Uncertain demand joint replenishment method based on hybrid intelligent algorithm

Abstract

The invention belongs to the technical field of joint replenishment, and particularly relates to a joint replenishment method for uncertain requirements based on a hybrid intelligent algorithm. The method comprises the following steps: step 1, constructing a joint replenishment model oriented to dynamic uncertain requirements; and 2, designing a joint replenishment model solution algorithm based on a hybrid intelligent algorithm. According to the method, the genetic algorithm and the ant colony algorithm are adopted for hybrid solving, the defects that the genetic algorithm is prone to falling into a local optimal solution and the ant colony algorithm is low in precision are overcome, and the efficiency and precision of solving the joint replenishment problem are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of joint replenishment, and in particular relates to a joint replenishment method for uncertain demand based on a hybrid intelligent algorithm. Background Art

[0002] The demand at the frontline demand points is dynamic and uncertain. On the one hand, the demand in each cycle changes dynamically, and on the other hand, sudden demand will occur according to wartime emergencies. Dynamic and uncertain demand requires the replenishment strategy of the front warehouse to be able to adapt to the dynamically changing material demand. Therefore, this technology uses a hybrid intelligent algorithm to solve the joint replenishment problem of dynamic and uncertain demand, which is more robust, adaptable and flexible.

[0003] The research on joint replenishment technology focuses on optimizing the replenishment and inventory links from the perspective of the supply chain, establishing a dual-objective optimization function of total cost and replenishment satisfaction, and constructing an extended JRP model that considers dynamic uncertain demand. Since most joint replenishment strategies and their extended models have been proven to be an NP-hard (non-deterministic polynomial hard) problem, it means that it is difficult to find an effective algorithm to solve these problems. At present, most studies on solving the JRP model use heuristic methods, but they have the disadvantages of slow convergence and easy to fall into local optimal solutions. In this context, swarm intelligence optimization algorithms can be used to solve difficult optimization problems. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] The technical problem to be solved by the present invention is: how to provide a joint replenishment method for uncertain demand based on a hybrid intelligent algorithm.

[0006] (II) Technical solution

[0007] In order to solve the above technical problems, the present invention provides a joint replenishment method for uncertain demand based on a hybrid intelligent algorithm. The method abstracts the replenishment problem from the rear warehouse to the front warehouse into a mathematical model, and then uses the hybrid intelligent algorithm to solve the model to determine the distribution relationship from the rear warehouse to the front warehouse.

[0008] The method comprises the following steps:

[0009] Step 1: Construction of joint replenishment model for dynamic uncertain demand;

[0010] Step 2: Design of algorithm for solving the joint replenishment model based on hybrid intelligent algorithm.

[0011] In step 1, a joint replenishment model for dynamic uncertain demand is constructed:

[0012] The rear warehouse stores the same type of materials, and the front warehouse is a comprehensive warehouse that stores a variety of materials. The rear warehouse replenishes the front warehouse, and the replenishment demand comes from the material demand of the front-line demand points; referring to the basic JRP model assumptions, the rear warehouse is equivalent to the supplier, and the front warehouse is equivalent to the central warehouse. The joint replenishment strategy, that is, replenishing the same central warehouse from the same supplier or different suppliers, and the various commodities are replenished jointly, so as to achieve the purpose of sharing the main ordering costs and saving the total replenishment costs, reduce replenishment and inventory costs, and at the same time consider the goods replenishment cycle, maximize replenishment satisfaction, and achieve overall optimization of the entire supply chain.

[0013] In step 1, the assumptions of the joint replenishment model for dynamic uncertain demand are as follows:

[0014] (1) In the joint replenishment process, multiple products are replenished. A primary ordering fee will be incurred when at least one product is ordered, and a secondary ordering fee will be incurred for each product ordered;

[0015] (2) When different types of goods are ordered at the same time, these goods will share the main ordering costs of this order, significantly reducing costs, and greater demand can also reduce the unit logistics costs of the goods;

[0016] (3) Consider the maximum replenishment time for each commodity, and try to meet the replenishment needs of each commodity within the maximum replenishment time to maximize replenishment satisfaction, thereby ensuring material supply;

[0017] (4) The assumptions also include that out-of-stock conditions are not allowed, there are no quantity discounts and resource constraints, and because dynamic uncertain demand is taken into account, the rate of change of demand is not a constant.

[0018] In step 1, the parameters in the joint replenishment model for dynamic uncertain demand are defined as follows:

[0019] i: product category, i=1,2,…,n;

[0020] D i : The predicted demand for commodity i;

[0021] The actual demand for commodity i;

[0022] S: main ordering cost per replenishment;

[0023] s i : The secondary ordering cost for each replenishment of commodity i;

[0024] h i : The unit time inventory cost of unit commodity i;

[0025] t i : The maximum replenishment time for each product i;

[0026] The decision variables are defined as follows:

[0027] T: basic replenishment cycle;

[0028] k i : The joint replenishment frequency of the basic replenishment cycle of commodity i;

[0029] The total ordering cost of a product includes the primary ordering cost and the secondary ordering cost, as shown below:

[0030]

[0031] The inventory holding cost of a commodity is based on the actual demand for the commodity, as shown in the following formula:

[0032]

[0033] The replenishment satisfaction of a product is as follows:

[0034]

[0035] In order to optimize both cost and satisfaction, a joint replenishment model is established as shown below:

[0036]

[0037] The above formula is the objective function of the model, where c1, c2, and c3 are weight variables that adjust the proportion of total ordering costs, inventory costs, and replenishment satisfaction in the objective function. In order to maximize replenishment satisfaction, the replenishment satisfaction is multiplied by -1 in the model.

[0038] In supply chain management, the normal distribution is often used to model demand, so the variable demand The mean and variance are μ i , σ i The normal distribution of .

[0039] In step 1, the constraints of the joint replenishment model for dynamic uncertain demand are as follows:

[0040] All needs must be met. That is, the predicted demand for product i needs to cover the actual demand for product i.

[0041] In step 2, the joint replenishment model solution algorithm is designed based on the hybrid intelligent algorithm:

[0042] The solution of the hybrid algorithm is defined as 0, u 11 ,…,u 1n ,0,uk1 ,…,u kn ,0,u ij The above is a 0-1 variable. If its value is 1, it means that materials are delivered from the rear warehouse i to the front warehouse j;

[0043] The hybrid intelligent algorithm is as follows:

[0044] (a) Input the initial solution according to the actual situation;

[0045] (b) Calculate the total cost and satisfaction of the initial solution;

[0046] (c) Randomly swap the positions of two values ​​in the current solution, or swap the positions of multiple solutions;

[0047] (d) extracting the best m solutions from the above solutions and the current solution;

[0048] (e) A feasible solution is selected, and the pheromone of the globally feasible connection and the pheromone of the connection of the feasible solution are weighted averaged to obtain a new pheromone;

[0049] (f) Use the ant colony algorithm to obtain m feasible solutions;

[0050] (g) Calculate the objective function of the feasible solution and determine whether the maximum number of iterations has been reached. If so, output the optimal solution. If not, jump to step (c) and continue the genetic mutation operation.

[0051] The method adopts a hybrid solution of genetic algorithm and ant colony algorithm, which overcomes the shortcomings of genetic algorithm easily falling into local optimal solution and low precision of ant colony algorithm, and improves the efficiency and precision of solving the joint replenishment problem.

[0052] (III) Beneficial effects

[0053] Compared with the prior art, the technical solution proposed by the present invention has the following advantages:

[0054] (1) Solved the joint replenishment problem of multi-center warehouses with multiple time windows and established a mathematical model based on the specific problem;

[0055] (2) Combining traditional heuristic algorithms: While ensuring the accuracy of the solution, the speed of the solution is also guaranteed;

[0056] (3) Combining a variety of heuristic intelligent algorithms, a hybrid solution process that integrates genetic algorithm and ant colony algorithm is proposed. This can solve the problems that the genetic algorithm is prone to fall into local optimality and the ant colony algorithm has low accuracy, thereby improving the efficiency and accuracy of the solution.

[0057] The key innovations of the present invention are:

[0058] (1) The process structure of the joint replenishment method for uncertain demand based on hybrid intelligent algorithm is to first establish a model through the problem and then solve the problem using hybrid intelligent algorithm;

[0059] (2) In the modeling process of multiple supply warehouses and multiple demand points, the actual problem is symbolized and digitized, abstracted into a mathematical problem that can be solved by machine calculation;

[0060] (3) Integrate genetic algorithm and ant colony algorithm into a hybrid intelligent algorithm, including the definition of solution or the overall process of intelligent algorithm;

[0061] The invention is intended to protect:

[0062] (1) Joint replenishment method process based on hybrid intelligent algorithm;

[0063] (2) Modeling process of multiple supply warehouses and multiple demand points;

[0064] (3) Hybrid intelligent algorithm that integrates genetic algorithm and ant colony algorithm. DETAILED DESCRIPTION

[0065] In order to make the purpose, content, and advantages of the present invention more clear, the specific implementation methods of the present invention are further described in detail below in conjunction with embodiments.

[0066] In order to solve the above technical problems, the present invention provides a joint replenishment method for uncertain demand based on a hybrid intelligent algorithm. The method abstracts the replenishment problem from the rear warehouse to the front warehouse into a mathematical model, and then uses the hybrid intelligent algorithm to solve the model to determine the distribution relationship from the rear warehouse to the front warehouse.

[0067] The method comprises the following steps:

[0068] Step 1: Construction of joint replenishment model for dynamic uncertain demand;

[0069] Step 2: Design of algorithm for solving the joint replenishment model based on hybrid intelligent algorithm.

[0070] In step 1, a joint replenishment model for dynamic uncertain demand is constructed:

[0071] The rear warehouse stores the same type of materials, and the front warehouse is a comprehensive warehouse that stores a variety of materials. The rear warehouse replenishes the front warehouse, and the replenishment demand comes from the material demand of the front-line demand points; referring to the basic JRP model assumptions, the rear warehouse is equivalent to the supplier, and the front warehouse is equivalent to the central warehouse. The joint replenishment strategy, that is, replenishing the same central warehouse from the same supplier or different suppliers, and the various commodities are replenished jointly, so as to achieve the purpose of sharing the main ordering costs and saving the total replenishment costs, reduce replenishment and inventory costs, and at the same time consider the goods replenishment cycle, maximize replenishment satisfaction, and achieve overall optimization of the entire supply chain.

[0072] In step 1, the assumptions of the joint replenishment model for dynamic uncertain demand are as follows:

[0073] (1) In the joint replenishment process, multiple products are replenished. A primary ordering fee will be incurred when at least one product is ordered, and a secondary ordering fee will be incurred for each product ordered;

[0074] (2) When different types of goods are ordered at the same time, these goods will share the main ordering costs of this order, significantly reducing costs, and greater demand can also reduce the unit logistics costs of the goods;

[0075] (3) Consider the maximum replenishment time for each commodity, and try to meet the replenishment needs of each commodity within the maximum replenishment time to maximize replenishment satisfaction, thereby ensuring material supply;

[0076] (4) The assumptions also include that out-of-stock conditions are not allowed, quantity discounts and resource constraints do not exist, and since dynamic uncertain demand is taken into account, the rate of change of demand is not a constant.

[0077] In step 1, the parameters in the joint replenishment model for dynamic uncertain demand are defined as follows:

[0078] i: product category, i=1,2,…,n;

[0079] D i : The predicted demand for commodity i;

[0080] The actual demand for commodity i;

[0081] S: main ordering cost per replenishment;

[0082] s i : The secondary ordering cost for each replenishment of commodity i;

[0083] h i : Unit time inventory cost of unit commodity i;

[0084] ti : The maximum replenishment time for each product i;

[0085] The decision variables are defined as follows:

[0086] T: basic replenishment cycle;

[0087] k i : The joint replenishment frequency of the basic replenishment cycle of commodity i;

[0088] The total ordering cost of a product includes the primary ordering cost and the secondary ordering cost, as shown below:

[0089]

[0090] The inventory holding cost of a commodity is based on the actual demand for the commodity, as shown in the following formula:

[0091]

[0092] The replenishment satisfaction of a product is as follows:

[0093]

[0094] In order to optimize both cost and satisfaction, a joint replenishment model is established as shown below:

[0095]

[0096] The above formula is the objective function of the model, where c1, c2, and c3 are weight variables that adjust the proportion of total ordering costs, inventory costs, and replenishment satisfaction in the objective function. In order to maximize replenishment satisfaction, the replenishment satisfaction is multiplied by -1 in the model.

[0097] In supply chain management, the normal distribution is often used to model demand, so the variable demand The mean and variance are μ i , σ i The normal distribution of .

[0098] In step 1, the constraints of the joint replenishment model for dynamic uncertain demand are as follows:

[0099] All needs must be met. That is, the predicted demand for product i needs to cover the actual demand for product i.

[0100] In step 2, the joint replenishment model solution algorithm is designed based on the hybrid intelligent algorithm:

[0101] The solution of the hybrid algorithm is defined as 0, u 11 ,…,u 1n ,0,uk1 ,…,u kn ,0,u ij The above is a 0-1 variable. If its value is 1, it means that materials are delivered from the rear warehouse i to the front warehouse j;

[0102] The hybrid intelligent algorithm is as follows:

[0103] (a) Input the initial solution according to the actual situation;

[0104] (b) Calculate the total cost and satisfaction of the initial solution;

[0105] (c) Randomly swap the positions of two values ​​in the current solution, or swap the positions of multiple solutions;

[0106] (d) extracting the best m solutions from the above solutions and the current solution;

[0107] (e) A feasible solution is selected, and the pheromone of the globally feasible connection and the pheromone of the connection of the feasible solution are weighted averaged to obtain a new pheromone;

[0108] (f) Use the ant colony algorithm to obtain m feasible solutions;

[0109] (g) Calculate the objective function of the feasible solution and determine whether the maximum number of iterations has been reached. If so, output the optimal solution. If not, jump to step (c) and continue the genetic mutation operation.

[0110] The method adopts a hybrid solution of genetic algorithm and ant colony algorithm, which overcomes the shortcomings of genetic algorithm easily falling into local optimal solution and low precision of ant colony algorithm, and improves the efficiency and precision of solving the joint replenishment problem.

[0111] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A joint replenishment method for uncertain demand based on hybrid intelligent algorithm, characterized in that: The method abstracts the replenishment problem from the rear warehouse to the front warehouse into a mathematical model, and then uses a hybrid intelligent algorithm to solve the model to determine the distribution relationship from the rear warehouse to the front warehouse.

2. The method for joint replenishment of uncertain demand based on hybrid intelligent algorithm according to claim 1, characterized in that: The method comprises the following steps: Step 1: Construction of joint replenishment model for dynamic uncertain demand; Step 2: Design of algorithm for solving the joint replenishment model based on hybrid intelligent algorithm.

3. The method for joint replenishment of uncertain demand based on hybrid intelligent algorithm according to claim 2, characterized in that: In step 1, a joint replenishment model for dynamic uncertain demand is constructed: The rear warehouse stores the same type of materials, and the front warehouse is a comprehensive warehouse that stores a variety of materials. The rear warehouse replenishes the front warehouse, and the replenishment demand comes from the material demand of the front-line demand points; referring to the basic JRP model assumptions, the rear warehouse is equivalent to the supplier, and the front warehouse is equivalent to the central warehouse. The joint replenishment strategy, that is, replenishing the same central warehouse from the same supplier or different suppliers, and the various commodities are replenished jointly, so as to achieve the purpose of sharing the main ordering costs and saving the total replenishment costs, reduce replenishment and inventory costs, and at the same time consider the goods replenishment cycle, maximize replenishment satisfaction, and achieve overall optimization of the entire supply chain.

4. The method for joint replenishment of uncertain demand based on hybrid intelligent algorithm according to claim 3, characterized in that: In step 1, the assumptions of the joint replenishment model for dynamic uncertain demand are as follows: (1) In the joint replenishment process, multiple products are replenished. A primary ordering fee will be incurred when at least one product is ordered, and a secondary ordering fee will be incurred for each product ordered; (2) When different types of goods are ordered at the same time, these goods will share the main ordering costs of this order, significantly reducing costs, and greater demand can also reduce the unit logistics costs of the goods; (3) Consider the maximum replenishment time for each commodity, and try to meet the replenishment needs of each commodity within the maximum replenishment time to maximize replenishment satisfaction, thereby ensuring material supply; (4) The assumptions also include that out-of-stock conditions are not allowed, there are no quantity discounts and resource constraints, and because dynamic uncertain demand is taken into account, the rate of change of demand is not a constant.

5. The method for joint replenishment of uncertain demand based on hybrid intelligent algorithm according to claim 4, characterized in that: In step 1, the parameters in the joint replenishment model for dynamic uncertain demand are defined as follows: i: product category, i=1, 2, ..., n; D i : The predicted demand for commodity i; The actual demand for commodity i; S: main ordering cost per replenishment; s i : The secondary ordering cost for each replenishment of commodity i; h i : The unit time inventory cost of unit commodity i; t i : The maximum replenishment time for each product i; The decision variables are defined as follows: T: basic replenishment cycle; k i : The joint replenishment frequency of the basic replenishment cycle of commodity i; The total ordering cost of a product includes the primary ordering cost and the secondary ordering cost, as shown below: The inventory holding cost of a commodity is based on the actual demand for the commodity, as shown in the following formula: The replenishment satisfaction of a product is as follows: In order to optimize both cost and satisfaction, a joint replenishment model is established as shown below: The above formula is the objective function of the model, where c1, c2, and c3 are weight variables that adjust the proportion of total ordering costs, inventory costs, and replenishment satisfaction in the objective function. In order to maximize replenishment satisfaction, the replenishment satisfaction is multiplied by -1 in the model. In supply chain management, the normal distribution is often used to model demand, so the variable demand The mean and variance are μ i , σ i The normal distribution of .

6. The method for joint replenishment of uncertain demand based on hybrid intelligent algorithm according to claim 5, characterized in that: In step 1, the constraints of the joint replenishment model for dynamic uncertain demand are as follows: All needs must be met. That is, the predicted demand for product i needs to cover the actual demand for product i.

7. The method for joint replenishment of uncertain demand based on hybrid intelligent algorithm according to claim 6, characterized in that: In step 2, the joint replenishment model solution algorithm is designed based on the hybrid intelligent algorithm: The solution of the hybrid algorithm is defined as 0, u 11， …,u 1n， 0,u k1 ,…,u kn ,0,u ij The above is a 0-1 variable. If its value is 1, it means that materials are delivered from the rear warehouse i to the front warehouse i; The hybrid intelligent algorithm is as follows: (a) Input the initial solution according to the actual situation; (b) Calculate the total cost and satisfaction of the initial solution; (c) Randomly swap the positions of two values ​​in the current solution, or swap the positions of multiple solutions; (d) extracting the best m solutions from the above solutions and the current solution; (e) A feasible solution is selected, and the pheromone of the globally feasible connection and the pheromone of the connection of the feasible solution are weighted averaged to obtain a new pheromone; (f) Use the ant colony algorithm to obtain m feasible solutions; (g) Calculate the objective function of the feasible solution and determine whether the maximum number of iterations has been reached. If so, output the optimal solution. If not, jump to step (c) and continue the genetic mutation operation.

8. The method for joint replenishment of uncertain demand based on hybrid intelligent algorithm according to claim 7, characterized in that: The method adopts a hybrid solution of genetic algorithm and ant colony algorithm, which overcomes the shortcomings of genetic algorithm easily falling into local optimal solution and low precision of ant colony algorithm, and improves the efficiency and precision of solving the joint replenishment problem.

9. The method for joint replenishment of uncertain demand based on hybrid intelligent algorithm according to claim 7, characterized in that: The method solves the joint replenishment problem of multi-center warehouses with multiple time windows, and establishes a mathematical model according to the specific problem.

10. The method for joint replenishment of uncertain demand based on hybrid intelligent algorithm according to claim 7, characterized in that: The combination of the traditional heuristic algorithm ensures the solution accuracy while ensuring the solution speed.

Images