Production workshop intelligent material distribution method based on time-varying station set

Abstract

The invention discloses a production workshop intelligent material distribution method based on a time-varying station set, and the method comprises the steps: setting a station reorder point value interval based on the station information and distribution vehicle information of a production workshop, calculating a station service time window, and dynamically dividing the time-varying station set of each distribution period according to a cross relation; overall planning of a distribution path is realized by coordinating distribution decision-making elements and carrying out collaborative optimization with an adjacent distribution station set; and meanwhile, constructing a distribution optimization model under the time-varying station set, and performing hierarchical solution by adopting a double-layer genetic algorithm to obtain a joint distribution scheme including station reorder points and corresponding material distribution plans. According to the method, the time-varying characteristics of the distribution station set are considered, and the distribution path is dynamically planned according to the distribution decision-making elements, so that the distribution scheme can plan the dynamic change of the material demand of each station, the total cost of material distribution is reduced, and intelligent planning of the distribution scheme is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of workshop material distribution optimization, and in particular relates to an intelligent material distribution method for a production workshop based on a time-varying workstation set. Background Art

[0002] With the innovative development and cross-integration of the new generation of information technology, the manufacturing industry, as the cornerstone of my country's industrial development, is entering a stage of high-quality development. Its production technology is becoming increasingly mature, and the focus of enterprises is gradually expanding from a single production field to a more comprehensive non-production field. Among them, production logistics, as a key link in the non-production field, has attracted much attention.

[0003] Production logistics is the foothold of the deep integration of manufacturing and logistics industries. It runs through the entire production process and covers the flow and management of materials, information and funds in the production process. In the current context of intelligent manufacturing, accurate and intelligent material distribution is the key to efficient and low-cost operation of production logistics. Its function has gradually evolved from traditional simple distribution to a key part with certain value-added functions. It accurately controls the smooth progress of the entire production process, which is of great significance to improving production efficiency and reducing production risks.

[0004] Traditional material distribution strategies usually formulate material distribution plans in advance based on production plans, and adopt a distribution model with a fixed set of workstations, with fixed distribution personnel distributing materials to designated workstations regularly according to the plan. Although this model performs well in a stable demand environment, the shortcomings of this model are gradually emerging in today's highly dynamic and complex production environment. For example, the inability to effectively handle demand fluctuations, the difficulty in adjusting distribution routes and distribution times in real time, and the lack of response to immediate demand on the production line all limit the intelligent development of production logistics. The present invention makes up for the shortcomings of traditional strategies by rationally planning the reorder points of workstations, determining the distribution workstation set and distribution-related decisions under time-varying states, and constructing a distribution optimization model and using a two-layer genetic algorithm for solution. Specifically: the rational planning of the reorder points of workstations enhances the visualization level and planning ability of material distribution; the determination of the distribution workstation set and distribution-related decisions enables the distribution process to meet the dynamic needs of workstations; the application of distribution optimization models and solution algorithms enables the distribution process to be closely integrated with actual application scenarios, and realizes intelligent planning of distribution plans. These innovative improvements enable the present invention to be better adapted to different production situations and application scenarios, and to achieve intelligent, flexible, and efficient material distribution planning. Summary of the invention

[0005] In view of the shortcomings of traditional distribution strategies, the present invention provides an intelligent material distribution method for production workshops based on time-varying workstation sets, which aims to build a more intelligent and flexible material distribution method, in order to achieve dynamic coordination between the distribution process and real-time production needs, thereby helping enterprises to achieve the goal of reducing costs and increasing efficiency.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: a production workshop intelligent material distribution method based on a time-varying workstation set, comprising the following steps: Step S1: Collect the workstation information and delivery vehicle information of the production workshop to determine the time range in which material delivery needs to be executed; Step S2: Based on the workstation information and delivery vehicle information collected in step S1, the value range of the workstation reorder point is limited; Step S3: Based on the workstation reorder point value range set in step S2, a reasonable workstation reorder point is randomly set for each workstation, and the combination becomes a workstation reorder point combination; Step S4: Based on the workstation reorder point combination and the workstation information collected in step S1, the workstation service time window is calculated, and the delivery workstation set of the current delivery cycle is determined based on the cross relationship of the workstation service time windows; Step S5: Determine a method for calculating the departure time of the delivery vehicle based on the specific number of delivery stations of the delivery vehicle, and determine a method for calculating the number of materials delivered to the station based on the arrival time of the station; Step S6: Based on the set of delivery stations determined in step S4 and the calculation method of the departure time of the delivery vehicle and the quantity of material delivered to the station determined in step S5, the forward method is used to calculate and evaluate the loading status of the delivery vehicle in real time, plan the material delivery path, and obtain the distribution planning result under the time-varying characteristics of the current delivery cycle; Step S7: Based on the distribution planning result under the time-varying characteristics of the current distribution cycle obtained in step S6, the service time window of the station and the distribution station set of the next distribution cycle are updated according to step S4, and the distribution station of the end distribution vehicle of the current distribution cycle is collaboratively optimized with the adjacent distribution station set by comparing the service time window of the distribution station of the end distribution vehicle of the current distribution cycle with the earliest departure time of the next distribution cycle; Step S8: Based on the workstation information and delivery vehicle information collected in step S1, a delivery optimization model under a time-varying workstation set is constructed; Step S9: Based on the distribution optimization model under the time-varying workstation set constructed in step S8, the distribution optimization model is solved by using the inner genetic algorithm, and steps S5 to S7 are repeated until all material distribution tasks within the time range determined in step S1 to be required for material distribution are completed, and the total distribution planning scheme under the current workstation re-order point is output; Step S10: Record the overall distribution plan obtained in step S9, use the outer genetic algorithm to update the workstation reorder point, repeat steps S4-S9 until the maximum number of iterations is reached, and output the optimal distribution plan including the workstation reorder point combination and distribution planning.

[0007] Furthermore, in step S1, the workstation information and delivery vehicle information of the production workshop are collected to determine the time range for executing material delivery; specifically: Workstation information includes: the number of workstations, the travel distance between workstations, the available time of materials per workstation, the service time of the workstation, the volume of the workstation material box, the maximum number of material boxes in the line inventory of the workstation, and the loadable number of material boxes per workstation; The delivery vehicle information includes: the driving speed of the delivery vehicle, the maximum loading volume of the delivery vehicle; The time frame within which material delivery needs to be performed is pre-set.

[0008] Furthermore, in step S2, based on the workstation information and delivery vehicle information collected in step S1, the value range of the workstation reorder point is limited; specifically: The minimum value level of the reorder point of the workstation is set as shown in formula (1): (1); The maximum value level of the reorder point of the workstation is set as shown in formula (2): (2); Where: For workstation reorder point; is the minimum reorder point level of the workstation; For workstation The maximum value level of the reorder point; * is the multiplication sign; Is the station number, belonging to the collection of all stations ; is the collection of all workstations; The number of the dispatched delivery vehicle, belonging to the collection of all delivery vehicles ; It is the collection of all delivery vehicles; For warehouse and workstation The distance between them, unit: m; is the driving speed of the delivery vehicle, unit: m / min; For workstation The usable time of a unit material, unit: s; For workstation Maximum number of bins in line side inventory; For workstation The amount of material that can be loaded into a unit container.

[0009] Furthermore, in step S3, based on the workstation reorder point value range set in step S2, a reasonable workstation reorder point is randomly set for each workstation, and the combination becomes a workstation reorder point combination; specifically: For each workstation , based on the value range of the reorder point of the workstation , randomly select an integer in the interval of the workstation reorder point as the workstation Reorder point , and combine the reorder points of each workstation to form a workstation reorder point combination .

[0010] Furthermore, in step S4, based on the workstation reorder point combination and the workstation information collected in step S1, the workstation service time window is calculated, and the delivery workstation set of the current delivery cycle is determined based on the cross relationship of the workstation service time window; specifically: Step S41: Calculation of workstation service time window; The workstation service time window consists of the delivery arrival time window and the delivery departure time window. Based on the time-varying characteristics, the setting standard of the workstation service time window is determined. The specific calculation method is as follows: The earliest delivery arrival time is set as shown in formula (3): (3); In the formula, For workstation No. The earliest arrival time of the delivery; For the Deliver materials to the workstation moment; For workstation No. The quantity of materials after the first delivery; For workstation A delivery belongs to the workstation The number of material delivery times ; For workstation The collection of material delivery times; The latest delivery arrival time is set as shown in formula (4): (4); In the formula, For workstation No. The latest arrival time of the delivery; For workstation service time; Based on the delivery arrival time window, the delivery departure time window of the workstation is calculated, which is divided into the earliest departure time of delivery and the latest departure time of delivery, which are expressed by formula (5) and formula (6) respectively: (5); (6); In the formula, For workstation No. The earliest departure time for the delivery; For workstation No. The latest departure time for the delivery; Step S42: dividing the delivery station set based on the intersection of time windows; Based on the cross-relationship between the workstation service time windows, the workstations that need to perform material distribution in the current distribution cycle are divided into distribution workstation sets. The division method is as follows: A collection of all workstations The earliest workstation to initiate a delivery request is set to , marking the earliest departure time as The latest departure time is ; For removing the workstation that initiated the earliest delivery request Each workstation outside If the workstation Current Earliest Departure Time Less than the latest departure time , then the workstation It is classified into the same delivery station set as the first station, and the expression is shown in formula (7): (7); In the formula, For the A set of delivery stations.

[0011] Furthermore, in step S5, a method for calculating the departure time of the delivery vehicle is determined based on the specific number of delivery stations of the delivery vehicle, and a method for calculating the number of materials delivered to the station is determined based on the arrival time of the station; specifically: Step S51: Calculation method of the delivery vehicle departure time; Select the delivery departure time window and delivery arrival time window as the basis to calculate the delivery vehicle departure time. Assuming that the delivery vehicle In the The departure time of the delivery cycle is The method for calculating the departure time of the delivery vehicle is divided into two cases: step S511 and step S512: Step S511: When the number of workstations to be delivered is equal to 1, it is assumed that the delivery workstation is workstation , the calculation method of the delivery vehicle departure time is shown in formula (8), which is No. The earliest departure time of a delivery cycle; (8); In the formula, For delivery vehicles In the The departure time of each delivery cycle; is a certain delivery cycle, belonging to the set of all delivery cycles ; is the collection of all delivery cycles; is a 0-1 variable: if Delivery cycle delivery vehicles Responsible for delivering materials to workstations but is 1, otherwise is 0; Is a 0-1 variable: If the workstation No. The second delivery belongs to Delivery cycle is 1, otherwise is 0; Step S512: When the number of workstations to be delivered is greater than 1, the latest arrival time of the delivery vehicle is used as the calculation basis, and the departure time of the delivery vehicle is obtained by the reverse calculation method; specifically: Step S5121, for each workstation , according to the workstation of the current delivery cycle No. The latest arrival time of the delivery The values ​​of are sorted in descending order, and the largest station number among the values ​​of the latest arrival time is , the second largest station number is , then the workstation No. Delivery cycle arrival time It is expressed as formula (9): (9); In the formula, For the Delivery cycle delivery vehicles Responsible for delivery workstation The latest arrival time; For workstation and workstation The distance between For workstation service time; Step S5122, remove the workstation , the original workstation Update to new workstation , from the latest arrival time Select the workstation in descending order The corresponding next station is used as the new station , execute formula (9) again until the arrival time of all delivery vehicles at all delivery stations of the delivery vehicle is calculated; Step S5123, based on the final calculated workstation No. Delivery cycle arrival time , determine the delivery vehicle In the Departure time of delivery cycle , expressed as formula (10): (10); In the formula, is the distance between the warehouse and workstation b; Step S52: a method for calculating the quantity of material delivered to a workstation; When delivery vehicle In the Departure time of delivery cycle After confirmation, the delivery frequency of each station is tracked and recorded to identify the first Delivery cycle to workstation The delivery belongs to the workstation The number of deliveries is then further determined based on the driving distance between workstations and the speed of the delivery vehicle. No. The arrival time of the delivery ; Workstation No. The time of initiating the material request is calculated by formula (11): (11); Where: For workstation No. The time when the material request was first initiated; Workstation No. The specific quantity of remaining materials at the time of the next delivery is calculated by formula (12): (12); Where: For workstation No. The specific quantity of remaining materials at the time of the next delivery arrival; Finally, the workstation No. The number of remaining available containers and workstations at the time of the next delivery No. The number of containers delivered each time is calculated using formula (13) and formula (14): (13); (14); In the formula, For workstation No. The number of available containers remaining at the time of the next delivery arrival. For workstation No. The number of boxes delivered.

[0012] Furthermore, in step S6, based on the set of delivery stations determined in step S4 and the calculation method of the departure time of the delivery vehicle and the quantity of material delivered to the station determined in step S5, the forward method is used to calculate and evaluate the loading status of the delivery vehicle in real time, plan the material delivery path, and obtain the distribution planning result under the time-varying characteristics of the current delivery cycle; specifically: Step S61, set the station set that needs to be delivered to , the number of workstations in the distribution workstation concentration is , the set of all delivery vehicles is , delivery vehicles The material distribution path is , the maximum loading volume of the delivery vehicle is , the delivery vehicle departure time is ; When starting path planning, the workstation The first delivery station and the initial number 1, delivery vehicle This is the first delivery vehicle and the initial number is 1, and the initial material delivery path of the delivery vehicle is empty; Step S62: Add delivery vehicle Material delivery path and calculate the corresponding delivery vehicle departure time and the current loading volume of the delivery vehicle ; Step S63, then evaluate the loading status of the delivery vehicle: Step S631: If and :The current loading volume of the delivery vehicle does not exceed the maximum loading volume of the delivery vehicle and there is a next workstation, then the workstation Add to delivery vehicle Update the material distribution path of for , and go to step S62; Step S632: If and : If the current loading volume of the delivery vehicle does not exceed the maximum loading volume of the delivery vehicle and there is no next station, the delivery vehicle is recorded The complete material distribution path of the distribution vehicle is obtained by obtaining the material distribution paths of all distribution vehicles in the current distribution cycle; Step S633: If and : If the current loading volume of the delivery vehicle exceeds the maximum loading volume of the delivery vehicle and there is a next workstation, the workstation will be removed from the current material delivery path. The remaining workstations are recorded as delivery vehicles. The complete material distribution path is updated at the same time for , the workstation As a delivery vehicle Delivery station, update for , and go to step S6; Step S634: If and : If the current loading volume of the delivery vehicle exceeds the maximum loading volume of the delivery vehicle and there is no next workstation, the workstation will be removed from the current material delivery path. The remaining workstations are recorded as delivery vehicles. Complete material distribution path, and the workstation As a delivery vehicle The distribution station can obtain the material distribution paths of all distribution vehicles in the current distribution cycle.

[0013] Furthermore, in step S7, based on the distribution planning result under the time-varying characteristics of the current distribution cycle obtained in step S6, the station service time window and the distribution station set of the next distribution cycle are updated according to step S4, and the distribution station of the terminal distribution vehicle of the current distribution cycle is collaboratively optimized with the adjacent distribution station set by comparing the service time window of the distribution station of the terminal distribution vehicle of the current distribution cycle with the earliest departure time of the next distribution cycle; specifically, as follows: Step S61, assuming that the delivery vehicle that departs from the warehouse the latest in the current delivery cycle is the delivery vehicle , delivery vehicles All workstations responsible for delivery are combined into a workstation group , workstation group The number of workstations in indivual; Step S62, define the delivery vehicles in the current delivery cycle Delivery team The latest departure time for each workstation is , the earliest departure time for the next delivery cycle is ; Step S63, when , then the delivery vehicle Delivery team Co-optimize with adjacent delivery station sets; Step S64: If the delivery vehicle Delivery team The latest departure time for each workstation Both are greater than the earliest departure time of the next delivery cycle , then the delivery vehicle Delivery team The various workstations in the distribution cycle are merged into the next distribution cycle for joint distribution, and the workstation data is updated; otherwise, the original material distribution path is maintained.

[0014] Furthermore, in step S8, based on the workstation information and delivery vehicle information collected in step S1, a delivery optimization model under a time-varying workstation set is constructed; specifically: Step S81: constructing the objective function; taking the delivery vehicle startup cost, delivery vehicle driving cost, line-side inventory cost, time window penalty cost and line-side inventory penalty cost into account in the total cost, with minimizing the total cost of material distribution as the overall goal; Step S811: The calculation formula for the delivery vehicle startup cost is shown in formula (15): (15); In the formula, start-up costs for delivery vehicles; The single startup cost of the delivery vehicle; is the total number of starts of the delivery vehicle; Step S812: The calculation formula of the delivery vehicle driving cost is shown in formula (16): (16); In the formula, The driving cost of the delivery vehicle; is the unit driving cost of the delivery vehicle; 0-1 variable: If the delivery vehicle In the Delivery cycle from workstation Travel to the workstation but is 1, otherwise is 0; For workstation With workstation The distance between Step S813: The calculation formula of line-side inventory cost is shown in formula (17): (17); In the formula, is the line-side inventory cost; is the unit volume cost of line-side inventory; For workstation No. The quantity of material when the first delivery arrives; For workstation The volume of the material box, in m 3 ; For the Deliver materials to the workstation moment; Step S814: The calculation formula of the time window penalty cost is shown in formula (18): (18); In the formula, Penalize costs for time windows; Penalty costs for early arrival; Penalty costs for late arrival; Step S815: The calculation formula of the line side inventory penalty cost is shown in formula (19): (19); In the formula, Line side inventory penalty costs; Penalty cost per unit volume of line-side inventory; In order to minimize the total cost of material distribution, the objective function is constructed as shown in formula (20): (20); Where: is the total delivery cost; Step S82: constructing constraint conditions; Based on the optimization process in line with the actual production operation conditions, the following constraints are given: Constraint 1: , indicating that the reorder point value of each workstation must not exceed the value range; Constraint 2: , which means that the delivery vehicle starts from the warehouse every time and returns to the warehouse after completing the delivery task; where, Indicates delivery vehicle In the A delivery cycle from the warehouse to the workstation ; Indicates delivery vehicle In the Delivery cycle from workstation Drive to the warehouse; Constraint three: , indicating that the material distribution path must be connected, that is, the distribution vehicle arrives at the workstation , must be from the workstation Leave; in the formula, Indicates delivery vehicle In the Delivery cycle from workstation Travel to the workstation ; Indicates delivery vehicle In the Delivery cycle from workstation Travel to the workstation ; Constraint 4: , indicating that in each delivery cycle, there is only 0 or 1 delivery vehicle as a workstation Services cannot be split into two or more vehicles; Constraint 5: , indicating that the total loading volume of the delivery vehicle for each delivery shall not be greater than the maximum loading volume of the delivery vehicle; where, The maximum loading volume of the delivery vehicle, in m 3 ; Constraint six: , indicating that the number of line-side material boxes at each workstation must not exceed the corresponding maximum number of line-side inventory material boxes; Constraint seven: , which means that the order between the two workstations delivered by the same delivery vehicle must be sequential. Only after the previous workstation is delivered, the next workstation can be delivered. In the formula, For the Delivery cycle to deliver materials to the workstation moment; For the Delivery cycle to deliver materials to the workstation moment; is an infinite value; Constraint 8: , indicating the workstation The material delivery must be completed within the specified time window.

[0015] Furthermore, in step S9, based on the distribution optimization model under the time-varying workstation set constructed in step S8, the distribution optimization model is solved by using the inner genetic algorithm, and steps S5 to S7 are repeated until all material distribution tasks within the distribution time range specified in S1 are completed, and the total distribution planning scheme under the current workstation reorder point is output; specifically: The inner genetic algorithm is part of the double-layer genetic algorithm. It is mainly responsible for solving the corresponding material distribution path, distribution vehicle departure time and total distribution cost for each set of reorder point plans based on the distribution optimization model under the time-varying workstation set. ,In the solution process, the encoding method selects the real number encoding method, and the genetic operations include the roulette selection strategy, the two-point crossover strategy and the two-point mutation strategy; The basic idea of ​​the inner genetic algorithm solution is: based on the combination of workstation reorder points , within the time range of material distribution pre-set by S1, according to the intersection between the service time windows of the workstations, the time-varying workstation set is determined when each distribution task is executed, and the material distribution path is planned by the inner genetic algorithm. Through continuous iterative calculation, the solution with the minimum total distribution cost is selected as the optimal material distribution path for the distribution cycle. At the same time, after the planning of each distribution cycle is determined, the workstation information is updated in real time, and the next distribution cycle is solved until all distribution tasks within the pre-set time range for material distribution are completed, and the workstation re-order point combination is output. The optimal material distribution path, distribution vehicle departure time, various distribution costs and total distribution cost for each distribution cycle under .

[0016] Further, in step S10, the overall material distribution planning scheme obtained in step S9 is recorded, the workstation reorder point is updated using the outer genetic algorithm, and steps S4 to S9 are repeated until the maximum number of iterations is reached, and the optimal distribution scheme including the workstation reorder point combination and the distribution plan is output; specifically: The outer genetic algorithm is part of the double-layer genetic algorithm and is mainly responsible for adjusting the reorder point combination of the workstation , as the basis of distribution planning, the real number encoding method is selected in the solution process, and the genetic operations include ternary tournament selection strategy, uniform crossover strategy and multi-point mutation strategy; The basic idea of ​​solving the outer genetic algorithm is: in each iteration, the value interval of the reorder point of the workstation is , using genetic operations to update the station reorder point combination , and use the inner genetic algorithm to determine the corresponding distribution plan, record the optimal workstation reorder point combination and the corresponding material distribution plan in each iteration, and compare the total distribution cost of the distribution plan with the optimal total distribution cost in the previous iteration. If the distribution plan under the new reorder point combination can reduce the total distribution cost, then update the current optimal reorder point combination to the new workstation reorder point, and repeat the iteration until the set maximum number of iterations is met. Finally, the optimal workstation reorder point combination in all iterations and the optimal material distribution path, distribution vehicle departure time, various distribution costs and total distribution cost of each distribution cycle of the corresponding distribution plan are obtained as the final distribution plan.

[0017] Compared with the prior art, the present invention has the following advantages: 1. The present invention combines the production workshop workstation information and the delivery vehicle information to set the interval of the reorder point of each workstation, and optimizes the reorder point of the workstation based on this. It not only ensures the dynamic balance of supply and demand of the production line, but also enhances the visualization level and planning ability of the distribution process. At the same time, by combining the production workshop workstation information and the delivery vehicle information, the setting interval of the workstation reorder point can be flexibly adjusted, so as to intelligently adapt to different production workshops and have stronger adaptability.

[0018] 2. Based on the time-varying characteristics of the distribution station set, the present invention dynamically divides the distribution station set of each distribution cycle according to the cross-relationship between the station service time windows. This enables material distribution planning to flexibly respond to the real-time needs of stations in different cycles, ensures the continuity of the production process, and achieves a high degree of consistency between material distribution and production needs.

[0019] 3. The present invention determines the specific delivery route by comprehensively coordinating the departure time, delivery quantity and dynamic changes of route planning in the delivery process and their mutual influence, and coordinates and optimizes the delivery station and adjacent time window of the terminal delivery vehicle based on the station service time window. This process effectively improves the dynamic response capability and flexibility of the delivery process, ensures more accurate delivery decisions, and helps improve the overall production efficiency of the workshop.

[0020] 4. The present invention constructs a distribution optimization model under a time-varying workstation set by comprehensively considering the distribution vehicle driving and line-side inventory management processes, and combining constraints such as the full-box feeding strategy and the workstation service time window. Through a two-layer genetic algorithm, the complex decision-making problem is decomposed into two interrelated and mutually supportive optimization levels, realizing the joint optimization of the workstation reorder point and the material distribution plan. Not only does it achieve a high degree of fit between material distribution and the actual production process, ensuring comprehensive and flexible planning of the distribution path, but it also makes the overall planning process more intelligent and scientific through the combination of dynamic planning and intelligent algorithms. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a schematic diagram of the overall framework of the production workshop intelligent material distribution method based on the time-varying workstation set in the present invention; Figure 2 It is an iterative graph of the optimal solution when a double-layer genetic algorithm is used to solve a problem in the present invention. DETAILED DESCRIPTION

[0022] The present invention is further described in detail below with reference to the accompanying drawings and embodiments.

[0023] See also Figure 1 The technical solution provided by the present invention is: an intelligent material distribution method for a production workshop based on a time-varying workstation set, comprising the following steps: Step S1: Collect the workstation information and delivery vehicle information of the production workshop to determine the time range in which material delivery needs to be executed; Step S2: Based on the workstation information and delivery vehicle information collected in step S1, the value range of the workstation reorder point is limited; Step S3: Based on the workstation reorder point value range set in step S2, a reasonable workstation reorder point is randomly set for each workstation, and the combination becomes a workstation reorder point combination; Step S4: Based on the workstation reorder point combination and the workstation information collected in step S1, the workstation service time window is calculated, and the delivery workstation set of the current delivery cycle is determined based on the cross relationship of the workstation service time windows; Step S5: Determine a method for calculating the departure time of the delivery vehicle based on the specific number of delivery stations of the delivery vehicle, and determine a method for calculating the number of materials delivered to the station based on the arrival time of the station; Step S6: Based on the set of delivery stations determined in step S4 and the calculation method of the departure time of the delivery vehicle and the quantity of material delivered to the station determined in step S5, the forward method is used to calculate and evaluate the loading status of the delivery vehicle in real time, plan the material delivery path, and obtain the distribution planning result under the time-varying characteristics of the current delivery cycle; Step S7: Based on the distribution planning result under the time-varying characteristics of the current distribution cycle obtained in step S6, the service time window of the station and the distribution station set of the next distribution cycle are updated according to step S4, and the distribution station of the end distribution vehicle of the current distribution cycle is collaboratively optimized with the adjacent distribution station set by comparing the service time window of the distribution station of the end distribution vehicle of the current distribution cycle with the earliest departure time of the next distribution cycle; Step S8: Based on the workstation information and delivery vehicle information collected in step S1, a delivery optimization model under a time-varying workstation set is constructed; Step S9: Based on the distribution optimization model under the time-varying workstation set constructed in step S8, the distribution optimization model is solved by using the inner genetic algorithm, and steps S5 to S7 are repeated until all material distribution tasks within the time range determined in step S1 to be required for material distribution are completed, and the total distribution planning scheme under the current workstation re-order point is output; Step S10: Record the overall distribution plan obtained in step S9, use the outer genetic algorithm to update the workstation reorder point, repeat steps S4-S9 until the maximum number of iterations is reached, and output the optimal distribution plan including the workstation reorder point combination and distribution planning.

[0024] Furthermore, in step S1, the workstation information and delivery vehicle information of the production workshop are collected to determine the time range for executing material delivery; specifically: Workstation information includes: the number of workstations, the travel distance between workstations, the available time of materials per workstation, the service time of the workstation, the volume of the workstation material box, the maximum number of material boxes in the line inventory of the workstation, and the loadable number of material boxes per workstation; The delivery vehicle information includes: the driving speed of the delivery vehicle, the maximum loading volume of the delivery vehicle; The time frame within which material delivery needs to be performed is pre-set.

[0025] Furthermore, in step S2, based on the workstation information and delivery vehicle information collected in step S1, the value range of the workstation reorder point is limited; specifically: The minimum value level of the reorder point of the workstation is set as shown in formula (1): (1); The maximum value level of the reorder point of the workstation is set as shown in formula (2): (2); Where: For workstation reorder point; is the minimum reorder point level of the workstation; For workstation The maximum value level of the reorder point; * is the multiplication sign; Is the station number, belonging to the collection of all stations ; is the collection of all workstations; The number of the dispatched delivery vehicle, belonging to the collection of all delivery vehicles ; It is the collection of all delivery vehicles; For warehouse and workstation The distance between them, unit: m; is the driving speed of the delivery vehicle, unit: m / min; For workstation The usable time of a unit material, unit: s; For workstation Maximum number of bins in line side inventory; For workstation The amount of material that can be loaded into a unit container.

[0026] Furthermore, in step S3, based on the workstation reorder point value range set in step S2, a reasonable workstation reorder point is randomly set for each workstation, and the combination becomes a workstation reorder point combination; specifically: For each workstation , based on the value range of the reorder point of the workstation , randomly select an integer in the interval of the workstation reorder point as the workstation Reorder point , and combine the reorder points of each workstation to form a workstation reorder point combination .

[0027] Furthermore, in step S4, based on the workstation reorder point combination and the workstation information collected in step S1, the workstation service time window is calculated, and the delivery workstation set of the current delivery cycle is determined based on the cross relationship of the workstation service time window; specifically: Step S41: Calculation of workstation service time window; The workstation service time window consists of the delivery arrival time window and the delivery departure time window. Based on the time-varying characteristics, the setting standard of the workstation service time window is determined. The specific calculation method is as follows: The earliest delivery arrival time is set as shown in formula (3): (3); In the formula, For workstation No. The earliest arrival time of the delivery; For the Deliver materials to the workstation moment; For workstation No. The quantity of materials after the first delivery; For workstation A delivery belongs to the workstation The number of material delivery times ; For workstation The collection of material delivery times; The latest delivery arrival time is set as shown in formula (4): (4); In the formula, For workstation No. The latest arrival time of the delivery; For workstation service time; Based on the delivery arrival time window, the delivery departure time window of the workstation is calculated, which is divided into the earliest departure time of delivery and the latest departure time of delivery, which are expressed by formula (5) and formula (6) respectively: (5); (6); In the formula, For workstation No. The earliest departure time for the delivery; For workstation No. The latest departure time for the delivery; Step S42: dividing the delivery station set based on the intersection of time windows; Based on the cross-relationship between the workstation service time windows, the workstations that need to perform material distribution in the current distribution cycle are divided into distribution workstation sets. The division method is as follows: A collection of all workstations The earliest workstation to initiate a delivery request is set to , marking the earliest departure time as The latest departure time is ; For removing the workstation that initiated the earliest delivery request Each workstation outside If the workstation Current Earliest Departure Time Less than the latest departure time , then the workstation It is classified into the same delivery station set as the first station, and the expression is shown in formula (7): (7); In the formula, For the A set of delivery stations.

[0028] Furthermore, in step S5, a method for calculating the departure time of the delivery vehicle is determined based on the specific number of delivery stations of the delivery vehicle, and a method for calculating the number of materials delivered to the station is determined based on the arrival time of the station; specifically: Step S51: Calculation method of the delivery vehicle departure time; Select the delivery departure time window and delivery arrival time window as the basis to calculate the delivery vehicle departure time. Assuming that the delivery vehicle In the The departure time of the delivery cycle is The method for calculating the departure time of the delivery vehicle is divided into two cases: step S511 and step S512: Step S511: When the number of workstations to be delivered is equal to 1, it is assumed that the delivery workstation is workstation , the calculation method of the delivery vehicle departure time is shown in formula (8), which is No. The earliest departure time of a delivery cycle; (8); In the formula, For delivery vehicles In the The departure time of each delivery cycle; is a certain delivery cycle, belonging to the set of all delivery cycles ; is the collection of all delivery cycles; is a 0-1 variable: if Delivery cycle delivery vehicles Responsible for delivering materials to workstations but is 1, otherwise is 0; Is a 0-1 variable: If the workstation No. The second delivery belongs to Delivery cycle is 1, otherwise is 0; Step S512: When the number of workstations to be delivered is greater than 1, the latest arrival time of the delivery vehicle is used as the calculation basis, and the departure time of the delivery vehicle is obtained by the reverse calculation method; specifically: Step S5121, for each workstation , according to the workstation of the current delivery cycle No. The latest arrival time of the delivery The values ​​of are sorted in descending order, and the largest station number among the values ​​of the latest arrival time is , the second largest station number is , then the workstation No. Delivery cycle arrival time It is expressed as formula (9): (9); In the formula, For the Delivery cycle delivery vehicles Responsible for delivery workstation The latest arrival time; For workstation and workstation The distance between For workstation service time; Step S5122, remove the workstation , the original workstation Update to new workstation , from the latest arrival time Select the workstation in descending order The corresponding next station is used as the new station , execute formula (9) again until the arrival time of all delivery vehicles at all delivery stations of the delivery vehicle is calculated; Step S5123, based on the final calculated workstation No. Delivery cycle arrival time , determine the delivery vehicle In the Departure time of delivery cycle , expressed as formula (10): (10); In the formula, is the distance between the warehouse and workstation b; Step S52: a method for calculating the quantity of material delivered to a workstation; When delivery vehicle In the Departure time of delivery cycle After confirmation, the delivery frequency of each station is tracked and recorded to identify the first Delivery cycle to workstation The delivery belongs to the workstation The number of deliveries is then further determined based on the driving distance between workstations and the speed of the delivery vehicle. No. The arrival time of the delivery ; Workstation No. The time of initiating the material request is calculated by formula (11): (11); Where: For workstation No. The time when the material request was first initiated; Workstation No. The specific quantity of remaining materials at the time of the next delivery is calculated by formula (12): (12); Where: For workstation No. The specific quantity of remaining materials at the time of the next delivery arrival; Finally, the workstation No. The number of remaining available containers and workstations at the time of the next delivery No. The number of containers delivered each time is calculated using formula (13) and formula (14): (13); (14); In the formula, For workstation No. The number of available containers remaining at the time of the next delivery arrival. For workstation No. The number of boxes delivered.

[0029] Furthermore, in step S6, based on the set of delivery stations determined in step S4 and the calculation method of the departure time of the delivery vehicle and the quantity of material delivered to the station determined in step S5, the forward method is used to calculate and evaluate the loading status of the delivery vehicle in real time, plan the material delivery path, and obtain the distribution planning result under the time-varying characteristics of the current delivery cycle; specifically: Step S61, set the station set that needs to be delivered to , the number of workstations in the distribution workstation concentration is , the set of all delivery vehicles is , delivery vehicles The material distribution path is , the maximum loading volume of the delivery vehicle is , the delivery vehicle departure time is ; When starting path planning, the workstation The first delivery station and the initial number 1, delivery vehicle This is the first delivery vehicle and the initial number is 1, and the initial material delivery path of the delivery vehicle is empty; Step S62: Add delivery vehicle Material delivery path and calculate the corresponding delivery vehicle departure time and the current loading volume of the delivery vehicle ; Step S63, then evaluate the loading status of the delivery vehicle: Step S631: If and :The current loading volume of the delivery vehicle does not exceed the maximum loading volume of the delivery vehicle and there is a next workstation, then the workstation Add to delivery vehicle Update the material distribution path of for , and go to step S62; Step S632: If and : If the current loading volume of the delivery vehicle does not exceed the maximum loading volume of the delivery vehicle and there is no next station, the delivery vehicle is recorded The complete material distribution path of the distribution vehicle is obtained by obtaining the material distribution paths of all distribution vehicles in the current distribution cycle; Step S633: If and : If the current loading volume of the delivery vehicle exceeds the maximum loading volume of the delivery vehicle and there is a next workstation, the workstation will be removed from the current material delivery path. The remaining workstations are recorded as delivery vehicles. The complete material distribution path is updated at the same time for , the workstation As a delivery vehicle Delivery station, update for , and go to step S6; Step S634: If and : If the current loading volume of the delivery vehicle exceeds the maximum loading volume of the delivery vehicle and there is no next workstation, the workstation will be removed from the current material delivery path. The remaining workstations are recorded as delivery vehicles. Complete material distribution path, and the workstation As a delivery vehicle The distribution station can obtain the material distribution paths of all distribution vehicles in the current distribution cycle.

[0030] Furthermore, in step S7, based on the distribution planning result under the time-varying characteristics of the current distribution cycle obtained in step S6, the station service time window and the distribution station set of the next distribution cycle are updated according to step S4, and the distribution station of the terminal distribution vehicle of the current distribution cycle is collaboratively optimized with the adjacent distribution station set by comparing the service time window of the distribution station of the terminal distribution vehicle of the current distribution cycle with the earliest departure time of the next distribution cycle; specifically, as follows: Step S61, assuming that the delivery vehicle that departs from the warehouse the latest in the current delivery cycle is the delivery vehicle , delivery vehicles All workstations responsible for delivery are combined into a workstation group , workstation group The number of workstations in indivual; Step S62, define the delivery vehicles in the current delivery cycle Delivery team The latest departure time for each workstation is , the earliest departure time for the next delivery cycle is ; Step S63, when , then the delivery vehicle Delivery team Co-optimize with adjacent delivery station sets; Step S64: If the delivery vehicle Delivery team The latest departure time for each workstation Both are greater than the earliest departure time of the next delivery cycle , then the delivery vehicle Delivery team The various workstations in the distribution cycle are merged into the next distribution cycle for joint distribution, and the workstation data is updated; otherwise, the original material distribution path is maintained.

[0031] Furthermore, in step S8, based on the workstation information and delivery vehicle information collected in step S1, a delivery optimization model under a time-varying workstation set is constructed; specifically: Step S81: constructing the objective function; taking the delivery vehicle startup cost, delivery vehicle driving cost, line-side inventory cost, time window penalty cost and line-side inventory penalty cost into account in the total cost, with minimizing the total cost of material distribution as the overall goal; Step S811: The calculation formula for the delivery vehicle startup cost is shown in formula (15): (15); In the formula, start-up costs for delivery vehicles; The single startup cost of the delivery vehicle; is the total number of starts of the delivery vehicle; Step S812: The calculation formula of the delivery vehicle driving cost is shown in formula (16): (16); In the formula, The driving cost of the delivery vehicle; is the unit driving cost of the delivery vehicle; 0-1 variable: If the delivery vehicle In the Delivery cycle from workstation Travel to the workstation but is 1, otherwise is 0; For workstation With workstation The distance between Step S813: The calculation formula of line-side inventory cost is shown in formula (17): (17); In the formula, is the line-side inventory cost; is the unit volume cost of line-side inventory; For workstation No. The quantity of material when the first delivery arrives; For workstation The volume of the material box, in m 3 ; For the Deliver materials to the workstation moment; Step S814: The calculation formula of the time window penalty cost is shown in formula (18): (18); In the formula, Penalize costs for time windows; Penalty costs for early arrival; Penalty costs for late arrival; Step S815: The calculation formula of the line side inventory penalty cost is shown in formula (19): (19); In the formula, Line side inventory penalty costs; Penalty cost per unit volume of line-side inventory; In order to minimize the total cost of material distribution, the objective function is constructed as shown in formula (20): (20); Where: is the total delivery cost; Step S82: constructing constraint conditions; Based on the optimization process in line with the actual production operation conditions, the following constraints are given: Constraint 1: , indicating that the reorder point value of each workstation must not exceed the value range; Constraint 2: , which means that the delivery vehicle starts from the warehouse every time and returns to the warehouse after completing the delivery task; where, Indicates delivery vehicle In the A delivery cycle from the warehouse to the workstation ; Indicates delivery vehicle In the Delivery cycle from workstation Drive to the warehouse; Constraint three: , indicating that the material distribution path must be connected, that is, the distribution vehicle arrives at the workstation , must be from the workstation Leave; in the formula, Indicates delivery vehicle In the Delivery cycle from workstation Travel to the workstation ; Indicates delivery vehicle In the Delivery cycle from workstation Travel to the workstation ; Constraint 4: , indicating that in each delivery cycle, there is only 0 or 1 delivery vehicle as a workstation Services cannot be split into two or more vehicles; Constraint 5: , indicating that the total loading volume of the delivery vehicle for each delivery shall not be greater than the maximum loading volume of the delivery vehicle; where, The maximum loading volume of the delivery vehicle, in m 3 ; Constraint six: , indicating that the number of line-side material boxes at each workstation must not exceed the corresponding maximum number of line-side inventory material boxes; Constraint seven: , which means that the order between the two workstations delivered by the same delivery vehicle must be sequential. Only after the previous workstation is delivered, the next workstation can be delivered. In the formula, For the Delivery cycle to deliver materials to the workstation moment; For the Delivery cycle to deliver materials to the workstation moment; is an infinite value; Constraint 8: , indicating the workstation The material delivery must be completed within the specified time window.

[0032] Furthermore, in step S9, based on the distribution optimization model under the time-varying workstation set constructed in step S8, the distribution optimization model is solved by using the inner genetic algorithm, and steps S5 to S7 are repeated until all material distribution tasks within the distribution time range specified in S1 are completed, and the total distribution planning scheme under the current workstation reorder point is output; specifically: The inner genetic algorithm is part of the double-layer genetic algorithm. It is mainly responsible for solving the corresponding material distribution path, distribution vehicle departure time and total distribution cost for each set of reorder point plans based on the distribution optimization model under the time-varying workstation set. ,In the solution process, the encoding method selects the real number encoding method, and the genetic operations include the roulette selection strategy, the two-point crossover strategy and the two-point mutation strategy; The basic idea of ​​the inner genetic algorithm solution is: based on the combination of workstation reorder points , within the time range of material distribution pre-set by S1, according to the intersection between the service time windows of the workstations, the time-varying workstation set is determined when each distribution task is executed, and the material distribution path is planned by the inner genetic algorithm. Through continuous iterative calculation, the solution with the minimum total distribution cost is selected as the optimal material distribution path for the distribution cycle. At the same time, after the planning of each distribution cycle is determined, the workstation information is updated in real time, and the next distribution cycle is solved until all distribution tasks within the pre-set time range for material distribution are completed, and the workstation re-order point combination is output. The optimal material distribution path, distribution vehicle departure time, various distribution costs and total distribution cost for each distribution cycle under .

[0033] Further, in step S10, the overall material distribution planning scheme obtained in step S9 is recorded, the workstation reorder point is updated using the outer genetic algorithm, and steps S4 to S9 are repeated until the maximum number of iterations is reached, and the optimal distribution scheme including the workstation reorder point combination and the distribution plan is output; specifically: The outer genetic algorithm is part of the double-layer genetic algorithm and is mainly responsible for adjusting the reorder point combination of the workstation , as the basis of distribution planning, the real number encoding method is selected in the solution process, and the genetic operations include ternary tournament selection strategy, uniform crossover strategy and multi-point mutation strategy; The basic idea of ​​solving the outer genetic algorithm is: in each iteration, the value interval of the reorder point of the workstation is , using genetic operations to update the station reorder point combination , and use the inner genetic algorithm to determine the corresponding distribution plan, record the optimal workstation reorder point combination and the corresponding material distribution plan in each iteration, and compare the total distribution cost of the distribution plan with the optimal total distribution cost in the previous iteration. If the distribution plan under the new reorder point combination can reduce the total distribution cost, then update the current optimal reorder point combination to the new workstation reorder point, and repeat the iteration until the set maximum number of iterations is met. Finally, the optimal workstation reorder point combination in all iterations and the optimal material distribution path, distribution vehicle departure time, various distribution costs and total distribution cost of each distribution cycle of the corresponding distribution plan are obtained as the final distribution plan.

[0034] The following is a specific example of material distribution within a single shift of a product assembly line in a machine branch factory of a manufacturing enterprise to illustrate the distribution planning effect of the invention method: The company's work system is a four-shift three-operation system. The material distribution adopts the whole box feeding mode. The product assembly line layout is linear, including 1 material warehouse and 14 workstations, each corresponding to a material type. The material distribution task is carried out by a uniform distribution vehicle according to the hard time window standard. Due to the workshop road restrictions, the distribution vehicle speed is constant at 83.33m / min. The distribution vehicle can carry up to 3 tooling vehicles, and the total maximum loadable volume is 6.75m 3 , and the service time of the delivery vehicle at each station is 2 minutes, and the loading time in the warehouse is 0.5 minutes per box. The cost parameters in the delivery process are: the single startup cost of the delivery vehicle =30 yuan / time; driving cost =0.1 yuan / m; line side inventory unit volume cost =0.2 yuan / min; early arrival penalty cost =1000 yuan / min; penalty cost for late arrival =1000 yuan / min; penalty cost per unit volume of line inventory =0.4 yuan / min.

[0035] At the same time, the following related assumptions are made for the research questions: (1) Delivery-related assumptions: All delivery vehicles are of the same type, with a constant driving speed, a fixed start-up cost per delivery vehicle, and no consideration of traffic jams or vehicle breakdowns during the delivery process.

[0036] (2) Workstation-related assumptions: The initial line inventory of each workstation is in a full inventory state, and the service time and material consumption rate of each workstation are known. At the same time, in each delivery, materials at the same workstation shall not be split for delivery.

[0037] (3) Material-related assumptions: The capacity of the delivery vehicle and the line-side inventory capacity are both expressed in units of bins. When materials are in use, they can only be replaced at the next delivery after the materials in the bins are exhausted.

[0038] In addition to the above parameters, the other relevant parameters of each station are shown in Table 1: including the volume of the material box of each station 、Quantity that can be loaded per unit box , Maximum number of material boxes at the line side And the usable time of each unit of material ; The driving distance between each workstation is shown in Table 2: It shows the distance between each workstation and between the workstation and the material warehouse, where 0 represents the material warehouse and 1-14 represents each workstation.

[0039] Table 1 Workstation related parameters

[0040] Table 2 Travel distance between workstations

[0041] The double-layer genetic algorithm is used to solve the above example. The parameter setting is divided into the upper genetic algorithm and the lower genetic algorithm. The upper genetic algorithm parameters are: population size U N is 50, the maximum number of iterations U G is 200, and the crossover probability U PM is 0.9, the mutation probability U PC is 0.1; the parameters of the lower genetic algorithm are set as: population size L N is 50, the maximum number of iterations is L G is 50, the crossover probability is L PM is 0.85, and the crossover probability L PC is 0.05.

[0042] Algorithm iteration diagram see Figure 2 ,With the gradual increase of the number of algorithm iterations, the total ,distribution cost shows a decreasing trend. After 157 iterations, the optimal solution of the ,distribution scheme is obtained. By decoding the optimal solution of the ,distribution scheme, the calculation results are shown in Table 3.

[0043] Table 3 Optimal solution of algorithm iteration

[0044] The calculation results in Table 3 include three parts: the top part shows the optimal reorder point combination of workstations, ensuring that material distribution can meet the material needs of each workstation at a lower cost; the middle part shows the material distribution path arrangement and the departure time of the distribution vehicle in each cycle based on the reorder point of the workstation. Each material distribution is sorted according to the workstation and executed according to the departure time in brackets. For example: 0-6-4-0 (68.918), which means that the distribution vehicle departs from the warehouse at 68.918 minutes, delivers materials to workstations 6 and 4 in turn, and returns to the warehouse after completing the distribution. From the results obtained in this section, it can be seen that based on the corresponding workstation reorder points, material distribution within a shift can be divided into 14 delivery cycles, including 27 delivery tasks in total, and the material distribution workstation set of the delivery vehicles in each cycle presents time-varying characteristics, highlighting the flexibility and effectiveness of the algorithm in responding to changes in actual demand; the bottom part shows that the total cost of material distribution within a shift is 4173.78 yuan, the number of delivery vehicles required for delivery is 4, and the time window penalty cost value in the cost component item is 0, which shows that each delivery can complete the delivery task according to the delivery requirements within the specified time, meet various constraints, and achieve the optimization of the delivery plan.

[0045] In order to illustrate the superiority of the method, the results in the above table are compared and analyzed with the traditional material distribution strategy adopted by the workshop. At present, the material distribution strategy of the workshop adopts a fixed-station distribution mode. Specifically, the production workshop is equipped with 5 distribution vehicles, each of which is responsible for the material distribution task of a fixed station. The station allocation remains unchanged. At the same time, the reorder point of each station is set to the material loading volume of the corresponding station unit material box. When a station issues a material distribution request, the distribution staff responsible for the corresponding station will comprehensively check the line inventory of each station in charge, and determine the material replenishment quantity of each station based on the overall inventory situation. Finally, unified material distribution is implemented according to the time when the distribution request is initiated.

[0046] In addition, since the traditional distribution strategy does not consider the time window association between each workstation when setting the delivery departure time, and only executes material delivery according to the time when the earliest delivery request is initiated, the time window penalty cost is not considered in the comparison. The final comparison results are shown in Table 4. The results on the left are the traditional material distribution strategy, and the results on the right are the material distribution strategy in the present invention.

[0047] Table 4 Comparison of results under two distribution strategies

[0048] As can be seen from the above table, the present invention shows significant superiority over traditional strategies. In terms of delivery vehicle driving, the present invention reduces the startup cost and driving cost of delivery vehicles by 12.90% and 8.72% respectively by optimizing the transportation route and delivery vehicle scheduling, and successfully reduces the number of delivery vehicles required for delivery, significantly improving transportation efficiency. In terms of online inventory management, although the online inventory penalty cost has increased slightly, compared with the substantial decrease of 21.12% in the online inventory cost, the advantages of the present invention in inventory management are more prominent. Through the dual optimization of the delivery vehicle driving process and the online inventory management process, the total material distribution cost has been reduced from 4746.7 yuan to 4173.78 yuan, a reduction of 12.07%, which fully demonstrates that the present invention can effectively respond to fluctuations in production demand by flexibly adjusting the material distribution plan, and ultimately help enterprises achieve the goal of reducing costs and increasing efficiency, showing high adaptability and flexibility.

[0049] Overall, the present invention sets the workstation reorder point, considers the dynamic coordination of distribution decision elements, combines the time-varying workstation set, and takes the minimization of the total distribution cost as the goal. It uses a double-layer genetic algorithm to jointly optimize the setting problem of the workstation reorder point and the corresponding material distribution plan. While reducing the material distribution cost, the distribution plan can meet the dynamic needs of each workstation, improving the efficiency and dynamism of the distribution process. In addition, by combining time-varying characteristics with actual production data, the present invention can adapt to different production situations and application scenarios through intelligent decision-making, which can not only meet the multiple constraints brought by the complex environment of the workshop, but also improve the fit between material distribution and the production process, bringing new perspectives and methods to the research of workshop material distribution.

[0050] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent material distribution in a production workshop based on a time-varying workstation set, characterized in that: The following steps are involved: Step S1: Collect the workstation information and delivery vehicle information of the production workshop to determine the time range in which material delivery needs to be executed; Step S2: Based on the workstation information and delivery vehicle information collected in step S1, the value range of the workstation reorder point is limited; Step S3: Based on the workstation reorder point value range set in step S2, a reasonable workstation reorder point is randomly set for each workstation, and the combination becomes a workstation reorder point combination; Step S4: Based on the workstation reorder point combination and the workstation information collected in step S1, the workstation service time window is calculated, and the delivery workstation set of the current delivery cycle is determined based on the cross relationship of the workstation service time windows; Step S5: Determine a method for calculating the departure time of the delivery vehicle based on the specific number of delivery stations of the delivery vehicle, and determine a method for calculating the number of materials delivered to the station based on the arrival time of the station; Step S6: Based on the set of delivery stations determined in step S4 and the calculation method of the departure time of the delivery vehicle and the quantity of material delivered to the station determined in step S5, the forward method is used to calculate and evaluate the loading status of the delivery vehicle in real time, plan the material delivery path, and obtain the distribution planning result under the time-varying characteristics of the current delivery cycle; Step S7: Based on the distribution planning result under the time-varying characteristics of the current distribution cycle obtained in step S6, the service time window of the station and the distribution station set of the next distribution cycle are updated according to step S4, and the distribution station of the end distribution vehicle of the current distribution cycle is collaboratively optimized with the adjacent distribution station set by comparing the service time window of the distribution station of the end distribution vehicle of the current distribution cycle with the earliest departure time of the next distribution cycle; Step S8: Based on the workstation information and delivery vehicle information collected in step S1, a delivery optimization model under a time-varying workstation set is constructed; Step S9: Based on the distribution optimization model under the time-varying workstation set constructed in step S8, the distribution optimization model is solved by using the inner genetic algorithm, and steps S5 to S7 are repeated until all material distribution tasks within the time range determined in step S1 to be required for material distribution are completed, and the total distribution planning scheme under the current workstation re-order point is output; Step S10: Record the overall distribution plan obtained in step S9, use the outer genetic algorithm to update the workstation reorder point, repeat steps S4-S9 until the maximum number of iterations is reached, and output the optimal distribution plan including the workstation reorder point combination and distribution planning.

2. According to claim 1, a production workshop intelligent material distribution method based on a time-varying workstation set is characterized by: In step S1, the workstation information and delivery vehicle information of the production workshop are collected to determine the time range for material delivery; specifically: Workstation information includes: the number of workstations, the travel distance between workstations, the available time of materials per workstation, the service time of the workstation, the volume of the workstation material box, the maximum number of material boxes in the line inventory of the workstation, and the loadable number of material boxes per workstation; The delivery vehicle information includes: the driving speed of the delivery vehicle, the maximum loading volume of the delivery vehicle; The time frame within which material delivery needs to be performed is pre-set.

3. According to claim 2, a production workshop intelligent material distribution method based on a time-varying workstation set is characterized by: In step S2, based on the workstation information and delivery vehicle information collected in step S1, the value range of the workstation reorder point is limited; specifically: The minimum value level of the reorder point of the workstation is set as shown in formula (1): （1）； The maximum value level of the reorder point of the workstation is set as shown in formula (2): （2）； Where: For workstation reorder point; It is the minimum reorder point level of the workstation; For workstation The maximum value level of the reorder point; * is the multiplication sign; Is the station number, belonging to the collection of all stations ; is the collection of all workstations; The number of the dispatched delivery vehicle, belonging to the collection of all delivery vehicles ; It is the collection of all delivery vehicles; For warehouse and workstation The distance between them, unit: m; is the driving speed of the delivery vehicle, unit: m / min; For workstation The usable time of a unit material, unit: s; For workstation Maximum number of bins in line side inventory; For workstation The amount of material that can be loaded into a unit container.

4. According to claim 3, a production workshop intelligent material distribution method based on a time-varying workstation set is characterized in that: In step S3, based on the workstation reorder point value range set in step S2, a reasonable workstation reorder point is randomly set for each workstation, and the combination becomes a workstation reorder point combination; specifically: For each workstation , based on the value range of the reorder point of the workstation , randomly select an integer in the interval of the workstation reorder point as the workstation Reorder point , and combine the reorder points of each workstation to form a workstation reorder point combination .

5. According to claim 4, a method for intelligent material distribution in a production workshop based on a time-varying workstation set is characterized in that: In step S4, based on the workstation reorder point combination and the workstation information collected in step S1, the workstation service time window is calculated, and the delivery workstation set of the current delivery cycle is determined based on the cross relationship of the workstation service time window; specifically: Step S41: Calculation of workstation service time window; The workstation service time window consists of the delivery arrival time window and the delivery departure time window. Based on the time-varying characteristics, the setting standard of the workstation service time window is determined. The specific calculation method is as follows: The earliest delivery arrival time is set as shown in formula (3): （3）； In the formula, For workstation No. The earliest arrival time of the delivery; For the Deliver materials to the workstation moment; For workstation No. The quantity of materials after the first delivery; For workstation A delivery belongs to the workstation The number of material delivery times ; For workstation The collection of material delivery times; The latest arrival time of delivery is set as shown in formula (4): （4）； In the formula, For workstation No. The latest arrival time of the delivery; For workstation service time; Based on the delivery arrival time window, the delivery departure time window of the workstation is calculated, which is divided into the earliest departure time of delivery and the latest departure time of delivery, which are expressed by formula (5) and formula (6) respectively: （5）； （6）； In the formula, For workstation No. The earliest departure time for the delivery; For workstation No. The latest departure time for the delivery; Step S42: dividing the delivery station set based on the intersection of time windows; Based on the cross-relationship between the workstation service time windows, the workstations that need to perform material distribution in the current distribution cycle are divided into distribution workstation sets. The division method is as follows: A collection of all workstations The earliest workstation to initiate a delivery request is set to , marking the earliest departure time as The latest departure time is ; To remove the workstation that initiated the earliest delivery request Each workstation outside If the workstation Current Earliest Departure Time Less than the latest departure time , then the workstation It is classified into the same delivery station set as the first station, and the expression is shown in formula (7): （7）； In the formula, For the A set of delivery stations.

6. The method for intelligent material distribution in a production workshop based on a time-varying workstation set according to claim 5, characterized in that: In step S5, the method for calculating the departure time of the delivery vehicle is determined based on the specific number of delivery stations of the delivery vehicle, and the method for calculating the number of materials delivered to the station is determined based on the arrival time of the station; specifically: Step S51: Calculation method of the delivery vehicle departure time; Select the delivery departure time window and delivery arrival time window as the basis to calculate the delivery vehicle departure time. Assuming that the delivery vehicle In the The departure time of the delivery cycle is The method for calculating the departure time of the delivery vehicle is divided into two cases: step S511 and step S512: Step S511: When the number of workstations to be delivered is equal to 1, it is assumed that the delivery workstation is workstation , the calculation method of the delivery vehicle departure time is shown in formula (8), which is No. The earliest departure time of a delivery cycle; （8）； In the formula, For delivery vehicles In the The departure time of each delivery cycle; is a certain delivery cycle, belonging to the set of all delivery cycles ; is the collection of all delivery cycles; is a 0-1 variable: if Delivery cycle delivery vehicles Responsible for delivering materials to workstations but is 1, otherwise is 0; Is a 0-1 variable: If the workstation No. The second delivery belongs to Delivery cycle is 1, otherwise is 0; Step S512: When the number of workstations to be delivered is greater than 1, the latest arrival time of the delivery vehicle is used as the calculation basis, and the departure time of the delivery vehicle is obtained by the reverse calculation method; specifically: Step S5121, for each workstation , according to the workstation of the current delivery cycle No. The latest arrival time of the delivery The values ​​of are sorted in descending order, and the largest station number among the values ​​of the latest arrival time is , the second largest station number is , then the workstation No. Delivery cycle arrival time It is expressed as formula (9): （9）； In the formula, For the Delivery cycle delivery vehicles Responsible for delivery workstation The latest arrival time; For workstation and workstation The distance between For workstation service time; Step S5122, remove the workstation , the original workstation Update to new workstation , from the latest arrival time Select the workstation in descending order The corresponding next station is used as the new station , execute formula (9) again until the arrival time of all delivery vehicles at all delivery stations of the delivery vehicle is calculated; Step S5123, based on the final calculated workstation No. Delivery cycle arrival time , determine the delivery vehicle In the Departure time of delivery cycle , expressed as formula (10): （10）； In the formula, is the distance between the warehouse and workstation b; Step S52: a method for calculating the quantity of material delivered to a workstation; When delivery vehicle In the Departure time of delivery cycle After confirmation, the delivery frequency of each workstation is tracked and recorded to identify the Delivery cycle to workstation The delivery belongs to the workstation The number of deliveries is then further determined based on the driving distance between workstations and the speed of the delivery vehicle. No. The arrival time of the delivery ; Workstation No. The time of initiating the material request is calculated by formula (11): （11）； Where: For workstation No. The time when the material request was first initiated; Workstation No. The specific quantity of remaining materials at the time of the next delivery is calculated by formula (12): （12）； Where: For workstation No. The specific quantity of remaining materials at the time of the next delivery arrival; Finally, the workstation No. The number of remaining available containers and workstations at the time of the next delivery No. The number of containers delivered each time is calculated using formula (13) and formula (14): （13）； （14）； In the formula, For workstation No. The number of available containers remaining at the time of the next delivery arrival. For workstation No. The number of boxes delivered.

7. The method for intelligent material distribution in a production workshop based on a time-varying workstation set according to claim 6, characterized in that: In step S6, based on the set of delivery stations determined in step S4 and the calculation method of the departure time of the delivery vehicle and the quantity of material delivered to the station determined in step S5, the forward method is used to calculate and evaluate the loading status of the delivery vehicle in real time, plan the material delivery path, and obtain the distribution planning result under the time-varying characteristics of the current distribution cycle; specifically: Step S61, set the station set that needs to be delivered to , the number of workstations in the distribution workstation concentration is , the set of all delivery vehicles is , delivery vehicles The material distribution path is , the maximum loading volume of the delivery vehicle is , the delivery vehicle departure time is ; When starting path planning, the workstation The first delivery station and the initial number 1, delivery vehicle This is the first delivery vehicle and the initial number is 1, and the initial material delivery path of the delivery vehicle is empty; Step S62: Add delivery vehicle Material delivery path and calculate the corresponding delivery vehicle departure time and the current loading volume of the delivery vehicle ; Step S63, then evaluate the loading status of the delivery vehicle: Step S631: If and :The current loading volume of the delivery vehicle does not exceed the maximum loading volume of the delivery vehicle and there is a next workstation, then the workstation Add to delivery vehicle Update the material distribution path of for , and go to step S62; Step S632: If and : If the current loading volume of the delivery vehicle does not exceed the maximum loading volume of the delivery vehicle and there is no next station, the delivery vehicle is recorded The complete material distribution path of the distribution vehicle is obtained by obtaining the material distribution paths of all distribution vehicles in the current distribution cycle; Step S633: If and : If the current loading volume of the delivery vehicle exceeds the maximum loading volume of the delivery vehicle and there is a next workstation, the workstation will be removed from the current material delivery path. The remaining workstations are recorded as delivery vehicles. The complete material distribution path is updated at the same time for , the workstation As a delivery vehicle Delivery station, update for , and go to step S6; Step S634: If and : If the current loading volume of the delivery vehicle exceeds the maximum loading volume of the delivery vehicle and there is no next workstation, the workstation will be removed from the current material delivery path. The remaining workstations are recorded as delivery vehicles. Complete material distribution path, and the workstation As a delivery vehicle The distribution station can obtain the material distribution paths of all distribution vehicles in the current distribution cycle.

8. The method for intelligent material distribution in a production workshop based on a time-varying workstation set according to claim 7, characterized in that: In step S7, based on the distribution planning result under the time-varying characteristics of the current distribution cycle obtained in step S6, the station service time window and the distribution station set of the next distribution cycle are updated according to step S4, and the distribution station of the end distribution vehicle of the current distribution cycle is collaboratively optimized with the adjacent distribution station set by comparing the service time window of the distribution station of the end distribution vehicle of the current distribution cycle with the earliest departure time of the next distribution cycle; specifically, as follows: Step S61, assuming that the delivery vehicle that departs from the warehouse the latest in the current delivery cycle is the delivery vehicle , delivery vehicles All workstations responsible for delivery are combined into a workstation group , workstation group The number of workstations in indivual; Step S62, define the delivery vehicles in the current delivery cycle Delivery team The latest departure time for each workstation is , the earliest departure time for the next delivery cycle is ; Step S63, when , then the delivery vehicle Delivery team Co-optimize with adjacent delivery station sets; Step S64: If the delivery vehicle Delivery team The latest departure time for each workstation Both are greater than the earliest departure time of the next delivery cycle , then the delivery vehicle Delivery team The various workstations in the distribution cycle are merged into the next distribution cycle for joint distribution, and the workstation data is updated; otherwise, the original material distribution path is maintained.

9. The method for intelligent material distribution in a production workshop based on a time-varying workstation set according to claim 8, characterized in that: In step S8, based on the workstation information and delivery vehicle information collected in step S1, a delivery optimization model under a time-varying workstation set is constructed; specifically: Step S81: constructing the objective function; taking the delivery vehicle startup cost, delivery vehicle driving cost, line-side inventory cost, time window penalty cost and line-side inventory penalty cost into account in the total cost, with minimizing the total cost of material distribution as the overall goal; Step S811: The calculation formula for the delivery vehicle startup cost is shown in formula (15): （15）； In the formula, start-up costs for delivery vehicles; The single startup cost of the delivery vehicle; is the total number of starts of the delivery vehicle; Step S812: The calculation formula of the delivery vehicle driving cost is shown in formula (16): （16）； In the formula, The driving cost of the delivery vehicle; is the unit driving cost of the delivery vehicle; 0-1 variable: If the delivery vehicle In the Delivery cycle from workstation Travel to the workstation but is 1, otherwise is 0; For workstation With workstation The distance between Step S813: The calculation formula of line-side inventory cost is shown in formula (17): （17）； In the formula, is the line-side inventory cost; is the unit volume cost of line-side inventory; For workstation No. The quantity of material when the first delivery arrives; For workstation The volume of the material box, in m 3 ; For the Deliver materials to the workstation moment; Step S814: The calculation formula of the time window penalty cost is shown in formula (18): （18）； In the formula, Penalize costs for time windows; Penalty costs for early arrival; Penalty costs for late arrival; Step S815: The calculation formula of the line side inventory penalty cost is shown in formula (19): （19）； In the formula, Line side inventory penalty costs; Penalty cost per unit volume of line-side inventory; In order to minimize the total cost of material distribution, the objective function is constructed as shown in formula (20): （20）； Where: is the total delivery cost; Step S82: constructing constraint conditions; Based on the optimization process in line with the actual production operation conditions, the following constraints are given: Constraint 1: , indicating that the reorder point value of each workstation must not exceed the value range; Constraint 2: , which means that the delivery vehicle starts from the warehouse every time and returns to the warehouse after completing the delivery task; where, Indicates delivery vehicle In the A delivery cycle from the warehouse to the workstation ; Indicates delivery vehicle In the Delivery cycle from workstation Drive to the warehouse; Constraint three: , indicating that the material distribution path must be connected, that is, the distribution vehicle arrives at the workstation , must be from the workstation Leave; in the formula, Indicates delivery vehicle In the Delivery cycle from workstation Travel to the workstation ; Indicates delivery vehicle In the Delivery cycle from workstation Travel to the workstation ; Constraint 4: , indicating that in each delivery cycle, there is only 0 or 1 delivery vehicle as a workstation Services cannot be split into two or more vehicles; Constraint 5: , which means that the total loading volume of the delivery vehicle for each delivery shall not be greater than the maximum loading volume of the delivery vehicle; where, The maximum loading volume of the delivery vehicle, in m 3 ; Constraint six: , indicating that the number of line-side material boxes at each workstation must not exceed the corresponding maximum number of line-side inventory material boxes; Constraint seven: , which means that the order between the two workstations delivered by the same delivery vehicle must be sequential. Only after the previous workstation is delivered, the next workstation can be delivered. In the formula, For the Delivery cycle to deliver materials to the workstation moment; For the Delivery cycle to deliver materials to the workstation moment; is an infinite value; Constraint 8: , indicating the workstation The material delivery must be completed within the specified time window.

Images