a multi-objective optimization system for medical supplies emergency planning based on demand forecasting, comprising:

[0166] The forecasting module is used to predict the demand for medical supplies in each future period based on the historical data of medical supply demand in the same period;

[0167] A setting module is used to set input parameters of a hybrid non-dominated sorting genetic algorithm based on neighborhood search according to the demand for medical supplies, combined with supplier supply capacity data, supplier supply type data of medical supplies and demand point storage data;

[0168] The solution module is used to obtain the global optimal solution based on the hybrid non-dominated sorting genetic algorithm according to the input parameters, and decode and obtain the set and quantity of medical supplies ordered from suppliers in each future period and the amount of medical supplies outsourced.

[0169] It can be understood that the multi-objective optimization system for medical supplies emergency plan based on demand forecast provided in the embodiment of the present invention corresponds to the multi-objective optimization method for medical supplies emergency plan based on demand forecast provided in the embodiment of the present invention. The explanation, examples and beneficial effects of the relevant contents can refer to the corresponding parts in the multi-objective optimization method for medical supplies emergency plan, and will not be repeated here.

[0170] In summary, compared with the existing technology, the present invention has the following beneficial effects:

[0171] 1. The embodiment of the present invention performs multi-objective optimization on the medical supplies demand plan predicted based on the gray model while taking into account the inconsistent capabilities of suppliers. It comprehensively considers the minimum and maximum response time and minimum cost of suppliers, solves the problem through a hybrid intelligent algorithm and obtains an approximate optimal solution. It can effectively coordinate and optimize and design the order quantity of multiple cycles, thereby maximizing the use of supplier resources and reducing the overall cost of the emergency supply chain, thereby improving the management capability and operational efficiency of the emergency medical supplies supply chain.

[0172] 2. This embodiment of the present invention designs a hybrid non-dominated sorting genetic algorithm based on neighborhood search, combining the global optimization capabilities of the non-dominated sorting genetic algorithm with the local search capabilities of the variable neighborhood search algorithm. A corresponding adaptive selection mechanism is designed in the non-dominated sorting genetic algorithm and the variable neighborhood search algorithm, along with a special crossover operator and local search neighborhood structure, effectively improving the algorithm's solution efficiency and solution quality.

[0173] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0174] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A multi-objective optimization method for emergency planning of medical supplies based on demand forecasting, characterized in that: Including: S1. Predict the demand for medical supplies in each future period based on the historical data of the same period of medical supply demand; S2. Set the input parameters of the hybrid non-dominated sorting genetic algorithm based on neighborhood search according to the medical supply demand, combined with the supplier supply capacity data, the types of medical supplies supplied by the supplier, and the demand point warehousing data; S3. Based on the input parameters, obtain the global optimal solution based on the hybrid non-dominated sorting genetic algorithm, and decode to obtain the set of medical supplies ordered from the supplier, the quantity, and the amount of outsourced medical supplies in each future period; The input parameters in S2 include: Define the cycle index i = 1, 2,..., l; the supplier index j = 1, 2,..., m; the medical supply index k = 1, 2,..., n; A collection of types of medical supplies supplied by suppliers {Q 11 , ..., Q jk , ..., Q mn }, where Q jk It means that supplier j supplies medical supplies k; Initial medical supplies storage at demand point {SQ 11 , ..., SQ ik , ..., SQ ln }, where SQ ik represents the inventory of medical supplies k in period i; Supplier production capacity set {C 11 , C 12 , ..., C 1k , ..., C jk , ..., C mn }, where C jk represents the ability of supplier j to supply medical supplies k; Supplier medical supplies purchase unit price collection {[c i11 , ..., c i1k , ..., c i1n ],......,[c ij1 , ..., c ijk , ..., c ijn ],......,[c im1 , ..., c imk , ..., c imn ]}, where c ijk represents the purchase price of medical supplies k supplied by supplier j in period i; Medical supplies outsourcing unit price collection in represents the unit price of medical supplies outsourcing for medical supplies k in period i; Medical supplies storage unit price collection in represents the storage unit price of medical supplies k in period i; The demand for medical supplies in each period is {[n 11 , ..., n 1k , ..., n 1n ],......,[n i1 , ..., n ik , ..., n in ],......,[n l1 , ..., n 1k ,...,n ln ]}, where n ik represents the demand for medical supplies k in period i; S3 includes: S31, set the execution parameters of the non-dominated sorting hybrid genetic algorithm based on neighborhood search, including: the maximum number of iterations Iter1, set iter1 = 1; the population size N of the algorithm; the mutation probability P v ; Maximum number of iterations of local search Iter2; Initial cycle selection probability PT i =1 / T; mutation probability P v ; Neighborhood set Each neighborhood structure NS l The initial selection probability P l =1 / l max ; S32, generating an initial population according to the input parameters, wherein the individuals in the population are represented by two solutions, SN-MT-OQ and MT-EC; decoding the solutions X1 and X2 to obtain the non-dominated order and crowding value of the individuals, and recording the current optimal solution as the global optimal solution τ best ; Among them, the number in SN in solution X1 represents the supplier number, the number in MT represents the number of the medical material category, OQ represents the order quantity of medical materials, and the length is Solution: In X2, MT represents the category number of medical supplies and EC represents the outsourced quantity of medical supplies, with a length of n*l; S33. Perform a crossover operation on the individuals in the current population based on the crossover operator to generate a population with the same population size, and merge it with the current population to obtain a new population; S34. Based on the non-dominated sorting and crowding degree values of the individuals in the merged new population, perform a selection operation using the minimum domination degree and maximum crowding degree strategy to obtain a new population with the same population size; S35. Decode and non-dominated sort the individuals in the current population, calculate the crowding value of each individual, mark the individual with the largest crowding value in the same non-dominated level except the head and tail as τ, and select a neighborhood structure NS based on the roulette method. l , generate a neighborhood solution τ′ according to the neighborhood structure; S36, using the neighborhood structure NS l Perform local search on the solution τ′ to obtain the local optimal solution τ″, compare the congestion of τ and τ″, if τ″ is better than τ, set τ = τ″, and record the number of successful searches for the local optimal solution under the neighborhood structure, otherwise record the number of failed searches for the local optimal solution under the neighborhood structure, and update the selection probability of the neighborhood structure; S37, compare the optimal solution τ in the current neighborhood with the worst individual in the previous generation population. If the optimal solution τ is better than the worst solution in the previous generation population, replace the worst solution with the optimal solution τ, and compare the optimal solution τ in the current neighborhood with the global optimal solution τ. best By comparison, if τ is better than τ best , then let τ best = τ; S38, determine whether iter1≤Iter1 holds. If so, set iter1=iter1+1, and use the updated population as the input solution for the next iteration and return to step S32. Otherwise, output the global optimal solution τ best , and decode to obtain the set, quantity and outsourced quantity of medical supplies ordered from suppliers in each future period; The decoding process in S32 is as follows: S3202a, generate the quantity OQ of medical supplies k ordered from supplier j in each period from solution X1 ijk ; Calculate the cost of ordering medical supplies using the following formula: Procurement cost Among them: OQ ijk Indicates the quantity of medical supplies ordered from suppliers during the period; S3202b, calculate the medical supplies storage quantity SQ of the demand point in each cycle by solving X1 and the initial medical supplies storage quantity ik , where SQ i+1，k =SQ ik +OQ i+1，k -n ik ; Use the following formula to calculate the cost of medical supplies storage: S3202c, generate the outsourcing quantity EC of medical supplies k in each cycle from solution X2 ik ; Use the following formula to calculate the cost of outsourcing medical supplies: S3202d. Calculate the total cost and the maximum response time within the cycle according to solution X1 and solution X2; use the following formula to calculate: in Indicates the time of outsourcing medical supplies within the cycle.

2. The multi-objective optimization method for emergency plan of medical supplies as claimed in claim 1, characterized in that: In S1, the medical supply prediction model based on grey theory predicts the demand for medical supplies in each future period.

3. The multi-objective optimization method for emergency plan of medical supplies as claimed in claim 1, characterized in that: S33 includes: S3301, decoding to obtain the minimum and maximum response time and total cost of each individual in the population, respectively recorded as f1(x)=C max and f2(x)=TC; S3302. A new population X1 generated by operating on the current population X based on the crossover operator of the hybrid optimization algorithm, the number of the new population X1 is N; merge the initial population X and the new population X1, and the number of the merged population is 2N; mainly includes the following operations: S3302a, select individual X from the current population s and individual X t , randomly select a cycle in the interval (1, T) according to the roulette strategy, perform the crossover operation in the cycle, and record the number of selections CT in the cycle i ; S3302b, within this period, in the interval Randomly generate two integers Cut1 and Cut2 as intersection points, where New individual X′ s Copy individual X t The SN-MT-OQ is located between Cut1 and Cut2, and the remaining codes are reserved for individual X s The encoding in; the new individual X′ t Copy individual X s The SN-MT-OQ is located between Cut1 and Cut2, and the remaining codes are reserved for individual X t The encoding in; S3302c, compare individual X s With X′ s , X t With X′ t , if X′ s Better than X s or X′ t Better than X t If there is at least one, then record the number of successes of this cycle ST i =ST i +1, and calculate the success rate of this cycle according to the formula PS i , formula such as: PS i =ST i / CT i ; S3302d, determine whether the crossover number d is greater than popsize / 2, if so, terminate the crossover operation and update the cycle selection probability Otherwise, go to S3302a.

4. The multi-objective optimization method for emergency plan of medical supplies as claimed in claim 3, characterized in that: S34 includes: S3401. Define the non-dominated sorting rule: for individual x s and individual x t , for any s=1,2,...,N, f1(x s )≤f1(x t ), and there exists at least one f1(x s ) <f1(x t ) holds true, then the individual x s Dominant individual x t , and set s=1; S3402. For all t=1, 2, ..., N and t≠s, compare the individual x according to the definition. s and individual x t The non-dominance relationship between them; if there is no individual x t Better than individual x s , then label the individual x s is a non-dominated individual and is added to the set F1; S3403. Let s = s + 1, go to S3402, until all non-dominated individuals are found, go to S3404; S3404. Calculate the difference between the population X and the set F1 and use it as the new set X′, go to S3402, and repeat non-dominated selection for the individuals in the set X′ until all individuals in the population are classified according to the non-dominated relationship, go to S3405; S3405. Calculate the crowding degree of individuals in the same non-dominated level. r , the formula includes: Among them, m represents the objective function index; C r [s] m represents the individual x in the mth objective function s The degree of congestion; x max Represents individual x s The maximum value under the m function; x min Represents individual x s The minimum value under the m function; S3406. Select the levels in the order from low to high of the non-dominated levels, and select individuals in the same level in the order from large to small of the crowding degree as the individuals of the next generation population until N population individuals are selected.

5. The multi-objective optimization method for emergency plan of medical supplies as claimed in claim 1, characterized in that: The non-dominated sorting hybrid genetic algorithm designs 6 neighborhood structures: Neighborhood structure 1: Randomly select a cycle, define variables a and b, randomly generate two integers in the interval (1, m) and assign them to a and b, where a < b, and exchange the encoding between positions a and b; Neighborhood structure 2: Randomly select a cycle, define variables a and b, randomly generate two integers in the interval (1, m) and assign them to a and b, where a < b, and reverse the encoding between positions a and b in the sequence SN - OQ; Neighborhood structure 3: Randomly select a cycle, define variables a and b, randomly generate two integers in the interval (1, m) and assign them to a and b, where a < b, and exchange the encoding to the left of position a and to the right of position b in the sequence SN - OQ; Neighborhood Structure 4: Randomly select a supplier in the sequence SN-OQ, define variables a and b, randomly generate two integers in the interval (1, T) and assign them to a and b, where a < b, and swap the codes between positions a and b; Neighborhood Structure 5: Randomly select a supplier in the sequence SN-OQ, define variables a and b, randomly generate two integers in the interval (1, T) and assign them to a and b, where a < b, and reverse the order of the codes between positions a and b; Neighborhood Structure 6: Randomly select a supplier in the sequence SN-OQ, define variables a and b, randomly generate two integers in the interval (1, T) and assign them to a and b, where a < b, and swap the codes to the left of position a and to the right of position b.

6. The multi-objective optimization method for emergency plan of medical supplies as claimed in claim 1, characterized in that: The S36 includes: S3601, for the neighborhood structure NS l , in the operation of performing variable neighborhood local search, record the number of successful searches for the local optimal solution in each neighborhood structure S l and the number of failures F l , update the neighborhood selection probability according to the number of successful searches, the formula is as follows: SR l Represents the neighborhood structure NS l The local search success rate, SR l =S l / (S l +F l )+ε, ε is a sufficiently small positive number; S3602, define variable V as neighborhood selection variable, where V is a number randomly generated in the interval (0, 1), determine the range of V, if it satisfies SR l-1 <V<SR l , then select the neighborhood structure NS l Performs a neighborhood search operation.

7. A multi-objective optimization system for emergency planning of medical supplies based on demand forecasting, characterized in that: For executing the multi-objective optimization method of the medical material emergency plan as described in claim 1, including: A prediction module, configured to predict the demand for medical materials in each future period according to the historical data of the same period of medical material demand; A setting module, configured to set the input parameters of the hybrid non-dominated sorting genetic algorithm based on neighborhood search according to the demand for medical materials, in combination with the supplier supply capacity data, the data of the types of medical materials supplied by the suppliers, and the demand point storage data; A solving module, configured to obtain the global optimal solution based on the hybrid non-dominated sorting genetic algorithm according to the input parameters, and decode to obtain the set of medical materials, the quantity, and the amount of outsourced medical materials ordered from the suppliers in each future period.

Images