Emergency order scheduling method and system based on improved VNS-GSA algorithm

Abstract

The invention provides an emergency order scheduling method and system based on an improved VNS-GSA algorithm, a storage medium and electronic equipment, and relates to the field of emergency order scheduling. On the basis of the actual manufacturing background of a certain battery manufacturing enterprise, the problem of distribution and scheduling of emergency orders is solved, all emergency orders and normal orders are distributed to all factories, and the order processing sequence of all factories is determined according to the production deadline; the emergency order must be completed before the agreed delivery time, the normal order can be delayed after the delivery time is up, but the punishment cost caused by the delay can be continuously increased along with time increase. The aim of the invention is to improve the total production revenue by efficiently sorting normal and emergency orders. Besides, according to production mode characteristics of products such as a battery pack, a corresponding elite reverse learning strategy, a neighborhood structure and a gravitational search formula are designed, and a reasonable scheduling decision is made for each main enterprise in limited time.

Description

Technical Field

[0001] The present invention relates to the field of emergency order scheduling, and in particular to an emergency order scheduling method, system, storage medium and electronic device based on an improved VNS-GSA algorithm. Background Art

[0002] With the rapid development of new generation information technology, batteries, as an indispensable component of modern high-end equipment, are also constantly improving their manufacturing processes. Battery production involves a variety of complex manufacturing processes and technical requirements. From battery cell production to battery pack assembly, each link requires efficient scheduling and resource allocation. In the battery production system, factories usually need to produce a variety of different models of batteries and related parts, and manage multiple production workshops at the same time. Each workshop has unique resources and equipment, and will accept orders with different needs. It is crucial to make effective battery production scheduling decisions and resource allocation strategies, which can not only reduce the production costs of enterprises and improve production efficiency, but also enhance the competitiveness of enterprises in the market.

[0003] In the related art, the traditional scheduling algorithm is based on a specific production situation and cannot effectively respond to the scheduling needs in the real-time environment. The resources owned by each workshop in the factory and the specific environment in which they are located are not taken into account. For example, the tasks in the real-time system may have strict time limits, but the traditional algorithm may not be able to guarantee the timely completion of the tasks. Under the current multi-workshop manufacturing conditions, it is necessary to consider the processing and resource allocation of all different types of workpieces in order to complete the processing as much as possible within the specified time, thereby increasing production capacity and reducing costs. In previous studies, the allocation of workpieces in the actual production process rarely considered the scheduling optimization under the constraints of the production deadline, which led to some research algorithms in the actual production process. Deviations from the theory will occur, and good results cannot be obtained. In addition, there is also the problem of changes in battery orders in the battery production process, that is, in addition to receiving normal production orders, important urgent orders with clear delivery dates will be received at the same time, which will cause the original normal order production plan to be disrupted.

[0004] In view of this, it is necessary to provide a technical solution that can effectively respond to scheduling needs in a real-time environment, especially to solve the problem of emergency order allocation and scheduling. Summary of the invention

[0005] 1. Technical issues to be resolved

[0006] In view of the deficiencies in the prior art, the present invention provides an emergency order scheduling method, system, storage medium and electronic device based on an improved VNS-GSA algorithm, which solves the technical problem of emergency order allocation and scheduling.

[0007] (II) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0009] An urgent order scheduling method based on an improved VNS-GSA algorithm comprises:

[0010] Obtaining a set of orders to be processed and a set of factories; wherein the set of orders includes a number of normal orders and urgent orders;

[0011] Based on the production deadline of each normal order, the real cost of the factory producing the normal order is solved to obtain the total revenue of the factory; with the goal of maximizing the total revenue and limiting each factory to be assigned at most one emergency order, a mathematical model is constructed;

[0012] The improved VNS-GSA algorithm is used to solve the mathematical model and obtain the optimal emergency order scheduling solution.

[0013] Preferably, the actual cost of the factory producing the normal order is solved based on the production deadline of each normal order; it is expressed as:

[0014] D mn =d mn +α*max(0,T mn -t n ) (1)

[0015] Among them, D mn represents the true cost of the mth factory producing the nth normal order, and the cost will continue to rise as the deadline is exceeded; α is the penalty parameter; max is the maximization function; T mn represents the production completion time of the nth normal order of the mth factory; t n Indicates the production deadline for the nth normal order.

[0016] Preferably, the mathematical model includes:

[0017]

[0018]

[0019]

[0020]

[0021] Wherein, formula (2) represents the maximum profit in the production process; W represents the total profit; the number of factories is M, the number of normal orders is N, the number of emergency orders is E, and the number of processing positions in the mth factory is R m , M≥E; a mrirepresents the revenue of the mth factory and the rth processing position producing the ith order; d me represents the cost of producing the e-th emergency order at the i-th factory;

[0022] Constraint (3) indicates that urgent orders need to be produced before their deadline; represents the production completion time of the e-th emergency order at the r-th location of the m-th factory; t e represents the production deadline of the e-th emergency order;

[0023] Constraint (4) means that the production start time of a normal order is greater than the production completion time of the previous order and less than the production completion time of the current normal order; They represent the production start time and production completion time of the nth normal order produced at the rth position of the mth factory respectively; It represents the production completion time of the nth normal order produced at the r-1th position of the mth factory;

[0024] Constraint (5) indicates that the production start time of the emergency order is greater than the production completion time of the previous normal order and less than the production completion time of the current emergency order; They represent the production start time and production completion time of the e-th emergency order at the r-th location of the m-th factory, respectively; It indicates the production completion time of the nth normal order produced at the r-1th position of the mth factory.

[0025] Preferably, the improved VNS-GSA algorithm is used to solve the mathematical model to obtain the optimal emergency order scheduling solution; including:

[0026] S31. Initialize the input parameters of the algorithm, including the number of normal orders, the number of urgent orders, and the production cost, production revenue, and production time of the orders in each factory, and encode all workpiece allocations using integer encoding; the encoding rules are as follows:

[0027] Solution vector 1 is denoted as V e =(v1,v2,v3,v e ,...,v E ), where v1,v2,v3,...,v E is any integer from 1 to M and is different for each. e The value of represents the number of the corresponding factory to which the e-th emergency order is assigned; the solution vector 2 is represented by V N =(v1,v2,v3,v n ,...,v N ), where v1,v2,v3,...,v N is any integer from 1 to M, vn The value of represents the number of the corresponding factory to which the nth normal order is assigned;

[0028] S32, according to the preset heuristic algorithm, obtain the workpiece processing order of each factory, and set the fitness function based on formula (2) to obtain the initial solution X s ;

[0029] S33, setting the execution parameters of the VNS algorithm, wherein the execution parameters include the initial solution s , the maximum number of iterations I max , current iteration number I, minimum neighborhood structure index value o min , current neighborhood structure index value o, maximum neighborhood structure index value o max ; The local search algorithm includes the maximum number of iterations it max , the current iteration number it=1, the local search algorithm population size Q; the global optimal solution X best =X s ;

[0030] S34, set I=0;

[0031] S35. Set o=o min ;

[0032] S36, at the current global optimal solution X best Randomly select Q individuals from the o-th neighborhood of to obtain the individual set S;

[0033] S37, using the elite reverse learning strategy for the initial population S, to obtain the population individual set S′ in the local search based on the local search algorithm;

[0034] S38, execute the GSA algorithm to obtain the local approximate optimal solution X′; determine that X′ is better than the current global optimal solution X best , then assign X′ to X best , and go to S34; otherwise go to S39;

[0035] S39, assign o+1 to o, and determine whether o is not greater than o max , if established, go to S36, otherwise go to S310;

[0036] S310, assign I+1 to I, and judge I≤I max Is it true? If so, execute S39; otherwise, go to S311;

[0037] S311, assign it+1 to it, and judge it≤it max Is it established? If so, execute S34; otherwise, output the global optimal solution and its fitness value, and decode to obtain the optimal emergency order scheduling plan.

[0038] Preferably, the S32 includes:

[0039] S321. Arrange the normal order sequence of the current factory in non-descending order according to the production deadline, n=1, C n =0, C e =0; where C n , C e They represent the production start time of the nth normal order and the eth emergency order respectively;

[0040] S322, if the normal order n satisfies C n-1 +T mn +T me ≤t e , then C n =C n-1 +T mn , n=n+1; where T mn , T me They represent the production time required for the nth normal order and the eth emergency order in the mth factory respectively;

[0041] S323, if the normal order n satisfies C n-1 +T mn +T me >t e , then C e =C n-1 +T me , C n =C e +T e , n=n+1, otherwise go to step S322;

[0042] S324、C n =C n-1 +T mn , n = n + 1;

[0043] S325. If n=N+1, calculate the total factory revenue W based on formula (2); otherwise, go to step S324.

[0044] Preferably, the VNS algorithm designs the following three neighborhood structures:

[0045] (1) Neighborhood structure 1, changing a random element in solution vector 1 to an unused element in the element range to obtain a new solution; including:

[0046] Step a1: Define the element range m, obtain the input initial solution X, and randomly obtain the index x of an element from the element sequence;

[0047] Step a2: randomly generate a random integer y in the range of 1 to m;

[0048] Step a3: If y is equal to any element in the initial solution X, return to step a2; otherwise, execute step a4;

[0049] Step a4: Change the element at index x in the initial solution X to y to obtain the solution X′;

[0050] Step a5: Output the solution X′;

[0051] (2) Neighborhood structure 2, reversing the order of elements between two random elements in solution vector 2 to obtain a new solution; including:

[0052] Step b1: define variables x, y, g = 1, obtain the input initial solution X, variable o;

[0053] Step b2: Randomly generate a random integer ranging from 1 to N and assign it to x and y;

[0054] Step b3: swap the xth element and the yth element in the solution X;

[0055] Step b4: x=x+1, y=y-1, if x≥y, a new solution X′ is obtained and step b5 is executed; otherwise, return to step b3;

[0056] Step b5: g=g+1, if g≤o holds, return to step b2; otherwise, execute step b5;

[0057] Step b6: Output the solution X′;

[0058] (3) Neighborhood structure 3, moving a random workpiece from the factory with the highest profit to another factory; including:

[0059] Step c1: Get factory revenue W j The largest factory j randomly generates a random integer a between 1 and m that is not j;

[0060] Step c2: Change a random j in solution vector 2 to a to generate a new solution X′;

[0061] Step c3: Output the solution X′.

[0062] Preferably, the S37 includes:

[0063] Step 371: Randomly select Q solutions as the individual set S according to the current global optimal solution;

[0064] Step 372: Arrange the individuals in the set S in descending order according to their fitness values, and form the elite population P with the first n / 2 individuals;

[0065] Step 373, forming an elite reverse population O according to the reverse solutions of the individuals in the elite population P; The reverse solution corresponding to an elite individual in N-dimensional space is expressed as in They represent the number of the corresponding factory assigned to the nth normal order in the i-th individual of the e-th generation population and its reverse solution;

[0066] The reverse solution of individuals in the elite population P is calculated according to the following formula, where rand is a random number in [0,1], and lb j andub j are the lower and upper bounds of the population space respectively;

[0067]

[0068] At the same time, in order to prevent the elite reverse solution from crossing the boundary, cross-boundary judgment and update are also performed:

[0069]

[0070] Step 374: merge the initial populations S and O, select the first n individuals with higher fitness according to the fitness value rule from high to low, and establish a new population S′.

[0071] Preferably, the S38 includes:

[0072] Step 381: The population S′ obtained by the elite reverse learning strategy algorithm is used as the initial solution of the local search algorithm; the jth individual of the oth generation population of the local search algorithm is defined as in Represents the element on the nth dimension of the jth individual in the oth generation population; let o = 1;

[0073] Step 382: Calculate the fitness value of each individual in the oth generation population, f j (o) represents the fitness value of the jth individual; the worst individual worst(o) and the best individual best(o) in the oth generation population are obtained;

[0074] Step 383: Obtain the mass of each individual in the population according to the following formula: M j (o) represents the mass of the jth individual in the population;

[0075]

[0076]

[0077] Step 384, obtain the acceleration of each individual in the population in each dimension according to the following formula; Expressed as the acceleration of the jth individual in the dth dimension;

[0078]

[0079]

[0080] Step 385: Obtain the position and velocity of each individual in each dimension of the current population according to the following formula; represents the d-th dimension velocity of the j-th individual;

[0081]

[0082]

[0083] Step 386: If o≤o max , then o=o+1, and go to step 382; otherwise, end the algorithm and output the local optimal approximate solution X′, judging that X′ is better than the current global optimal solution X vest , then assign X′ to X best , and go to S34; otherwise go to S39.

[0084] An emergency order scheduling system based on an improved VNS-GSA algorithm comprises:

[0085] An acquisition module is used to acquire a set of orders to be processed and a set of factories; wherein the set of orders includes a number of normal orders and urgent orders;

[0086] A construction module is used to solve the real cost of the factory producing the normal order based on the production deadline of each normal order to obtain the total revenue of the factory; a mathematical model is constructed with the goal of maximizing the total revenue and limiting each factory to be assigned at most one emergency order;

[0087] The solution module is used to solve the mathematical model by using the improved VNS-GSA algorithm to obtain the optimal emergency order scheduling solution.

[0088] A computer-readable storage medium stores a computer program, which can implement the emergency order scheduling method based on the improved VNS-GSA algorithm as described above when a processor executes the computer program.

[0089] An electronic device, comprising:

[0090] One or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, the programs including a method for executing the emergency order scheduling method based on the improved VNS-GSA algorithm as described above.

[0091] (III) Beneficial effects

[0092] The present invention provides an emergency order scheduling method, system, storage medium and electronic device based on an improved VNS-GSA algorithm. Compared with the prior art, the present invention has the following beneficial effects:

[0093] In the present invention, firstly, a set of orders to be processed and a set of factories are obtained; wherein the set of orders includes several normal orders and emergency orders; then, based on the production deadline of each normal order, the real cost of the factory producing the normal order is solved to obtain the total revenue of the factory; with the goal of maximizing the total revenue, and limiting each factory to at most one emergency order, a mathematical model is constructed; finally, the improved VNS-GSA algorithm is used to solve the mathematical model to obtain the optimal emergency order scheduling solution. The present invention effectively sorts normal orders and emergency orders to improve the total production revenue, which is closer to the real production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0095] Figure 1 A block diagram of an emergency order scheduling method based on an improved VNS-GSA algorithm provided by an embodiment of the present invention;

[0096] Figure 2 A flowchart based on an improved VNS-GSA algorithm is provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0097] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0098] The embodiments of the present application solve the technical problem of emergency order allocation and scheduling by providing an emergency order scheduling method, system, storage medium and electronic device based on an improved VNS-GSA algorithm.

[0099] The technical solution in the embodiment of the present application is to solve the above technical problems, and the overall idea is as follows:

[0100] The embodiment of the present invention is based on the actual manufacturing background of a battery manufacturing enterprise, solves the allocation and scheduling problem of emergency orders, allocates each emergency order and normal order to each factory, and determines the order processing order of each factory according to the size of the production deadline; emergency orders must be completed before the agreed delivery date, and normal orders can be postponed after the delivery deadline, but the penalty cost caused by the delay will continue to increase with time. The goal is to increase the total production revenue by effectively sorting normal orders and emergency orders. The purpose of the present invention is to design corresponding elite reverse learning strategies, neighborhood structures, and gravitational search formulas according to the characteristics of the production model of battery packs, so as to make reasonable scheduling decisions for each main enterprise within a limited time, or provide decision support for the decision makers of the enterprise, so as to improve the overall production capacity of the manufacturing system and reduce production costs.

[0101] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0102] Embodiment 1:

[0103] like Figure 1 As shown, an embodiment of the present invention provides an emergency order scheduling method based on an improved VNS-GSA algorithm, comprising:

[0104] S1. Obtain a set of orders to be processed and a set of factories; wherein the set of orders includes a number of normal orders and urgent orders;

[0105] S2. Based on the production deadline of each normal order, solve the real cost of the factory producing the normal order to obtain the total revenue of the factory; take maximizing the total revenue as the goal and limit each factory to be assigned at most one emergency order, and construct a mathematical model;

[0106] S3. Use the improved VNS-GSA algorithm to solve the mathematical model and obtain the optimal emergency order scheduling solution.

[0107] The present invention effectively sorts normal orders and urgent orders to increase total production revenue and is closer to the actual production process.

[0108] The embodiment of the present invention realizes the manufacturing and optimization of products by designing a production scheduling algorithm based on a variable neighborhood search algorithm and a gravitational search algorithm. Figure 2 As shown, Figure 2 The flowchart of the improved VNS-GSA algorithm is given.

[0109] Next, we will combine Figure 2 The following are the steps of the above solution:

[0110] In step S1, a set of orders to be processed and a set of factories are obtained; wherein the order set includes a number of normal orders and urgent orders.

[0111] This step obtains data information such as the order set and factory set to be processed for subsequent mathematical modeling, and gives the symbol description shown in Table 1:

[0112] Table 1

[0113]

[0114] In step S2, based on the production deadline of each normal order, the actual cost of the factory producing the normal order is solved to obtain the total revenue of the factory; with the goal of maximizing the total revenue and limiting each factory to be assigned at most one emergency order, a mathematical model is constructed.

[0115] First of all, the model is based on the following assumptions:

[0116] (1) The arrival time of all orders is 0;

[0117] (2) The processing time of normal production orders and emergency orders is known in advance;

[0118] (3) The processing time for the same order assigned to different factories may not be the same;

[0119] (4) The production costs of the same order in different factories may not be the same;

[0120] (5) The penalty cost parameters for all factory production orders are the same;

[0121] (6) Each factory can only process one order at any time and the processing should be continuous and cannot be interrupted by other orders.

[0122] In particular, the embodiment of the present invention considers that the real cost of a factory producing a normal order is related to its production deadline, and specifically sets that the cost will continue to increase as the time exceeds the deadline, which is expressed as:

[0123]

[0124] Among them, max is the maximization function.

[0125] On this basis, this step constructs a mathematical model, including:

[0126]

[0127]

[0128]

[0129]

[0130] Among them, formula (2) represents the maximum profit in the production process.

[0131] Constraint (3) indicates that urgent orders need to be produced before their deadline.

[0132] Constraint (4) indicates that the production start time of a normal order is greater than the production completion time of the previous order and less than the production completion time of the current normal order.

[0133] Constraint (5) indicates that the production start time of the emergency order is greater than the production completion time of the previous normal order and less than the production completion time of the current emergency order.

[0134] In step S3, the improved VNS-GSA algorithm is used to solve the mathematical model to obtain the optimal emergency order scheduling solution.

[0135] The variable neighborhood search algorithm (VNS) diversifies the search direction by changing the neighborhood structure, thereby enhancing the search capability and optimizing the computing performance; and the gravitational search algorithm is inspired by the law of universal gravitation. This algorithm treats all particles as objects with mass, which can move without resistance. Each particle will be affected by the universal gravitation of other particles in the solution space, and will generate acceleration to move toward particles with larger mass. Since the mass of a particle is related to the fitness value of the particle, particles with large fitness values ​​will also have larger masses. Therefore, particles with small mass gradually approach the optimal solution in the optimization problem as they approach large mass. The gravitational search algorithm has the advantages of high efficiency, simplicity, and suitability for parallelization in global parallel search, and has been applied to many fields.

[0136] On this basis, the embodiment of the present invention takes into account the impact of emergency order insertion on the overall order distribution, and designs an improved GSA algorithm mixed with the VNS algorithm VNS-GSA algorithm with the goal of maximizing the total benefit. The total benefit of different schemes is used as the mass of the individual, and the individuals with heavier mass are found according to the gravity formula. The algorithm is continuously iterated in the multi-dimensional space to find the global optimal solution that meets the requirements.

[0137] Specifically, this step includes:

[0138] S31. Initialize the input parameters of the algorithm, including the number of normal orders, the number of urgent orders, and the production cost, production revenue, and production time of the orders in each factory, and encode all workpiece allocations using integer encoding; the encoding rules are as follows:

[0139] Solution vector 1 is denoted as V e=(v1,v2,v3,v e ,...,v E ), where v1,v2,v3,...,v E is any integer from 1 to M and is different for each. e The value of represents the number of the corresponding factory to which the e-th emergency order is assigned; the solution vector 2 is represented by V N =(v1,v2,v3,v n ,...,v N ), where v1,v2,v3,...,v N is any integer from 1 to M, v n The value of represents the number of the corresponding factory to which the nth normal order is assigned.

[0140] For example, V N =(1,,4,2,1,3,4) means that normal orders 1 to 6 are allocated to factory 1, factory 4, factory 2, factory 1, factory 3 and factory 4 respectively.

[0141] S32, according to the preset heuristic algorithm, obtain the workpiece processing order of each factory, and set the fitness function based on formula (2) to obtain the initial solution X s ;include:

[0142] S321. Arrange the normal order sequence of the current factory in non-descending order according to the production deadline, n=1, C n =0, C e =0; where C n , C e They represent the production start time of the nth normal order and the eth emergency order respectively;

[0143] S322, if the normal order n satisfies C n-1 +T mn +T me ≤t e , then C n =C n-1 +T mn , n=n+1; where T mn , T me They represent the production time required for the nth normal order and the eth emergency order in the mth factory respectively;

[0144] S323, if the normal order n satisfies C n-1 +T mn +T me >t e , then C e =C n-1 +T me , C n =Ce +T e , n=n+1, otherwise go to step S322;

[0145] S324、C n =C n-1 +T mn , n = n + 1;

[0146] S325. If n=N+1, calculate the total factory revenue W based on formula (2); otherwise, go to step S324.

[0147] S33, setting the execution parameters of the VNS algorithm, wherein the execution parameters include the initial solution s , the maximum number of iterations I max , current iteration number I, minimum neighborhood structure index value o min , current neighborhood structure index value o, maximum neighborhood structure index value o max ; The local search algorithm includes the maximum number of iterations it max , the current iteration number it=1, the local search algorithm population size Q; the global optimal solution X best =X s .

[0148] S34. Set I=0.

[0149] S35. Set o=o min .

[0150] S36, at the current global optimal solution X best Randomly select Q individuals from the o-th neighborhood of and obtain the individual set S.

[0151] Exemplarily, the following three neighborhood structures are designed here:

[0152] (1) Neighborhood structure 1, changing a random element in solution vector 1 to an unused element in the element range to obtain a new solution; including:

[0153] Step a1: Define the element range m, obtain the input initial solution X, and randomly obtain the index x of an element from the element sequence;

[0154] Step a2: randomly generate a random integer y in the range of 1 to m;

[0155] Step a3: If y is equal to any element in the initial solution X, return to step a2; otherwise, execute step a4;

[0156] Step a4: Change the element at index x in the initial solution X to y to obtain the solution X′;

[0157] Step a5: Output the solution X′.

[0158] (2) Neighborhood structure 2, reversing the order of elements between two random elements in solution vector 2 to obtain a new solution; including:

[0159] Step b1: define variables x, y, g = 1, obtain the input initial solution X, variable o;

[0160] Step b2: Randomly generate a random integer ranging from 1 to N and assign it to x and y;

[0161] Step b3: swap the xth element and the yth element in the solution X;

[0162] Step b4: x=x+1, y=y-1, if x≥y, a new solution X′ is obtained and step b5 is executed; otherwise, return to step b3;

[0163] Step b5: g=g+1, if g≤o holds, return to step b2; otherwise, execute step b5;

[0164] Step b6: Output the solution X′.

[0165] (3) Neighborhood structure 3, moving a random workpiece from the factory with the highest profit to another factory; including:

[0166] Step c1: Get factory revenue W j The largest factory j randomly generates a random integer a between 1 and m that is not j;

[0167] Step c2: Change a random j in solution vector 2 to a to generate a new solution X′;

[0168] Step c3: Output the solution X′.

[0169] S37, using the elite reverse learning strategy for the initial population S, to obtain a population individual set S′ in the local search based on the local search algorithm; including:

[0170] Step 371: Randomly select Q solutions as the individual set S according to the current global optimal solution;

[0171] Step 372: Arrange the individuals in the set S in descending order according to their fitness values, and form the elite population P with the first n / 2 individuals;

[0172] Step 373, forming an elite reverse population O according to the reverse solutions of the individuals in the elite population P; The reverse solution corresponding to an elite individual in N-dimensional space is expressed as in They represent the number of the corresponding factory assigned to the nth normal order in the i-th individual of the e-th generation population and its reverse solution;

[0173] The reverse solution of individuals in the elite population P is calculated according to the following formula, where rand is a random number in [0,1], and lb j andub j are the lower and upper bounds of the population space respectively;

[0174]

[0175] At the same time, in order to prevent the elite reverse solution from crossing the boundary, cross-boundary judgment and update are also performed:

[0176]

[0177] Step 374: merge the initial populations S and O, select the first n individuals with higher fitness according to the fitness value rule from high to low, and establish a new population S′.

[0178] S38, execute the GSA algorithm to obtain the local approximate optimal solution X′; determine that X′ is better than the current global optimal solution X best , then assign X′ to X best , and go to S34; otherwise go to S39; including:

[0179] Step 381: The population S′ obtained by the elite reverse learning strategy algorithm is used as the initial solution of the local search algorithm; the jth individual of the oth generation population of the local search algorithm is defined as in Represents the element on the nth dimension of the jth individual in the oth generation population; let o = 1;

[0180] Step 382: Calculate the fitness value of each individual in the oth generation population, f j (o) represents the fitness value of the jth individual; the worst individual worst(o) and the best individual best(o) in the oth generation population are obtained;

[0181] Step 383: Obtain the mass of each individual in the population according to the following formula: M j (o) represents the mass of the jth individual in the population;

[0182]

[0183]

[0184] Step 384, obtain the acceleration of each individual in the population in each dimension according to the following formula; Expressed as the acceleration of the jth individual in the dth dimension;

[0185]

[0186]

[0187] Step 385: Obtain the position and velocity of each individual in each dimension of the current population according to the following formula; represents the d-th dimension velocity of the j-th individual;

[0188]

[0189]

[0190] Step 386: If o≤o max , then o=o+1, and go to step 382; otherwise, end the algorithm and output the local optimal approximate solution X′, judging that X′ is better than the current global optimal solution X best , then assign X′ to X best , and go to S34; otherwise go to S39.

[0191] S39, assign o+1 to o, and determine whether o is not greater than o max If so, go to S36, otherwise go to S310.

[0192] S310, assign I+1 to I, and judge I≤I max Is it true? If so, execute S39; otherwise, go to S311.

[0193] S311, assign it+1 to it, and judge it≤it max Is it established? If so, execute S34; otherwise, output the global optimal solution and its fitness value, and decode to obtain the optimal emergency order scheduling plan.

[0194] So far, the embodiment of the present invention has completed the complete process of the emergency order scheduling method based on the improved VNS-GSA algorithm.

[0195] Embodiment 2:

[0196] The embodiment of the present invention provides an emergency order scheduling system based on an improved VNS-GSA algorithm, including:

[0197] An acquisition module is used to acquire a set of orders to be processed and a set of factories; wherein the set of orders includes a number of normal orders and urgent orders;

[0198] A construction module is used to solve the real cost of the factory producing the normal order based on the production deadline of each normal order to obtain the total revenue of the factory; a mathematical model is constructed with the goal of maximizing the total revenue and limiting each factory to be assigned at most one emergency order;

[0199] The solution module is used to solve the mathematical model by using the improved VNS-GSA algorithm to obtain the optimal emergency order scheduling solution.

[0200] Embodiment 3:

[0201] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When a processor executes the computer program, the emergency order scheduling method based on the improved VNS-GSA algorithm as described in the above embodiment 1 can be implemented.

[0202] Embodiment 4:

[0203] An embodiment of the present invention provides an electronic device, including:

[0204] One or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the programs include a method for executing an emergency order scheduling method based on an improved VNS-GSA algorithm as described in Example 1.

[0205] It can be understood that the emergency order scheduling system, storage medium and electronic device based on the improved VNS-GSA algorithm provided in the embodiments of the present invention correspond to the emergency order scheduling method based on the improved VNS-GSA algorithm provided in the embodiments of the present invention, and the explanations, examples and beneficial effects of the relevant contents can refer to the corresponding parts in the emergency order scheduling method, which will not be repeated here.

[0206] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0207] 1. The embodiment of the present invention designs an optimization algorithm of a variable neighborhood search algorithm and an improved gravitational search algorithm for scheduling problems of production delivery time and resource allocation of orders. First, the allocation of workpieces to be processed is determined by coding, and then the processing order and total revenue of each workshop workpiece are determined according to the scheduling algorithm. Finally, the approximate optimal solution is found through iteration of the variable neighborhood search algorithm and the improved gravitational search algorithm. This not only avoids the defect of the gravitational search algorithm that is easy to fall into the local optimum, but also retains the diversity of population characteristics, and also accelerates the convergence of the algorithm.

[0208] 2. The embodiment of the present invention is based on the actual production situation of the battery pack production process of a large-scale new energy power lithium battery manufacturing enterprise. It takes into account the problem of changing the production sequence of all orders after the insertion of emergency production orders, and provides corresponding heuristic algorithms to solve the corresponding problems. Finally, it can be applied in the actual production process.

[0209] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0210] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An urgent order scheduling method based on an improved VNS-GSA algorithm, characterized in that: include: Get the order set and factory set to be processed; The order set includes a number of normal orders and urgent orders; Based on the production deadline of each normal order, solving the real cost of the factory producing the normal order to obtain the total revenue of the factory; A mathematical model is constructed with the goal of maximizing total revenue and limiting each factory to be assigned at most one emergency order; The improved VNS-GSA algorithm is used to solve the mathematical model and obtain the optimal emergency order scheduling solution.

2. The method for dispatching urgent orders according to claim 1, characterized in that: Based on the production deadline of each normal order, the real cost of the factory producing the normal order is solved; it is expressed as: D mn =d mn +α*max(0,T mn -t n ) (1) Among them, D mn represents the true cost of the mth factory producing the nth normal order, and the cost will continue to rise as the deadline is exceeded; α is the penalty parameter; max is the maximization function; T mn represents the production completion time of the nth normal order of the mth factory; t n Indicates the production deadline for the nth normal order.

3. The method for dispatching urgent orders according to claim 2, characterized in that: The mathematical model includes: Wherein, formula (2) represents the maximum profit in the production process; W represents the total profit; the number of factories is M, the number of normal orders is N, the number of emergency orders is E, and the number of processing positions in the mth factory is R m , M≥E; a mri represents the revenue of the mth factory and the rth processing position producing the ith order; d me represents the cost of producing the e-th emergency order at the i-th factory; Constraint (3) indicates that urgent orders need to be produced before their deadline; represents the production completion time of the e-th emergency order at the r-th location of the m-th factory; t e represents the production deadline of the e-th emergency order; Constraint (4) means that the production start time of a normal order is greater than the production completion time of the previous order and less than the production completion time of the current normal order; They represent the production start time and production completion time of the nth normal order produced at the rth position of the mth factory respectively; It represents the production completion time of the nth normal order produced at the r-1th position of the mth factory; Constraint (5) indicates that the production start time of the emergency order is greater than the production completion time of the previous normal order and less than the production completion time of the current emergency order; They represent the production start time and production completion time of the e-th emergency order at the r-th location of the m-th factory, respectively; It indicates the production completion time of the nth normal order produced at the r-1th position of the mth factory.

4. The method for dispatching urgent orders according to claim 3, characterized in that: The improved VNS-GSA algorithm is used to solve the mathematical model to obtain the optimal emergency order scheduling solution; including: S31. Initialize the input parameters of the algorithm, including the number of normal orders, the number of urgent orders, and the production cost, production revenue, and production time of the orders in each factory, and encode all workpiece allocations using integer encoding; the encoding rules are as follows: Solution vector 1 is denoted as V e =(v1,v2,v3,v e ,...,v E ), where v1,v2,v3,...,v E is any integer from 1 to M and is different for each. e The value of represents the number of the corresponding factory to which the e-th emergency order is assigned; the solution vector 2 is represented by V N =(v1,v2,v3,v n ,...,v N ), where v1,v2,v3,...,v N is any integer from 1 to M, v n The value of represents the number of the corresponding factory to which the nth normal order is assigned; S32. According to the preset heuristic algorithm, the processing order of the workpieces of each factory is obtained, and the fitness function is set based on formula (2) to obtain the initial solution X s ; S33, setting the execution parameters of the VNS algorithm, wherein the execution parameters include the initial solution s , the maximum number of iterations I max , current iteration number I, minimum neighborhood structure index value o min , current neighborhood structure index value o, maximum neighborhood structure index value o max ; The local search algorithm includes the maximum number of iterations it max , the current number of iterations it = 1, the local search algorithm population size Q; the global optimal solution X best =X s ; S34, set I=0; S35. Set o=o min ; S36, at the current global optimal solution X best Randomly select Q individuals from the o-th neighborhood of to obtain the individual set S; S37, using the elite reverse learning strategy for the initial population S, to obtain the population individual set S′ in the local search based on the local search algorithm; S38, execute the GSA algorithm to obtain the local approximate optimal solution X′; determine that X′ is better than the current global optimal solution X best , then assign X′ to X best , and go to S34; otherwise go to S39; S39, assign o+1 to o, and determine whether o is not greater than o max , if established, go to S36, otherwise go to S310; S310, assign I+1 to I, and judge I≤I max Is it true? If so, execute S39; otherwise, go to S311; S311, assign it+1 to it, and judge it≤it max Is it established? If so, execute S34; otherwise, output the global optimal solution and its fitness value, and decode to obtain the optimal emergency order scheduling plan.

5. The method for dispatching urgent orders according to claim 4, characterized in that: The S32 includes: S321. Arrange the normal order sequence of the current factory in non-descending order according to the production deadline, n=1, C n =0, C e =0; where C n , C e They represent the production start time of the nth normal order and the eth emergency order respectively; S322, if the normal order n satisfies C n-1 +T mn +T me ≤t e , then C n =C n-1 +T mn , n=n+1; where T mn 、T me They represent the production time required for the nth normal order and the eth emergency order in the mth factory respectively; S323, if the normal order n satisfies C n-1 +T mn +T me >t e , then C e =C n-1 +T me , C n =C e +T e , n=n+1, otherwise go to step S322; S324、C n =C n-1 +T mm ,n=n+1; S325, if n=N+1, calculate the total factory revenue W based on formula (2), otherwise go to step S324; And / or the VNS algorithm designs the following three neighborhood structures: (1) Neighborhood structure 1, changing a random element in solution vector 1 to an unused element in the element range to obtain a new solution; including: Step a1: Define the element range m, obtain the input initial solution X, and randomly obtain the index x of an element from the element sequence; Step a2: randomly generate a random integer y in the range of 1 to m; Step a3: If y is equal to any element in the initial solution X, return to step a2; otherwise, execute step a4; Step a4: Change the element at index x in the initial solution X to y to obtain the solution X′; Step a5: Output the solution X′; (2) Neighborhood structure 2, reversing the order of elements between two random elements in solution vector 2 to obtain a new solution; including: Step b1: define variables x, y, g = 1, obtain the input initial solution X, variable o; Step b2: Randomly generate a random integer ranging from 1 to N and assign it to x and y; Step b3: swap the xth element and the yth element in the solution X; Step b4: x=x+1, y=y-1, if x≥y, a new solution X′ is obtained and step b5 is executed; otherwise, return to step b3; Step b5: g=g+1, if g≤o holds, return to step b2; otherwise, execute step b5; Step b6: Output the solution X′; (3) Neighborhood structure 3, moving a random workpiece from the factory with the highest profit to another factory; including: Step c1: Get factory revenue W j The largest factory j randomly generates a random integer a between 1 and m that is not j; Step c2: Change a random j in solution vector 2 to a to generate a new solution X′; Step c3: Output the solution X′.

6. The method for dispatching urgent orders according to claim 5, characterized in that: The S37 includes: Step 371: Randomly select Q solutions as the individual set S according to the current global optimal solution; Step 372: Arrange the fitness values ​​of the individuals in the set S in descending order, and form the elite population P with the first n / 2 individuals; Step 373, forming an elite reverse population O according to the reverse solutions of the individuals in the elite population P; The reverse solution corresponding to an elite individual in N-dimensional space is expressed as in They represent the number of the corresponding factory assigned to the nth normal order in the i-th individual of the e-th generation population and its reverse solution; The reverse solution of individuals in the elite population P is calculated according to the following formula, where rand is a random number in [0,1], and lb j andub j are the lower and upper bounds of the population space respectively; At the same time, in order to prevent the elite reverse solution from crossing the boundary, cross-boundary judgment and update are also performed: Step 374: merge the initial populations S and O, select the first n individuals with higher fitness according to the fitness value rule from high to low, and establish a new population S′.

7. The method for dispatching urgent orders according to claim 6, characterized in that: The S38 includes: Step 381: The population S′ obtained by the elite reverse learning strategy algorithm is used as the initial solution of the local search algorithm; the jth individual of the oth generation population of the local search algorithm is defined as 1≤o≤o max , j = 1, 2, ..., Q, n = 1, 2, ..., N, where Represents the element on the nth dimension of the jth individual in the oth generation population; let o = 1; Step 382: Calculate the fitness value of each individual in the oth generation population, f j (o) represents the fitness value of the jth individual; the worst individual worst(o) and the best individual best(o) in the oth generation population are obtained; Step 383: Obtain the mass of each individual in the population according to the following formula: M j (o) represents the mass of the jth individual in the population; Step 384, obtain the acceleration of each individual in the population in each dimension according to the following formula; Expressed as the acceleration of the jth individual in the dth dimension; Step 385: Obtain the position and velocity of each individual in each dimension of the current population according to the following formula; represents the d-th dimension velocity of the j-th individual; Step 386: If o≤o max , then o=o+1, and go to step 382; otherwise, end the algorithm and output the local optimal approximate solution X′, judging that X′ is better than the current global optimal solution X best , then assign X′ to X best , and go to S34; otherwise go to S39.

8. An emergency order scheduling system based on an improved VNS-GSA algorithm, characterized in that: include: The acquisition module is used to obtain the order set and factory set to be processed; The order set includes a number of normal orders and urgent orders; A building module is used to solve the real cost of the factory producing the normal order based on the production deadline of each normal order to obtain the total revenue of the factory; A mathematical model is constructed with the goal of maximizing total revenue and limiting each factory to be assigned at most one emergency order; The solution module is used to solve the mathematical model by using the improved VNS-GSA algorithm to obtain the optimal emergency order scheduling solution.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the processor executes the computer program, the emergency order scheduling method based on the improved VNS-GSA algorithm as described in any one of claims 1 to 7 can be implemented.

10. An electronic device, characterized in that: include: one or more processors; Memory; And one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the programs include a method for executing the emergency order scheduling method based on the improved VNS-GSA algorithm as described in any one of claims 1 to 7 above.

Images