Search

CN-116362403-B - Yield allocation optimization method and system based on mathematical model

CN116362403BCN 116362403 BCN116362403 BCN 116362403BCN-116362403-B

Abstract

The invention provides a capacity allocation optimization method and a system based on a mathematical model, wherein the method comprises the steps of constructing a corresponding relation between an operation machine group and a process route, acquiring standard operation time of each operation site on the process route, acquiring available production capacity of each operation machine group in a future period, acquiring work-in-process distribution condition of each operation site and a film throwing plan in the future period, and constructing a capacity allocation model as follows: And setting a coordination coefficient to obtain an optimal solution of the capacity allocation scheme. According to the invention, the capacity allocation is optimized based on the mathematical model, and a production decision maker can set the coordination coefficient according to the requirement to obtain the optimal solution of the capacity allocation scheme, so that the problem that the production decision maker can not consider all production indexes easily because of manually making a daily capacity allocation scheme according to personal production experience is avoided.

Inventors

  • Zhang Jiajiong
  • ZHANG FENG
  • CHENG JIE
  • CAO XINYU
  • Wei Mengru

Assignees

  • 上海华力微电子有限公司

Dates

Publication Date
20260512
Application Date
20230331

Claims (10)

  1. 1. The capacity allocation optimization method based on the mathematical model is characterized by comprising the following steps of: constructing a corresponding relation between the operation machine group and the process route; Obtaining standard operation time of each operation site on the process route; acquiring available production capacity of each working machine group in a future period of time; acquiring the work-in-process distribution condition of each operation site and a film throwing plan of a future period of time; the capacity allocation model is constructed as follows: F= (1) constraint conditions: (2) (3) (4) (5) (6) ; (7) In the formula, To be a co-ordination of yield level and throughput; For the number of job sites, Represent the first A plurality of job sites; for the total number of product types, Represent the first The individual product types; in order to produce the planned period of time, Represent the first A day; is the first The individual products are of the first category On the first day The number of jobs at each job site; is the first The individual products are of the first category Day arrive at the first The work-in-process number of the individual job sites; is the first The individual products are of the first category On the first day The work-in-process number of the individual job sites; is the first The variety of the individual products is at the beginning of the period Work-in-process inventory of individual job sites; is the first The individual products are of the first category Wafer input amount per day; For the number of work machines to be grouped, Represent the first A plurality of work machines; is the first Day 3 The available capacity of the individual work machine group; is the first The individual products are of the first category Time spent working by the individual working sites; Represent if No. 1 The job site is formed by The operation of each operation machine group is 1, otherwise, the operation machine group is 0; and setting a coordination coefficient to obtain an optimal solution of the capacity allocation scheme.
  2. 2. The capacity allocation optimization method based on the mathematical model according to claim 1, wherein an optimal solution of the capacity allocation model is obtained by adopting a branch-and-bound algorithm.
  3. 3. The capacity allocation optimization method based on the mathematical model according to claim 1, wherein the formula (2) is used to represent the flow balance between any product and the upstream and downstream operation sites.
  4. 4. The capacity allocation optimization method based on the mathematical model of claim 1, wherein the formula (3) is used for representing dynamic balance among initial process route start job site dosage, initial work-in-process level, current work-in-process output and work-in-process level.
  5. 5. The capacity allocation optimization method based on the mathematical model of claim 1, wherein the formula (4) is used for representing a dynamic balance among an initial work-in-process level, a current work-in-process output and a work-in-process level of each work site except the initial work site of the preliminary process route.
  6. 6. The capacity allocation optimization method based on the mathematical model according to claim 1, wherein the formulas (5) and (6) are used for representing the dynamic balance among the work-in-process level, the current input amount, the current output amount and the tablet feeding amount of all the working sites of the process route between the previous and the next production days.
  7. 7. The capacity allocation optimization method based on the mathematical model of claim 1, wherein the formula (7) is used to represent that the capacity requirement of each work machine group does not exceed the available capacity level of each product at each production day.
  8. 8. A capacity allocation optimization system based on a mathematical model, comprising: The construction module is configured to construct the corresponding relation between the working machine group and the process route; the data acquisition module is configured to acquire standard operation time of each operation site on the process route, available production capacity of each operation machine group in a future period of time, work-in-process distribution condition of each operation site and a film throwing plan in the future period of time; the capacity allocation module is configured to construct a capacity allocation model as follows: F= (1) constraint conditions: (2) (3) (4) (5) (6) ; (7) In the formula, To be a co-ordination of yield level and throughput; For the number of job sites, Represent the first A plurality of job sites; for the total number of product types, Represent the first The individual product types; in order to produce the planned period of time, Represent the first A day; is the first The individual products are of the first category On the first day The number of jobs at each job site; is the first The individual products are of the first category Day arrive at the first The work-in-process number of the individual job sites; is the first The individual products are of the first category On the first day The work-in-process number of the individual job sites; is the first The variety of the individual products is at the beginning of the period Work-in-process inventory of individual job sites; is the first The individual products are of the first category Wafer input amount per day; For the number of work machines to be grouped, Represent the first A plurality of work machines; is the first Day 3 The available capacity of the individual work machine group; is the first The individual products are of the first category Time spent working by the individual working sites; Represent if No. 1 The job site is formed by The operation of each operation machine group is 1, otherwise, the operation machine group is 0; The capacity allocation module is further configured to set a coordination coefficient to obtain an optimal solution of the capacity allocation scheme.
  9. 9. The capacity allocation optimization system based on the mathematical model of claim 8, wherein an optimal solution of the capacity allocation model is obtained by using a branch-and-bound algorithm.
  10. 10. A readable storage medium, on which a computer program is stored, characterized in that the computer program, when executed, is capable of implementing a mathematical model-based capacity allocation optimization method according to any one of claims 1-7.

Description

Yield allocation optimization method and system based on mathematical model Technical Field The invention relates to the technical field of semiconductor manufacturing, in particular to a capacity allocation optimization method and system based on a mathematical model. Background The integrated circuit manufacturing system mainly comprises 1) scheduling of a film throwing plan and a delivery plan, 2) allocation of daily equipment capacity (workload allocation of a plurality of production sections of the same working machine group in a certain time period), and 3) scheduling and scheduling of the working machine group. The scheduling of the film feeding plan and the delivery plan is formulated according to the work-in-process level control of one to two months in the future of the whole factory, the allocation of the daily equipment capacity is a short-term (generally one day) production target formulated according to the factors of the exchange period target, the production period, the output requirement and the like, and the cluster scheduling and scheduling is to implement the short-term (10 minutes or real-time) production plan according to the factors of the capacity allocation target of the equipment, the capacity utilization requirement of the equipment and the like. The allocation of daily plant capacity plays an important role as a linking module for the upper production plan and the lower execution plan. Therefore, research on capacity allocation has received much attention. Such as Chung et al, have studied the goal of maximizing throughput by balancing the production plan to create a daily capacity allocation scheme for a bottleneck plant population. In addition, lee and the like divide all process routes into sections with a section separated from each other according to the photoetching layers, and estimate the target work-in-process level corresponding to each photoetching layer by utilizing a statistical analysis mode, so as to continuously reduce the difference between the target work-in-process level and the current work-in-process level by utilizing a real-time control mode, and realize the production line balance. Similarly Bureauetal, etc. are also achieved by narrowing the gap between the target and current work-in-process levels of each photolithographic layer, where the target work-in-process level is obtained by means of simulation. Regardless of the manner in which it is employed, accurate acquisition of the target work in progress level is often difficult to achieve. Therefore, a learner designs a new combined scheduling rule by adopting a scheduling rule mode, such as Zhou and the like, and comprehensively considering scheduling rules such as traffic priority, shortest processing time priority and the like so as to realize comprehensive allocation of capacity and coordination among a plurality of production indexes of a production line. In summary, most of the existing researches comprehensively adopt modes of statistical analysis, simulation rules, heuristic scheduling rules and the like to realize the allocation of daily capacity of a production line, however, due to the complex production operation characteristics of integrated circuit manufacturing, the modes are difficult to consider all production indexes. Disclosure of Invention The invention aims to provide a capacity allocation optimization method and system based on a mathematical model, which at least solve one of the technical problems existing in the prior art. In order to achieve the above object, the present invention provides a capacity allocation optimization method based on a mathematical model, comprising: constructing a corresponding relation between the operation machine group and the process route; Obtaining standard operation time of each operation site on the process route; acquiring available production capacity of each working machine group in a future period of time; acquiring the work-in-process distribution condition of each operation site and a film throwing plan of a future period of time; the capacity allocation model is constructed as follows: constraint conditions: Xgtl＝Ygt(l-1) g=1,2,...G;t=1,2,...T;l=2,3,...,L (2) Wg11＝IWg1+Rg1-Yg11 g=1,2,...G; (3) Wg1l＝IWgl-Yg1l g=1,2,...G;l=2,3,...,L (4) Wgt1＝Wg(t-1)1+Rgt-Ygt1 g=1,2,...G;t=2,3,...T (5) Wgtl＝Wg(t-1)l+Ygtl+Xgtl g=1,2,...G;t=2,3,...T;l=2,3,...,L (6) Wherein delta is a coordination coefficient of the yield level and the production amount, L is the number of operation sites, L is the first operation site, G is the total product type, G is the G product type, T is the production planning period, T is the T day, Y gtl is the operation number of the G product type at the first operation site on the T day, X gtl is the work-in number of the G product type reaching the first operation site on the T day, W gtl is the work-in number of the G product type at the first operation site on the T day, IW gl is the work-in stock quantity of the G product type at