Multi-objective production order allocation device and multi-objective production order allocation method

Abstract

The present invention provides a multi-objective production order allocation device and a multi-objective production order allocation method, the purpose of which is to make the benefits of the entire group company reach global optimization. The multi-objective production order allocation device supplies products from multiple factories to multiple customers, and is equipped with: a data acquisition unit, which acquires factory data, logistics data and customer demand data; an allocation method determination unit, which determines the allocation method according to the customer demand data; a multi-objective optimization model establishment unit, which establishes a multi-objective optimization model including multiple objective functions and constraints according to the allocation method; and a calculation unit, which solves the multi-objective optimization model established by the multi-objective optimization model establishment unit based on the NSGA-Ⅱ algorithm.

Description

Technical Field

[0001] The invention relates to a multi-objective production order allocation device and a multi-objective production order allocation method based on an improved genetic algorithm. Background Art

[0002] In the context of group companies supplying products to multiple customers, it is often necessary to consider the overall profit maximization of multiple regional branches within the group. The allocation method of production orders has a very significant impact on production costs and efficiency. The quality of the allocation method is related to the benefits of the entire group company. Therefore, it is urgent to establish a multi-factory production order allocation optimization model at the group headquarters level, uniformly allocate resources and products of regional branches within the group company, and make full use of existing resources and equipment to improve the company's benefits based on market conditions and the production capacity of each factory within the group, so as to achieve the global optimal benefits of the entire group company.

[0003] In this regard, Patent Document 1 (CN107451679A) provides a cloud processing method for production order allocation. The method solves the objective function based on an immune algorithm to provide an order allocation method that minimizes production costs when a customer selects multiple different manufacturers for production.

[0004] However, the prior art such as Patent Document 1 has the following technical problems.

[0005] First, the existing technologies such as Patent Document 1 can only allocate one demand at a time, and cannot calculate the allocation of multiple demands or multiple orders at the same time. In actual manufacturing, there are production tasks that need to provide products to multiple customers within a cycle. The group headquarters needs to coordinate all production tasks at the same time and formulate allocation plans according to the different demand types of multiple customers to optimize the overall benefits of the cycle.

[0006] Secondly, the existing technologies such as Patent Document 1 only aim to optimize the cost of the purchaser, without taking into account the profit balance of multiple suppliers, that is, the profit balance of multiple factories within the group. Against the background of rising domestic labor costs and land costs, the fixed costs of factories have risen. When the group headquarters allocates production tasks, it is necessary to consider the balance of interests of each branch to avoid factories having no orders to do and affecting production enthusiasm.

[0007] In addition, the prior art such as Patent Document 1 only considers allocating a batch, and cannot obtain more specific allocation results, such as when and where a certain order is produced. In the production and manufacturing of large and medium-sized equipment, products often have certain customization requirements, and the cycle from design to final output is long; and each device contains many parts, and early production will form a large amount of parts, semi-finished products and finished products. Inventory, resulting in various risks such as capital backlog. Therefore, the allocation result that specifically considers when and where a certain order is produced is of great practical significance for avoiding the above-mentioned risks.

[0008] This shows that the existing technology cannot meet the global optimal efficiency of the entire group company, and a new production order allocation method is needed. Summary of the invention

[0009] The present invention aims at the above-mentioned technical problems in the prior art, and its purpose is to provide a multi-objective production order allocation device and a multi-objective production order allocation method that can achieve global optimal benefits for the entire group company.

[0010] One embodiment of the present invention provides a multi-objective production order allocation device, which is a multi-objective production order allocation device for supplying products from multiple factories to multiple customers, and is characterized in that it comprises: a data acquisition unit, which acquires factory data, logistics data and customer demand data, wherein the customer demand data is demand total amount information including at least the total amount of each demand of multiple customers or at least the order information including multiple orders corresponding to multiple customers; an allocation method determination unit, which determines the allocation method according to the customer demand data, and when the customer demand data is demand total amount information, the allocation method is determined as a first allocation method in which the total amount of each demand of multiple customers is respectively allocated to at least a part of the multiple factories; when the customer demand data is order information, the allocation method is determined as a second allocation method in which each of the multiple orders is respectively allocated to one of the multiple factories; a multi-objective optimization model establishment unit, which establishes a multi-objective optimization model corresponding to the first allocation method or the second allocation method according to the allocation method determined by the allocation method determination unit, and the multi-objective optimization model includes multiple objective functions and at least one constraint condition; and a calculation unit, which solves the multi-objective optimization model established by the multi-objective optimization model establishment unit based on the NSGA-Ⅱ algorithm, i.e., an improved non-dominated sorting genetic algorithm, so as to obtain a solution of the multi-objective optimization model as a production order allocation scheme.

[0011] As a result, the production order allocation can be optimized according to the different types of customer needs to meet the different needs of different customers. In addition, the allocation problem of multiple needs of multiple customers can be calculated at the same time, so that the overall benefit of the group company can reach the global optimum within a cycle.

[0012] The above-mentioned multi-objective production order allocation device can also be that the factory data acquired by the data acquisition unit includes at least the production capacity, production cost and break-even point of each of the multiple factories, and the logistics data acquired by the data acquisition unit includes at least the transportation cost between the multiple factories and the multiple customers; in the multi-objective optimization model established by the multi-objective optimization model establishment unit, the multiple objective functions include at least a cost objective function established according to the production cost, transportation cost and the allocation method determined by the allocation method determination unit, a break-even objective function established according to the break-even point and the allocation method determined by the allocation method determination unit, and a capacity utilization deviation objective function established according to the capacity and the allocation method determined by the allocation method determination unit, and at least one constraint condition is a constraint condition determined according to the factory data, logistics data and customer demand data.

[0013] In this way, both cost optimization and profit balance are taken into account, thereby optimizing the distribution of benefits among factories and enhancing production enthusiasm.

[0014] The above-mentioned multi-objective production order allocation device can also be that, in the multi-objective optimization model established by the multi-objective optimization model establishment unit, the cost objective function objective minimizes the sum of the production costs of multiple factories and the sum of the transportation costs from multiple factories to multiple customers; the break-even objective function objective makes the product quantity supplied from each of the multiple factories to multiple customers exceed the break-even point of each factory; the capacity utilization deviation objective function objective makes the deviation between the capacity utilization rates of each of the multiple factories not exceed the specified capacity utilization deviation upper limit, and the capacity utilization rate is the ratio of the product quantity supplied from a factory among the multiple factories to multiple customers to the capacity of the factory.

[0015] As a result, the order allocation plan accurately takes into account the overall cost optimization and profit balance of each factory, so that the overall benefits of the group company can reach the global optimum.

[0016] The above-mentioned multi-objective production order allocation device may also be that, in the multi-objective optimization model established by the multi-objective optimization model establishment unit, at least one constraint condition includes ensuring that the quantity of products supplied from each of the multiple factories to multiple customers does not exceed the production capacity of each factory.

[0017] As a result, the order allocation plan accurately defines the constraints determined by the production capacity of each factory, thereby improving the reliability of the order allocation plan.

[0018] The above-mentioned multi-objective production order allocation device may also be that, when the allocation method is determined to be the first allocation method, at least one constraint condition includes making the quantity of products supplied by multiple factories to a certain customer among multiple customers equal to the total demand of the customer.

[0019] As a result, the order allocation plan accurately defines the constraints determined according to customer needs, thereby improving the reliability of the order allocation plan.

[0020] The above-mentioned multi-objective production order allocation device may also be that the logistics data acquired by the data acquisition unit also includes the transportation time between multiple factories and multiple customers, the order information includes the agreed delivery time of multiple orders, and when the allocation method is determined to be the second allocation method, the second allocation method includes the allocation method of the production time of multiple orders, and the multiple objective functions also include a reverse production objective function established according to the transportation time, the agreed delivery time and the second allocation method.

[0021] As a result, the production time of multiple orders is further taken into account, so that the order allocation plan can enable multiple orders to be produced in a reverse production manner, that is, the actual delivery date is as close to the agreed delivery date as possible, thereby reducing various possible risks such as capital backlog caused by a large inventory of parts, semi-finished products and finished products.

[0022] The above-mentioned multi-objective production order allocation device may also be that, when the allocation method is determined to be the second allocation method, at least one constraint condition includes ensuring that a single order among multiple orders is produced only by a single factory among multiple factories, and ensuring that the actual delivery period obtained based on the transportation time and the second allocation method does not exceed the agreed delivery period.

[0023] As a result, the order allocation plan accurately defines the constraints determined according to customer needs, thereby improving the reliability of the order allocation plan.

[0024] The above-mentioned multi-objective production order allocation device can also be that, in the operation unit, a degree function is constructed for at least one constraint in the constraint conditions, and a penalty function is constructed according to the sum of the degree functions of all constraints and the penalty coefficient specified by the evolutionary algebra, and at least one constraint condition is processed by combining the penalty function with multiple objective functions, thereby converting the multi-objective production order allocation problem containing constraints into a multi-objective production order allocation problem without constraints.

[0025] Therefore, the operation based on NSGA-Ⅱ algorithm can correctly take the constraints into account and improve the feasibility of the operation.

[0026] The multi-objective production order allocation device may also be configured such that, in the operation unit, a predetermined genetic algebra is set, and the operation result after the predetermined genetic algebra is performed on the solution of the multi-objective optimization model based on the NSGA-II algorithm is used as the solution of the objective function.

[0027] Therefore, the optimal solution can be obtained based on the NSGA-Ⅱ algorithm with limited computational effort, thus improving the reliability of the order allocation scheme.

[0028] Another embodiment of the present invention provides a multi-objective production order allocation method, which is a multi-objective production order allocation method for supplying products from multiple factories to multiple customers, and is characterized by comprising: a data acquisition step, acquiring factory data, logistics data and customer demand data, wherein the customer demand data is demand total amount information including at least the total amount of each demand of multiple customers or at least order information including multiple orders corresponding to multiple customers; an allocation method determination step, determining the allocation method according to the customer demand data, when the customer demand data is demand total amount information, the allocation method is determined as a first allocation method in which the total amount of each demand of multiple customers is respectively allocated to at least a part of the multiple factories; when the customer demand data is order information, the allocation method is determined as a second allocation method in which each of the multiple orders is respectively allocated to one of the multiple factories; a multi-objective optimization model establishment step, establishing a multi-objective optimization model corresponding to the first allocation method or the second allocation method according to the allocation method determined in the allocation method determination step, the multi-objective optimization model including multiple objective functions and at least one constraint condition; and a calculation step, solving the multi-objective optimization model established in the multi-objective optimization model establishment step based on the NSGA-Ⅱ algorithm, i.e., an improved non-dominated sorting genetic algorithm, so as to obtain a solution of the multi-objective optimization model as a production order allocation scheme.

[0029] The above-mentioned specific methods and effects of the multi-target production order allocation method of the present invention can also be achieved through a multi-target production order allocation device, a multi-target production order allocation system, a multi-target production order allocation program that enables a computer to execute the multi-target production order allocation method, or a recording medium that stores the multi-target production order allocation program. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is a schematic diagram showing an overview of the multi-objective production order assignment problem.

[0031] Figure 2 1 is a block diagram showing the multi-objective production order assigning device 100 according to the present embodiment.

[0032] Figure 3A This is a schematic table showing an example of the data structure of the total demand information.

[0033] Figure 3B This is a schematic table showing an example of the data structure of order information.

[0034] Figure 4 1 is a flowchart showing a multi-objective production order allocating method executed by the multi-objective production order allocating device 100 according to the present embodiment.

[0035] Figure 5AThis is a schematic table showing an example of production cost of each factory in the first specific example of the embodiment.

[0036] Figure 5B This is a schematic table showing an example of transportation cost from each factory to each customer in the first specific example of the embodiment.

[0037] Figure 5C This is a schematic diagram showing an example of the production capacity and break-even point of each factory in the first specific example of the embodiment.

[0038] Figure 5D This is a schematic table showing an example of the total demand amount information of each customer in the first specific example of the implementation method.

[0039] Figure 6 This is a flowchart showing the steps of solving the multi-objective optimization model by the calculation unit 104 in the first specific example of the embodiment.

[0040] Figure 7 This is a schematic table showing an example of the allocation result obtained in the first specific example of the embodiment.

[0041] Fig. 8A This is a schematic diagram showing an example of the production capacity and break-even point of each factory in the second specific example of the embodiment.

[0042] Figure 8B This is a schematic table showing an example of the transportation time from each factory to each customer in the second specific example of the embodiment.

[0043] Figure 8C This is a schematic table showing an example of order information of a plurality of orders corresponding to a plurality of customers in the second specific example of the embodiment.

[0044] Fig. 9 This is a schematic table showing an example of the allocation result obtained in the second specific example of the embodiment. DETAILED DESCRIPTION

[0045] The present invention is described in more detail below in conjunction with the accompanying drawings, embodiments and specific examples. The following description is only an example given to facilitate understanding of the present invention and is not intended to limit the scope of the present invention. In a specific embodiment, the components of the device may be changed, deleted or added according to actual conditions, and the steps of the method may be changed, deleted, added or changed in order according to actual conditions.

[0046] First, an overview of the multi-objective production order allocation problem is explained. The multi-objective production order allocation problem refers to the allocation problem of producing products from multiple factories and supplying the products to multiple customers according to their respective needs. Since the production costs, production capacities, geographical locations, hardware facilities, and staffing of multiple factories are different, and the types, quantities, and delivery dates of products required by multiple customers are different, there are various allocation methods. The multi-objective production order allocation problem aims to select an allocation method that achieves the goal of keeping the total cost as low as possible, making each factory as profitable as possible, and maximizing the capacity utilization of each factory among the various allocation methods, while satisfying certain constraints. Figure 1 is a schematic diagram showing the outline of the multi-objective production order allocation problem. Figure 1 As shown, products are supplied from four factories, namely, Factory 1, Factory 2, Factory 3 and Factory 4, to three customers, namely, Customer A, Customer B and Customer C, in the following manner: Figure 1 Indicated by the arrow. Figure 1 An example of four factories and three customers is shown in FIG. 1 , but the specific numbers of factories and customers are not limited thereto.

[0047] The embodiment of the present invention will be described in detail below. First, a multi-objective production order allocation device 100 according to an embodiment of the present invention will be described. Figure 2 is a block diagram showing the multi-objective production order allocation device 100 of this embodiment, wherein arrows indicate the direction of data flow. Figure 2 As shown, the multi-objective production order allocation device 100 of the present embodiment comprises a data acquisition unit 101, an allocation method determination unit 102, a multi-objective optimization model establishment unit 103 and a calculation unit 104. Among them, the above-mentioned parts in the multi-objective production order allocation device 100 can be implemented in the form of software modules, or can also be implemented in the form of dedicated hardware integrated circuits. The multi-objective production order allocation device 100 can be implemented by a single computer or server, or by a plurality of separate devices, such as a cloud computing server composed of a plurality of servers. In addition, the functions of the above-mentioned parts of the multi-objective production order allocation device 100 can also be realized by executing a multi-objective production order allocation program pre-stored in a recording medium through a computer processor, thereby implementing the present embodiment as a multi-objective production order allocation device. The following specifically describes the various parts of the multi-objective production order allocation device 100.

[0048] The data acquisition unit 101 acquires factory data, logistics data and customer demand data. Figure 2As shown, the data acquisition unit 101 acquires factory data from the factory database 2, logistics data from the logistics database 3, and customer demand data from the customer demand database 4. Of course, the method of acquiring factory data, logistics data, and customer demand data is not limited to this. The data acquisition unit 101 may acquire factory data, logistics data, and customer demand data from the same database, or may acquire them through manual input by a user, or may acquire them from a storage unit (not shown) such as a portable storage medium.

[0049] The factory data acquired by the data acquisition unit 101 may include at least the production capacity, production cost and break-even point of each of the multiple factories. Production capacity refers to the maximum number of qualified products that can be produced by the factory within a unit period under the existing technology, equipment and personnel conditions. Depending on the specific allocation problem, it can be the factory capacity within different periods such as days, weeks, months and quarters; cost refers to the cost of producing a unit product, including material costs, manufacturing costs, packaging costs and management costs; the break-even point can be calculated in terms of amount or in kind. In this embodiment, it is calculated in terms of physical quantity, indicating the number of sales products that reach the factory's break-even point. The purpose of introducing this parameter is to fully consider the profits of each factory in the production allocation process.

[0050] The logistics data acquired by the data acquisition unit 101 may at least include the transportation costs between multiple factories and multiple customers. The transportation cost refers to the cost of transporting a unit product. If multiple products are transported in one batch, the cost needs to be allocated to each product.

[0051] In this embodiment, it is considered that there are two types of customer demand data acquired by the data acquisition unit 101, that is, the customer demand data is the total demand information including at least the total demand of multiple customers or at least the order information including multiple orders corresponding to multiple customers. The total demand information includes at least the total demand of each customer in the period. Figure 3A is a schematic diagram showing an example of the data structure of the total demand information. Figure 3A As shown, the total amount required by each customer in a cycle is different. For example, the total amount required by customer A in a cycle is 100 pieces. The cycle can be a day, week, month, quarter, etc. according to different products and needs. The order information at least includes the corresponding relationship between the numbers of multiple orders in the cycle and the customers to whom they belong. Figure 3B is a schematic diagram showing an example of the data structure of order information. Figure 3B As shown, for example, order O1 belongs to customer A.

[0052] The allocation method determination unit 102 determines the allocation method according to the customer demand data. That is, when the customer demand data is the total demand information, the allocation method is a first allocation method in which the total demand of each of the multiple customers is allocated to at least a part of the multiple factories. When the customer demand data is the order information corresponding to the multiple customers, the allocation method is a second allocation method in which each of the multiple orders is allocated to one of the multiple factories. The significance of determining two different allocation methods according to the above two types of customer demand data is that in the actual product production and sales process, there are usually two situations for customer demand. One situation is that a customer's demand needs to be supplied by the production of multiple factories. For example, when a customer needs a large number of products in a short period of time and the supply of one factory is difficult to meet its demand, the allocation method (first allocation method) of allocating the total demand of the customer to multiple factories for supply is applicable; the other situation is that a customer's demand (order) needs to be specified to be supplied by a factory. For example, when the product required by the customer has stricter dimensional tolerance requirements, and production by different factories may lead to many uncontrollable factors such as increased dimensional tolerance, the allocation method (second allocation method) in which a customer order is associated with only one factory is applicable. Therefore, this embodiment can determine different allocation formulas according to different types of customer needs, thereby meeting various needs of customers.

[0053] The allocation method determination unit 102 can determine the allocation method by, for example, a user manually judging the customer demand data; or it can determine the allocation method by, for example, a customer demand data judgment unit (not shown) built into the allocation method determination unit 102 and based on an automatic recognition algorithm, automatically judging the type of customer demand data.

[0054] The multi-objective optimization model building unit 103 builds a multi-objective optimization model corresponding to the first allocation method or the second allocation method according to the allocation method. The multi-objective optimization model in this embodiment includes multiple objective functions and at least one constraint condition. Thus, it is possible to build different multi-objective optimization models according to different allocation fractions, and optimize the production order allocation scheme in a targeted manner.

[0055] The multiple objective functions may include at least a cost objective function, a break-even objective function, and a capacity utilization rate deviation objective function. In this case, the multi-objective optimization model building unit 103 builds a cost objective function, a break-even objective function, and a capacity utilization rate deviation objective function corresponding to the first allocation method or the second allocation method.

[0056] The cost objective function is an objective function established based on the production cost and transportation cost acquired by the data acquisition unit 101 and the allocation method (first allocation method or second allocation method) determined by the allocation method determination unit 102. Its goal, for example, can be to minimize the sum of the production costs for multiple factories determined by the specific allocation method and the sum of the transportation costs between multiple factories and multiple customers, that is, to minimize the total cost. Thus, the goal is to minimize the total cost of the group company including all factories, and the goal is to achieve the most suitable interest distribution plan for the group company as a whole and multiple customers as a whole, that is, the allocation plan with the lowest total cost as possible. In addition, the goal of the cost objective function is not limited to this. According to the actual interest needs of the group company and the requirements of the customers, the goal setting of the cost objective function may be different. For example, the cost objective function can also be targeted at each of the multiple customers, and the goal is to minimize the maximum value of the costs including production costs and transportation costs required for each of the multiple customers to purchase from multiple factories. Thus, the goal is to achieve a distribution plan that satisfies each customer.

[0057] The break-even objective function is an objective function established based on the break-even point acquired by the data acquisition unit 101 and the allocation method (first allocation method or second allocation method) determined by the allocation method determination unit 102. Its goal, for example, can be to make the amount of products supplied from each of the multiple factories to multiple customers exceed the break-even point of each factory, that is, the actual production value of each factory is higher than the break-even point. Thus, the goal is to make the output value of each of the multiple factories as high as possible above the break-even point, and the goal is to achieve that multiple factories in the group company can make profits and balance interests, and there will be no situation where a factory has no orders to do, thereby improving production enthusiasm. In addition, the goal of the break-even objective function is not limited to this. According to the actual interests of the group company and the requirements of the customers, the goal setting of the break-even objective function may be different. For example, the goal of the break-even objective function can also be to minimize the maximum value of the difference between the break-even point of each of the multiple factories and the amount of products supplied to multiple customers, that is, to minimize the loss of the factory with the most serious loss (maximize the benefit). Thus, the goal is to ensure the interests of the factory with the worst benefits in the group company, and avoid the occurrence of unfavorable situations such as the factory closing down or going bankrupt due to losses.

[0058] The capacity utilization rate deviation objective function is an objective function established based on the capacity acquired by the data acquisition unit 101 and the allocation method (first allocation method or second allocation method) determined by the allocation method determination unit 102. Its goal, for example, can be to make the deviation between the capacity utilization rates of multiple factories not exceed the upper limit of the capacity utilization rate deviation, where the capacity utilization rate refers to the ratio of the amount of products supplied from a factory among multiple factories to multiple customers to the capacity of the factory, and the upper limit of the capacity utilization rate deviation is a constant pre-defined by the user, which can be set to any value between 0 and 1 according to the situation. Therefore, the goal is to make the deviation between the capacity utilization rates of factories as little as possible as not to exceed the specified upper limit, and optimize the balance of capacity utilization rates between factories. The balance of capacity utilization rate is of great significance to the geopolitical strategic development of the group company. For example, due to the higher transportation cost of the core customers in the western region of my country than in the eastern coastal region, the capacity utilization rate of the branch factories of the group company located in the western region may be low. By improving the balance of capacity utilization rate, the efficiency and production enthusiasm of the branch factories in the western region can be improved, which is of great positive significance for the group company to further explore potential customers in developing regions such as the western region. In addition, the goal of the capacity utilization rate deviation objective function is not limited to this. According to the actual interests of the group company and the requirements of customers, the goal setting of the capacity utilization rate deviation objective function can be different. For example, the goal of the capacity utilization rate deviation objective function can also be to minimize the maximum value of the capacity utilization rate deviation between two of the multiple factories. In this way, the goal is to avoid excessive imbalance in capacity utilization between factories.

[0059] The multi-objective optimization model building unit 103 establishes at least one constraint condition corresponding to the first allocation method or the second allocation method. The at least one constraint condition is a constraint condition determined by factory data, logistics data, and customer demand data. Specifically, the constraint condition is, for example, to make the allocation plan meet production capacity constraints, meet customer needs, meet delivery deadline constraints, etc. Depending on the specific situation, the constraint condition can be one or more. By setting the constraint condition, the reliability of the allocation plan can be ensured.

[0060] The operation unit 104 solves the multi-objective optimization model established by the multi-objective optimization model establishment unit based on the NSGA-Ⅱ algorithm, i.e., the improved non-dominated sorting genetic algorithm, so as to obtain a solution that satisfies the constraint conditions as a production order allocation scheme. Specifically, when the allocation method determined by the allocation method determination unit 102 is the first allocation method, the solution at least includes the production quantity of each factory for the total demand of each of the multiple customers; when the allocation method determined by the allocation method determination unit 102 is the second allocation method, the solution at least includes the corresponding relationship between each of the multiple orders and the multiple factories.

[0061] NSGA-Ⅱ (Non-dominated Sorting Genetic Algorithm-Ⅱ) is a multi-objective genetic algorithm. It has parallelism and does not require manual determination of the weights of each objective. It can find multiple Pareto optimal solutions in one run, and then the user can make judgments and choices based on the actual situation. Therefore, it is an effective method for dealing with multi-objective optimization problems. In addition, compared with other algorithms such as NSGA, NSGA-Ⅱ has the following advantages: it adopts fast non-dominated sorting to reduce the complexity of algorithm calculation; it proposes crowding and crowding comparison operators to replace the fitness sharing strategy that requires the specified sharing radius, and uses them as the winning criteria in the same-level comparison after fast sorting, so that individuals in the quasi-Pareto domain can be expanded to the entire Pareto domain and evenly distributed, maintaining the diversity of the population; it introduces an elite strategy to expand the sampling space, combine the parent population with the child population it generates, and compete to produce the next generation population, which is conducive to keeping the excellent individuals in the parent generation into the next generation, and by storing all individuals in the population in layers, the best individuals will not be lost, and the population level will be rapidly improved.

[0062] In this embodiment, the specific allocation scheme according to the allocation method (the first allocation method or the second allocation method) is used as a decision variable, and the chromosome in the NSGA-Ⅱ algorithm is composed of the decision variables and multiple objective functions. The specific chromosome encoding method is not limited, as long as the decision variables and objective functions are used as genes in the chromosome. In addition, since the traditional NSGA-Ⅱ algorithm only solves unconstrained multi-objective optimization problems, and the multi-objective optimization model of this embodiment involves constraints, it is necessary to convert the self-constrained multi-objective programming problem into a multi-objective programming problem without constraints. For example, a penalty function can be constructed according to all constraints in at least one constraint and the penalty coefficient specified by the evolutionary algebra, and a new objective function is constructed as the fitness function of the NSGA-Ⅱ algorithm by combining the constructed penalty function with multiple objective functions to process the constraints. Afterwards, according to the general process of the NSGA-Ⅱ algorithm, an initial population of a certain size is randomly generated, non-dominated sorting and crowding calculation are performed, and genetic operations are performed to generate offspring populations. After iterating several generations of populations, the final result is output according to the predetermined termination conditions, and the solution corresponding to the individual is the optimal solution as the production order allocation scheme. The termination condition is pre-set by the user according to actual needs. For example, a specified genetic generation can be set, and the calculation result after the specified genetic generation is taken as the optimal solution. In addition, the termination condition is not limited to this. For example, a specified convergence condition can also be set, and the calculation result after the individual difference between two consecutive generations of populations is less than a specified fixed value is taken as the optimal solution.

[0063] Thus, the multi-objective production order allocation device 100 of the embodiment of the present invention can output the final result as the production order allocation plan for the first allocation method or the second allocation method. Through one operation, the allocation problem of multiple demands of multiple customers is solved, so that the overall interests of the group company are globally optimized.

[0064] The following describes a multi-objective production order allocation method executed by the multi-objective production order allocation device 100 of this embodiment. Figure 4 1 is a flowchart showing a multi-objective production order allocating method executed by the multi-objective production order allocating device 100 according to the present embodiment. Figure 4 The process shown is executed based on one request from the user.

[0065] First, in the data acquisition step S101, the data acquisition unit 101 acquires factory data, logistics data, and customer demand data. The factory data at least includes the production capacity, production cost, and break-even point of each of the multiple factories; the logistics data at least includes the transportation cost between the multiple factories and the multiple customers; the customer demand data is the total demand information including the total demand of each of the multiple customers or the order information including multiple orders corresponding to the multiple customers. The process proceeds to step S102.

[0066] In the allocation method determination step S102, the allocation method determination unit 102 determines the allocation method based on the customer demand data. When the customer demand data acquired in the data acquisition step S101 is the total demand information, the allocation method is determined as a first allocation method that allocates the total demand of each of the plurality of customers to at least a part of the plurality of factories. The flow proceeds to step S103.

[0067] Next, in the multi-objective optimization model establishment step S103, the multi-objective optimization model establishment unit 103 establishes a multi-objective optimization model corresponding to the first allocation method. The multi-objective optimization model includes multiple objective functions and at least one constraint condition. The multiple objective functions may at least include a cost objective function established based on the production cost and transportation cost obtained in the data acquisition step S101 and the first allocation method determined in the allocation method determination step S102; a break-even objective function established based on the break-even point obtained in the data acquisition step S101 and the first allocation method determined in the allocation method determination step S102; and a capacity utilization rate deviation objective function established based on the capacity obtained in the data acquisition step S101 and the first allocation method determined in the allocation method determination step S102. The constraint condition may be a constraint condition determined based on the factory data, logistics data and total demand information obtained in the data acquisition step S101. The process proceeds to step S105.

[0068] Next, in the operation step S105, the operation unit 104 solves the multi-objective optimization model established in the multi-objective optimization model establishment step S103 based on the NSGA-Ⅱ algorithm, i.e., the improved non-dominated sorting genetic algorithm, so as to obtain a solution including at least the production quantity of each factory for the total demand of each of the multiple customers as the production order allocation plan. The process ends.

[0069] Furthermore, in the allocation method determination step S102, when the customer demand data acquired in the data acquisition step S101 is order information corresponding to a plurality of customers, the allocation method is determined to be a second allocation method of respectively allocating each of the plurality of orders to one of the plurality of factories. The flow proceeds to step S104.

[0070] Next, in the multi-objective optimization model establishment step S104, the multi-objective optimization model establishment unit 103 establishes a multi-objective optimization model corresponding to the second allocation method. The multi-objective optimization model includes multiple objective functions and at least one constraint condition. The multiple objective functions may at least include a cost objective function established based on the production cost and transportation cost obtained in the data acquisition step S101 and the second allocation method determined in the allocation method determination step S102; a break-even objective function established based on the break-even point obtained in the data acquisition step S101 and the second allocation method determined in the allocation method determination step S102; and a capacity utilization rate deviation objective function established based on the capacity obtained in the data acquisition step S101 and the second allocation method determined in the allocation method determination step S102. The constraint condition may be a constraint condition determined based on the factory data, logistics data and order information obtained in the data acquisition step S101. The process proceeds to step S106.

[0071] Next, in the operation step S106, the operation unit 104 solves the multi-objective optimization model established in the multi-objective optimization model establishment step S104 based on the NSGA-II algorithm, thereby obtaining a solution including at least the corresponding relationship between the multiple orders and the multiple factories as the production order allocation solution. The process ends.

[0072] By executing the above multi-objective production order allocation method, the production order allocation can be optimized according to different customer demand types to meet the different needs of different customers. In addition, the allocation problem of multiple demands of multiple customers can be calculated at the same time, so that the overall benefit of the group company can reach the global optimum within one cycle.

[0073] (First Specific Example of Implementation)

[0074] The following describes a first specific example of the implementation method. Figure 2The multi-objective production order allocation device 100 shown is a specific example in which the allocation method is the first allocation method. The multi-objective production order allocation method executed in the first specific example will be described in detail below.

[0075] In the first specific example of the implementation method, as an example, consider the case where products are supplied from four factories (F1 to F4) to 30 customers (C1 to C30) in one cycle. Of course, the number of factories and customers can also be other values. First, the data acquisition unit 101 acquires the cost of each unit product produced by the four factories. Figure 5A : is a schematic table showing an example of production costs of each factory in the first specific example of the implementation method. Figure 5A As shown, in this specific example, it is considered that the production cost of each unit product of each factory is fixed. In addition, the data acquisition unit 101 acquires the transportation cost of transporting unit products between the four factories and the 30 customers. Figure 5B : is a schematic table showing an example of the transportation cost from each factory to each customer in the first specific example of the implementation method. Figure 5B As shown, in this specific example, the transportation cost of transporting a unit product from each factory to each customer is considered to be fixed. For example, a batch of products supplied from factory f to customer j is transported in one cycle, and the transportation cost is allocated to each product, which is the data in the table. In addition, the data acquisition unit 101 obtains the production capacity and break-even point of each of the four factories in the cycle. Figure 5C : is a schematic diagram showing an example of the production capacity and break-even point of each factory in the first specific example of the implementation method. Figure 5C As shown, in this specific example, the physical quantity (pieces) of products produced within a cycle is used as a parameter for determining production capacity and profit and loss balance. The quantity of products produced by each factory within a cycle cannot exceed the data required by production capacity, and the number of products produced exceeds the number of break-even, which means profit. In addition, the data acquisition unit 101 obtains the total demand of each of the 30 customers. Figure 5D : is a schematic table showing an example of the total demand information of each customer in the first specific example of the implementation method. Figure 5D As shown, in this specific example, the total demand information is the number of products required by each customer during the period.

[0076] Next, the allocation method determination unit 102 determines the allocation method as a first allocation method that allocates the total demand of each of the 30 customers to at least a portion of the four factories, i.e., an allocation method in which the demand of one customer can be supplied by one or more factories, based on the total demand obtained by the data acquisition unit 101. In this specific example, for example, the user manually determines the information obtained by the data acquisition unit 101 and finds that the data contains the total demand of each of the 30 customers, so the first allocation method is selected.

[0077] Next, the multi-objective optimization model building unit 103 builds a multi-objective optimization model that matches the first allocation method determined by the allocation method determination unit 102. For the convenience of explanation, the parameters in the model are defined as follows: define a set of factories F, where F = {1, 2, ...f, ...}, and the number of elements in F is 4 in this specific example; define a set of customers J, where J = {1, 2, ...j, ...}, and the number of elements in J is 30 in this specific example; define the production cost W of a unit product produced by factory f f , where W f ≥0; defines the quantity Q of products produced by factory f and supplied to customer j jf , where Q jf ≥0 and is an integer, Q jf That is, the decision variables in this specific example are determined according to the first allocation method; define the unit product transportation cost V from factory f to customer j jf , where V jf ≥0; defines the capacity C of factory f within the period f , where C f ≥0; defines the break-even point S of factory f within the period f , where S f ≥0; define the demand d of customer j in the period j , where d j ≥0; defines the capacity utilization rate U of factory f during the period f ; The upper limit k of the capacity utilization deviation between any two factories, where 0≤k≤1.

[0078] Hereinafter, the method for establishing multiple objective functions in the first specific example of the implementation mode will be specifically described with reference to the above parameters. In this specific example, the cost objective function z 1 The decision variable Q jf As a function of the total cost, we calculate the total cost as a function of the variables. Specifically, we consider the quantity Q of products produced by factory f and supplied to customer j as jf and the corresponding production cost W f and transportation cost V jf The sum is multiplied and summed for all combinations of factories and customers to obtain the total cost. Cost objective function z 1 The goal is to make this total cost as low as possible. In this specific example, the break-even objective function z 2 The decision variable Q jf Specifically, the break-even point S of each factory f is used as a variable to calculate the maximum loss function. f The number of products supplied by the factory to all customers is ∑ j∈J Q jfPerform subtraction and find the maximum value for all factories as the maximum loss. Break-even objective function z 2 The goal is to minimize the maximum loss. In this specific example, the capacity utilization deviation objective function z 3 The decision variable Q jf The function of the total deviation of capacity utilization is calculated as a variable. Specifically, the capacity utilization ratio U between each factory f is f The absolute value of the difference between the two is subtracted from the upper limit k of the capacity utilization deviation, and the sum is calculated for all pairs of factories to obtain the total capacity utilization deviation. f means∑ j∈J Q jf With the factory capacity C f The ratio of capacity utilization rate deviation objective function z 3 The goal is to make the total amount of capacity utilization deviation as small as possible.

[0079] The following describes the constraint conditions in the first specific example of the implementation method. First, as shown in the following formula (1), the amount of products produced by factory f and supplied to multiple customers must not exceed the production capacity C of factory f. f The number of constraints is equal to the number of factories, which is 4 in this specific example.

[0080]

[0081] And, as shown in the following formula (2), the sum of the number of products produced by each factory and supplied to customer j should satisfy the demand d of customer j j The number of constraints is equal to the number of customers, which is 30 in this specific example.

[0082]

[0083] Next, the calculation unit 104 solves the multi-objective optimization model established by the multi-objective optimization model establishment unit 103 as described above based on the NSGA-II algorithm. Figure 6 1 is a flowchart showing the steps of solving the multi-objective optimization model by the operation unit 104 in the first specific example of the embodiment. Figure 6 The solution steps are explained in detail.

[0084] In step S201, integer coding is used for coding design, and the chromosome is represented by the decision variable Q jf and the above multiple objective functions. For the order allocation problem of j customers and f factories, the first j*f genes of each chromosome are decision variables, and the following l genes are the values ​​of the objective function, where l corresponds to the number of objective functions. That is, chromosome P i = {Q1 , Q 2 , ..., Q j*f , z 1 , ..., z l}, i = 1, 2, ... N, where i is an integer from 1 to N, and N represents the population size. In this specific example, for the above order allocation problem of 30 customers and 4 factories, the first 30*4 genes of each chromosome are decision variables, and the last 3 genes are cost objective function z 1 , break-even objective function z 2 and capacity utilization deviation objective function z 3 The value of .

[0085] In step S202, constraint processing is performed by constructing a penalty function. First, as shown in the following formula (3), the constraint conditions described in the above formulas (1) and (2) are converted into a standard form.

[0086]

[0087]

[0088] And, construct the function of the following formula (4):

[0089]

[0090]

[0091] Next, for the above order allocation problem with j customers and f factories, there are a total of j+f constraints, and these constraints are sorted. For individual q in the group, the degree of violation of the mth constraint can be expressed in the form of the following formula (5):

[0092]

[0093] The penalty function of individual q is constructed as shown in the following formula (6):

[0094]

[0095] Among them, (Ct) α is the penalty coefficient, which increases with the evolutionary generation, t is the evolutionary generation, C and α are positive constants.

[0096] Then, the penalty function is combined with the original objective function min z l (q) is combined to form a new objective function as the fitness function of the NSGA-Ⅱ algorithm.

[0097] In step S203, an initial population of a certain size is randomly generated to initialize the population. Specifically, the gene Q as a decision variable for each individual in the initial population is randomly generated. jf , and gene Q jf Substitute each objective function into the gene to obtain the value of the objective function.

[0098] In step S204, the initial population is non-dominated sorted, and the crowding degree of each individual in each non-dominated layer is calculated.

[0099] In step S205, it is specified that the population generation number Gen=1, that is, the initial population is the first generation.

[0100] In step S206, genetic operations including selection, crossover, and mutation are performed. Specifically, the binary tournament method that considers both fitness value and crowding degree is used to select the tth generation population P. t Select N / 2 individuals from the mating pool and put them into the mating pool. Perform simulated binary crossover and polynomial mutation operations on the individuals in the mating pool to generate a progeny population Q t .

[0101] In step S207, by merging P t and Q t , construct the intermediate generation population R t .

[0102] In step S208, the intermediate generation population R t , perform non-dominated sorting and calculate the crowding degree of each individual in each non-dominated layer.

[0103] In step S209, N individuals are selected using the elite strategy to form a new generation population P t+1 .

[0104] In step S210, the termination condition is determined. Specifically, it is determined whether the current population generation number Gen is less than a preset value. If Gen < the preset value (S210 is "yes"), step S211 is executed, the population generation number Gen increases by one generation, that is, Gen = Gen + 1, and the process returns to step S206. If Gen ≥ the preset value (S210 is "no"), the operation of the operation unit 104 is terminated, and the solution corresponding to the individual is output as the optimal solution to the order allocation problem. The solution includes the production quantity of each factory for the total demand of multiple customers, that is, Q jf The value of .

[0105] Figure 7 is a schematic table showing an example of the allocation result obtained in the first specific example of the implementation method. According to the relevant data of the above four factories and 30 customers acquired by the data acquisition unit 101, the following is obtained: Figure 7The specific allocation result of the global optimum is shown in Figure 7 In , the value represents the number of products supplied from each factory to each customer. The total demand of a customer can be supplied by multiple factories or only by one factory. Figure 7 The distribution scheme according to the first distribution method shown makes the group company's benefits reach the global optimal

[0106] (Second specific example of implementation method)

[0107] Next, a second specific example of the implementation method is described. Figure 2 The multi-objective production order allocation device 100 shown is a specific example when the allocation method is the second allocation method. The multi-objective production order allocation method executed in the first specific example is described in detail below. In addition, for the parts in the second specific example that are the same as the first specific example, the specific description is omitted or simplified.

[0108] In the second specific example of the implementation method, as an example, half a month (14 days) is taken as a cycle, and the production distribution of 40 orders (O1 to O40) from 4 customers (C1 to C4) in 4 factories (F1 to F4) within half a month is considered, wherein each order requires production in only one factory. Among them, the 4 customers (C1 to C4) are the same as the customers (C1 to C4) among the 30 customers in the first specific example, and the 4 factories (F1 to F4) are the same as the 4 factories (F1 to F4) in the first specific example, that is, the relevant production costs and transportation costs are the same as those in the first specific example. Figure 5A and Figure 5B The corresponding data in the data acquisition unit 101 is the same as the corresponding data in the data acquisition unit 101. In addition, the data acquisition unit 101 obtains the production capacity and break-even point of each of the four factories in the half month. Fig. 8A : is a schematic diagram showing an example of the production capacity and the break-even point of each factory in the second specific example of the embodiment. In addition, the data acquisition unit 101 acquires the transportation time (number of days) from each of the four factories to the four customers. Figure 8B This is a schematic table showing an example of the transportation time from each factory to each customer in the second specific example of the embodiment. Figure 8B The value in represents how many days it takes to transport a product from a certain factory to a certain customer. In addition, the data acquisition unit 101 acquires order information of a plurality of orders corresponding to the plurality of customers. Figure 8C : is a schematic table showing an example of order information of a plurality of orders corresponding to a plurality of customers in the second specific example of the implementation method. Figure 8CAs shown, the order information includes the corresponding relationship between each of the 40 orders (O1~O40) and the customers (C1~C4), as well as the remaining delivery days from the day when each of the 40 orders is executed to the delivery date specified by the customer. In this specific example, consider the situation that each order can be completed within one day and shipped to the corresponding customer the next day after the completion of production. In addition, in this specific example, for the sake of simplicity, consider the situation that each of the 40 orders (O1~O40) contains only one product. Of course, each order can also contain multiple products. In this case, it is only necessary for the data acquisition unit 101 to obtain the number of products contained in each order and perform adaptive calculations.

[0109] Next, the allocation method determination unit 102 determines the second allocation method of allocating each of the 40 orders to one of the four factories based on the order information obtained by the data acquisition unit 101, that is, the allocation method in which a customer's order demand can only be supplied by one factory.

[0110] Next, the multi-objective optimization model building unit 103 builds a multi-objective optimization model matching the second allocation method determined by the allocation method determining unit 102. For the parameters in the model, the F, J, and W defined in the first specific example are used. f 、V jf In addition, we further define the order set O: O = {1, 2, ..., o, ...}, in this specific example, the number of elements in O is 40; define the set of dates within the period D: D = {1, 2, ..., d, ...}, in this specific example, the number of elements in O is the number of days in the half month, that is, 14; define the relationship between orders and customers B oj , B oj It is a value that is either 1 or 0, 1 means that order o belongs to customer j, 0 means it does not belong to customer j; defines the agreed delivery date m of order o o , that is, on which day after the day of allocation, the delivery to the corresponding customer is required (including transportation time); define the transportation time T from factory f to customer j jf (days), where T jf ≥0 and is an integer; defines the daily production capacity C of factory f fd (Taiwan / day), among which C fd ≥0; define the daily break-even point S of factory f fd (Taiwan / day), where S fd ≥0; parameter z that defines whether order o is produced in factory f on day d ofd , z ofd It is a value that is either 1 or 0, 1 means production, 0 means no production, ofd As the decision variable in this specific example.

[0111] The following is a detailed description of the method for establishing multiple objective functions in the second specific example of the implementation method with reference to the above parameters. In this specific example, the cost objective function is to convert the decision variable z ofd As a function of the total cost calculated as a variable. Specifically, for all orders (O1 to O40), according to z ofd Determine the production plant f for each order o, thereby obtaining the production cost of the plant f, and summing up all the plants to obtain the total production cost; according to z ofd and B oj Get the production plant f and the corresponding customer j for each order o, and get the total transportation cost based on the transportation cost from the plant f to the customer j. Add the total production cost and the total transportation cost to get the total cost. The goal of the cost objective function is to make the total cost as low as possible. In this specific example, the break-even objective function is to set the decision variable z ofd The function of calculating the maximum loss as a variable. Specifically, according to the daily break-even point S fd The total break-even point of factory f in half a month is obtained by summing up all the time, and according to z ofd Get the total number of orders supplied by factory f, subtract the total breakeven point from the total number of orders, and take the maximum value of the breakeven point for all factories as the maximum loss. The goal of the breakeven objective function is to make the maximum loss as small as possible. In this specific example, the capacity utilization deviation objective function is to convert the decision variable z ofd The function of calculating the maximum deviation of capacity utilization as a variable. Specifically, according to z ofd Get the total number of orders generated by factory f on day d, which is equal to C fd The ratio of is taken as the capacity utilization rate of factory f on the dth day, and the maximum value of the difference in capacity utilization between each factory is taken within the period as the maximum capacity utilization deviation. The goal of the capacity utilization deviation objective function is to make the maximum capacity utilization deviation as small as possible.

[0112] In addition, in this specific example, the multiple objective functions also include a reverse production objective function. The purpose of setting the reverse production objective function is to make the time when each order o is delivered to the customer as close as possible to its delivery date, so as to avoid the production of products being piled up in the factory. The reverse production objective function is to convert the decision variable z ofd The function that calculates the maximum value of the difference between the agreed delivery date and the actual delivery date of each order o as a variable. Specifically, according to z ofd and B oj Get the production date d, production factory f and corresponding customer j of each order o, so as to calculate the production date d and transportation time T jf Calculate the actual delivery date and convert the agreed delivery date m oThe difference between the agreed delivery date and the actual delivery date is the maximum value for all orders. The goal of the reverse production scheduling objective function is to make the maximum value as small as possible. In addition, the goal of the reverse production scheduling objective function is not limited to this. For example, the goal can also be to minimize the sum of the differences between the agreed delivery date and the actual delivery date of all orders, as long as the goal is achieved in a certain sense.

[0113] The following describes the constraints in the second specific example of the implementation method. First, as shown in the following formula (7), the total amount of orders produced by the capacity constraint factory f on the dth day cannot exceed the capacity of factory f on that day. The number of constraints is equal to the product of the number of factories and the number of days in the cycle. In this specific example, it is 4*14, that is, 56.

[0114]

[0115] Secondly, as shown in the following formula (8), a single order is indivisible, that is, an order is produced by only one factory on a specific date. The number of this constraint is equal to the total number of orders, which is 40 in this specific example.

[0116]

[0117] Secondly, as shown in the following formula (9), the actual delivery date of each order shall not exceed the agreed delivery date, that is, the delivery on schedule is met. The number of this constraint is equal to the total number of orders, which is 40 in this specific example.

[0118] z ofd B oj (d+T fj )≤M o Formula (9)

[0119] Next, the operation unit 104 solves the multi-objective optimization model established by the multi-objective optimization model establishment unit 103 based on the NSGA-II algorithm. Figure 6 The process shown in is the same, and the solution steps will not be repeated here. In this specific example, for the above order allocation problem of 40 orders from 4 customers produced in 4 factories, the first 4*40*4 genes of each chromosome are decision variables, and the following 4 genes are the values ​​of the cost objective function, the break-even objective function, the capacity utilization deviation objective function and the reverse production objective function. In addition, similar to the first specific example, for the above three constraints in this specific example, their forms are standardized and a penalty function is constructed, thereby converting the multi-objective production order allocation problem with constraints into a multi-objective production order allocation problem without constraints. Therefore, the operation unit 104 outputs the solution corresponding to the individual based on the NSGA-Ⅱ algorithm as the optimal solution to the order allocation problem. The solution includes the corresponding relationship between multiple orders and multiple factories and the production time, that is, z ofd The value of .

[0120] Fig. 9 is a schematic table showing an example of the allocation result obtained in the second specific example of the implementation method. According to the 40 orders of the above 4 customers and the relevant data of the 4 factories obtained by the data acquisition unit 101, the following is obtained: Fig. 9 The specific allocation result of the global optimum is shown in Fig. 9 In the table, it is shown which factory produces each order on which day, and the specific actual delivery date and agreed delivery date information are shown. In this way, the group company can achieve the global optimal benefit.

[0121] The specific embodiments and specific examples of the present invention are described above with reference to the accompanying drawings. Among them, the specific embodiments and specific examples described above are only specific examples of the present invention, which are used to understand the present invention and are not used to limit the scope of the present invention. Those skilled in the art can make various deformations, combinations and reasonable omissions of elements to the specific embodiments and specific examples based on the technical ideas of the present invention, and the resulting methods are also included in the scope of the present invention. For example, the above-mentioned embodiments and specific examples can be combined with each other, and the embodiments formed by the combination are also included in the scope of the present invention.

Claims

1. A multi-target production order allocation device, which is a multi-target production order allocation device for supplying products from multiple factories to multiple customers, characterized in that: have: A data acquisition unit, which acquires factory data, logistics data and customer demand data, wherein the customer demand data is demand total amount information including at least the total amount of demand of each of the plurality of customers or order information including at least a plurality of orders corresponding to the plurality of customers; an allocation method determining unit, which determines an allocation method according to the customer demand data, and when the customer demand data is the total demand information, the allocation method is determined to be a first allocation method in which the total demand of each of the plurality of customers is respectively allocated to at least a part of the plurality of factories; In the case where the customer demand data is the order information, the allocation method is determined to be a second allocation method in which each of the multiple orders is allocated to one of the multiple factories; a multi-objective optimization model establishment unit, which establishes a multi-objective optimization model corresponding to the first allocation method or the second allocation method according to the allocation method determined by the allocation method determination unit, wherein the multi-objective optimization model includes a plurality of objective functions and at least one constraint condition; as well as The operation unit solves the multi-objective optimization model established by the multi-objective optimization model establishment unit based on the NSGA-Ⅱ algorithm, i.e., an improved non-dominated sorting genetic algorithm, so as to obtain a solution of the multi-objective optimization model as a production order allocation plan, The factory data acquired by the data acquisition unit at least includes the production capacity, production cost and break-even point of each of the multiple factories, and the logistics data acquired by the data acquisition unit at least includes the transportation cost between the multiple factories and the multiple customers; In the multi-objective optimization model established by the multi-objective optimization model establishment unit, the multiple objective functions include at least a cost objective function established according to the production cost, the transportation cost and the allocation method determined by the allocation method determination unit, a break-even objective function established according to the break-even point and the allocation method determined by the allocation method determination unit, and a capacity utilization rate deviation objective function established according to the capacity and the allocation method determined by the allocation method determination unit, and the at least one constraint condition is a constraint condition determined according to the factory data, the logistics data and the customer demand data. In the multi-objective optimization model established by the multi-objective optimization model establishment unit, the cost objective function objective minimizes the sum of the sum of the production costs of the plurality of factories and the sum of the transportation costs from the plurality of factories to the plurality of customers; The break-even objective function aims to ensure that the quantity of products supplied from each of the multiple factories to the multiple customers exceeds the break-even point of each factory respectively; the capacity utilization deviation objective function aims to ensure that the deviation between the capacity utilization rates of each of the multiple factories does not exceed the specified upper limit of the capacity utilization deviation, and the capacity utilization rate is the ratio of the quantity of products supplied from a factory among the multiple factories to the multiple customers to the capacity of the factory.

2. The multi-objective production order allocation device according to claim 1, characterized in that: In the multi-objective optimization model established by the multi-objective optimization model establishment unit, the at least one constraint condition includes ensuring that the quantity of products supplied from each of the multiple factories to the multiple customers does not exceed the production capacity of each factory.

3. The multi-objective production order allocation device according to claim 1, characterized in that: When the allocation method is determined to be the first allocation method, the at least one constraint condition includes making the quantity of products supplied by the multiple factories to a certain customer among the multiple customers equal to the total demand of the customer.

4. The multi-objective production order allocation device according to claim 1, characterized in that: The logistics data acquired by the data acquisition unit also includes the transportation time between the multiple factories and the multiple customers, and the order information includes the agreed delivery date of each of the multiple orders. When the allocation method is determined to be the second allocation method, the second allocation method includes the allocation method of the production time of each of the multiple orders, and the multiple objective functions also include a reverse production objective function established based on the transportation time, the agreed delivery period and the second allocation method.

5. The multi-objective production order allocation device according to claim 4, characterized in that: When the allocation method is determined to be the second allocation method, the at least one constraint condition includes ensuring that a single order among the multiple orders is produced only by a single factory among the multiple factories, and ensuring that the actual delivery period obtained based on the transportation time and the second allocation method does not exceed the agreed delivery period.

6. The multi-objective production order allocation device according to any one of claims 1 to 5, characterized in that: In the operation unit, a penalty function is constructed based on all the constraints in the at least one constraint and the penalty coefficient specified by the evolutionary algebra, and the at least one constraint is processed by combining the penalty function with the multiple objective functions, thereby converting the multi-objective production order allocation problem containing constraints into a multi-objective production order allocation problem without constraints.

7. The multi-objective production order allocation device according to claim 6, characterized in that: In the calculation unit, a predetermined genetic generation number is set, and a calculation result after performing the predetermined genetic generation number on the solution of the multi-objective optimization model based on the NSGA-II algorithm is used as the solution.

8. A multi-objective production order allocation method for supplying products from multiple factories to multiple customers, characterized in that: have: A data acquisition step, acquiring factory data, logistics data and customer demand data, wherein the customer demand data is demand total amount information including at least the total amount of demand of each of the plurality of customers or order information including at least a plurality of orders corresponding to the plurality of customers; an allocation method determination step, determining an allocation method according to the customer demand data, wherein when the customer demand data is the total demand information, the allocation method is determined to be a first allocation method in which the total demand of each of the plurality of customers is respectively allocated to at least a part of the plurality of factories; In the case where the customer demand data is the order information, the allocation method is determined to be a second allocation method in which each of the multiple orders is allocated to one of the multiple factories; a multi-objective optimization model establishment step, establishing a multi-objective optimization model corresponding to the first allocation method or the second allocation method according to the allocation method determined in the allocation method determination step, wherein the multi-objective optimization model includes a plurality of objective functions and at least one constraint condition; as well as A calculation step, based on the NSGA-Ⅱ algorithm, i.e., an improved non-dominated sorting genetic algorithm, solves the multi-objective optimization model established in the multi-objective optimization model establishment step, thereby obtaining a solution of the multi-objective optimization model as a production order allocation plan, The factory data acquired by the data acquisition step at least includes the production capacity, production cost and break-even point of each of the multiple factories, and the logistics data acquired by the data acquisition step at least includes the transportation cost between the multiple factories and the multiple customers; In the multi-objective optimization model established by the multi-objective optimization model establishment step, the multiple objective functions at least include a cost objective function established according to the production cost, the transportation cost and the allocation method determined by the allocation method determination step, a break-even objective function established according to the break-even point and the allocation method determined by the allocation method determination step, and a capacity utilization rate deviation objective function established according to the capacity and the allocation method determined by the allocation method determination step, and the at least one constraint condition is a constraint condition determined according to the factory data, the logistics data and the customer demand data. In the multi-objective optimization model established in the multi-objective optimization model establishment step, the cost objective function objective minimizes the sum of the sum of the production costs of the plurality of factories and the sum of the transportation costs from the plurality of factories to the plurality of customers; The break-even objective function aims to ensure that the quantity of products supplied from each of the multiple factories to the multiple customers exceeds the break-even point of each factory respectively; the capacity utilization deviation objective function aims to ensure that the deviation between the capacity utilization rates of each of the multiple factories does not exceed the specified upper limit of the capacity utilization deviation, and the capacity utilization rate is the ratio of the quantity of products supplied from a factory among the multiple factories to the multiple customers to the capacity of the factory.

Images