CN111445079B - Multi-fidelity simulation optimization method and equipment applied to workshop plan production - Google Patents

Multi-fidelity simulation optimization method and equipment applied to workshop plan production Download PDF

Info

Publication number
CN111445079B
CN111445079B CN202010245660.1A CN202010245660A CN111445079B CN 111445079 B CN111445079 B CN 111445079B CN 202010245660 A CN202010245660 A CN 202010245660A CN 111445079 B CN111445079 B CN 111445079B
Authority
CN
China
Prior art keywords
fidelity
low
production
solution
simulation model
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010245660.1A
Other languages
Chinese (zh)
Other versions
CN111445079A (en
Inventor
岳磊
管在林
张正敏
王创剑
周洋
田亚娟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huazhong University of Science and Technology
Original Assignee
Huazhong University of Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huazhong University of Science and Technology filed Critical Huazhong University of Science and Technology
Priority to CN202010245660.1A priority Critical patent/CN111445079B/en
Publication of CN111445079A publication Critical patent/CN111445079A/en
Application granted granted Critical
Publication of CN111445079B publication Critical patent/CN111445079B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/067—Enterprise or organisation modelling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/04—Manufacturing
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Abstract

The invention belongs to the technical field of production control, and particularly relates to a multi-fidelity simulation optimization method and equipment applied to workshop planning production. The method optimizes the workshop plan production problem by using two simulation models with fidelity levels in a matching way: firstly, establishing a high-fidelity and low-fidelity simulation model according to an actual production system; then, searching a solution space of the production plan problem by using a low-fidelity simulation model and combining the optimization iteration characteristic of the GA algorithm; sampling by using a sampling method with ordinal number conversion and optimal sampling according to a result obtained by the operation of the low-fidelity model to obtain an optimal sampling set; and finally, running the optimal sampling set through the high-fidelity simulation model, wherein the high-fidelity model has extremely high similarity with an actual production system of a workshop, and running the production scheme in the optimal sampling set by using the high-fidelity model can obtain a relatively accurate result, so that the result can be used as a basis for selecting the optimal production scheme.

Description

Multi-fidelity simulation optimization method and equipment applied to workshop plan production
Technical Field
The invention belongs to the field of production control, relates to a multi-fidelity simulation optimization method and equipment applied to workshop planning production, and more particularly relates to a multi-fidelity simulation optimization method based on a DBR theory and applied to workshop planning production.
Background
The reasonable production plan can make the workshop processing more smooth and easy, and the workshop resources are fully utilized. Taking a workshop planning production problem as an example, an improper workpiece production scheme may cause machining blockage in a workshop or a large number of machines are idle, so that optimizing the workshop production plan can improve the machining efficiency. Because the production process of the production system has irreversibility, a workshop manager often judges the feasibility of the scheme by using a simulation method before executing a production decision, and even searches for a better solution of the problem by using a simulation operation method. The traditional simulation technology usually adopts a single simulation model to simulate a physical space, and when large-scale problems are faced, the simulation speed is low, so that the traditional simulation technology is difficult to be applied to the current increasingly complex processing scenes. The result obtained by the operation of the simplified simulation model is probably different from the actual production system, and has no reference significance.
However, for a system under study, multiple simulation models of different levels of fidelity may typically be built. The effect of the solution can be accurately predicted by a high-fidelity simulation model with high system restoration degree, but the simulation calculation of a complex system is high in cost and can be very time-consuming; the low fidelity simulation model has a fast solving speed, but the reliability of the simulation result is relatively low, and the low fidelity simulation model may have significant deviation and randomness, and the effect of the candidate solution cannot be accurately evaluated.
In order to guarantee the accuracy and the solving speed of the model as much as possible, the method can be considered to provide the method for realizing the rapid solving of the problem by the matching use of different fidelity models, the method is called multi-precision modeling or multi-fidelity modeling, namely on the basis of reasonable abstraction and simplification, a low-fidelity model which is easy to calculate is established to solve the whole solution space, the solutions are sampled after being subjected to sequencing and other operations, the sampled result is used as the input of a high-fidelity simulation model to be calculated and evaluated to obtain a high-quality scheme, and the method can shorten the time for finding the optimal solution by using the high-fidelity model.
At present, research and application of a multi-fidelity simulation optimization method are mostly concentrated in continuous system simulation, and related application and research in discrete system simulation are few. On the other hand, for the complex production planning problem, even if the low-fidelity simulation model is used for operation, the operation time of the whole solution space is difficult to ignore, which results in longer search time of the solution space and influences the decision efficiency.
Disclosure of Invention
Aiming at the defects or the improvement requirements of the prior art, the invention provides a multi-fidelity simulation optimization method applied to workshop planning production, and aims to optimize a workshop product production plan by using a simulation optimization method based on a multi-fidelity simulation model, so that the technical problem that the actual workshop product production plan is only made according to the experience of a manager is solved.
In order to achieve the above object, according to an aspect of the present invention, there is provided a multi-fidelity simulation optimization method applied to a plant planning production, the method specifically includes the following steps:
step 1, establishing a high-fidelity simulation model and a low-fidelity simulation model of the actual production system production problem, recording an operation result obtained by an operation scheme X of the high-fidelity simulation model as h (X), and recording an operation result obtained by the operation scheme X of the low-fidelity simulation model as l (X);
step 2, performing GA search by using a low-fidelity simulation model, thereby performing low-fidelity approximate solution space search on specific production problems to obtain a low-fidelity search solution space set;
step 3, reordering the low-fidelity search solution space set in the step 2 according to the quality of the solution;
step 4, sampling from the reordered solution space set to form an optimal sampling subset;
step 5, operating the optimal sampling subset N by using a high-fidelity modeljAnd selecting the optimal scheme to obtain the optimal production plan.
Further, the step 1 comprises the following sub-steps:
step 1.1: establishing a high-fidelity simulation model of a real scene according to production operation and process flow characteristics of an actual production system, and setting a scheduling rule and related parameters in an operation process in the high-fidelity simulation model according to actual production;
step 1.2: simplifying production operation and process flow characteristics of an actual production system, reserving modeling of a bottleneck problem in the system, replacing other non-critical processes and resources by using infinite capacity, and obtaining a low-fidelity simulation model; the scheduling rules and related parameters in the low fidelity simulation model are set to be the same as in the high fidelity simulation model.
Further, the throughput of the non-bottleneck process is replaced by the pass time in the low-fidelity simulation model.
Further, the step 2 comprises the following sub-steps:
step 2.1: setting total budget number M of low fidelity searchmaxInitializing the evolution algebra to be r as 0 according to the GA population quantity P;
step 2.2: generating a solution set of an initial population { xr1，xr2，...，xrp}，xrpRepresenting the p solution obtained by the evolution of the r generation; run and get solution set { l (x) using low fidelity simulation modelr1)，l(xr2)，…，l(xrp)}，l(xrp) Is xrpA corresponding low fidelity simulation solution;
step 2.3: setting an evolution algebra as r ═ r + 1; solution set { l (x)r1)，l(xr2)，...，l(xrp) Choose cross population 1 using elite selection rule, then set solution { l (x)r1)，l(xr2)，...，l(xrp) Sequentially disorganizing, selecting a cross population 2 by using an elite selection rule, and carrying out genetic evolution by using the cross population 1 and the cross population 2 to obtain a new scheme solution set xr＝{xr1，xr2，...，xrp}；
Step 2.4: judging whether the low fidelity search budget is used up, namely judging whether r P < MmaxIf the inequality is true, repeating the step 2.3, otherwise, turning to the step 2.5;
step (ii) of2.5: collecting all the schemes in the evolution process to form a low fidelity search scheme set x1，x2，...，xr，...，xMM is the total number of solutions, and obtains the corresponding low-fidelity search solution space set { l (x)1)，l(x2)，...，l(xr)，...，l(xM)}。
Further, the elite selection rule in step 2.3 is as follows: sequentially selecting two unselected individuals in the population, comparing the results of the two individuals, and selecting a better individual until all the individuals in the population are selected;
the genetic evolution method in step 2.3 is as follows: and (3) carrying out cross operation on the two schemes by using chromosome crossing and mutation ideas in a genetic algorithm and using a two-point crossing mode to realize global search, and carrying out mutation operation on the two schemes by using a single-point mutation mode to realize neighborhood search.
Further, step 3 comprises the following sub-steps:
step 3.1: ordinal number conversion
Searching the low fidelity for a solution space set { l (x)1)，l(x2)，...，l(xM) Sorting according to the size of the solution to form a solution space set { l (x) with the result converted from good to second ordinal numbersOT1)，l(xOT2)，...，l(xOTM)}；
Step 3.2: set of ordinal transformed solution space { l (x)OT1)，l(xOT2)，...，l(xOTM) Divide evenly into K subsets ΘjJ 1.. K, then each subset ΘjThe method includes N solutions, and M is N x K.
Further, the following optimal sampling strategy is adopted for sampling in step 4:
step 4.1: setting a high fidelity search budget NmaxNumber of initial samples N0And total incremental sample number Δ; setting an evolution algebra as r to be 0;
step 4.2: at each subset ΘjIn the random selection
Figure BDA0002433912470000046
Running the samples through a high-fidelity simulation model to obtain a running result;
step 4.3: if it is not
Figure BDA0002433912470000041
Jump to step 4.5, otherwise increase Δ budgets, according to the formula
Figure BDA0002433912470000042
j ≠ l ≠ b and
Figure BDA0002433912470000043
calculating to obtain new budget result of each subset
Figure BDA0002433912470000044
j, n, l ═ 1.·, K; the new budget amount for each subset is
Figure BDA0002433912470000045
δb，jRepresentation subset ΘbAnd ΘjMean difference between, δb，lRepresentation subset ΘbAnd ΘlMean difference between, σjDenotes ΘjStandard deviation of (a) ("SigmalDenotes ΘlStandard deviation of (a)bDenotes ΘbStandard deviation of (1), NlRepresentation assignment to ΘlThe number of high fidelity assessment budgets;
step 4.4: from the subset ΘjIn the random selection of NrjA sample is obtained; setting an evolution algebra as r ═ r + 1;
step 4.5: obtain each subset ΘjCorresponding optimal sampling subset Nj。
Further, step 5 comprises the following sub-steps:
step 5.1: n in step 4.5jCollected to form a high fidelity search best sample set
Figure BDA0002433912470000051
j 1.. K, using the high fidelity of step 1.1The model and corresponding rules and parameters are operated to obtain a high-fidelity solution space set
Figure BDA0002433912470000052
And step 5.2: the solution with the smallest run time is selected from the solution space set in step 5.1 as the final solution, i.e. the best commissioning plan.
To achieve the above object, according to another aspect of the present invention, there is provided a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method as described in any one of the preceding claims.
To achieve the above object, according to another aspect of the present invention, there is provided a multi-fidelity simulation optimization apparatus applied to plant planning production, including the computer-readable storage medium as described above and a processor for calling and processing a computer program stored in the computer-readable storage medium.
Generally speaking, compared with the prior art, the technical scheme of the invention firstly establishes a high fidelity and low fidelity simulation model according to an actual production system; and then, searching a solution space of the production planning problem by using a low-fidelity simulation model and combining the optimization iteration characteristic of the GA algorithm, and sampling by using a sampling method with ordinal number conversion and optimal sampling according to a result obtained by the operation of the low-fidelity simulation model to obtain an optimal sampling set. And finally, running the optimal sampling set through the high-fidelity simulation model, wherein the high-fidelity model has extremely high similarity with an actual production system of a workshop, and running the production scheme in the optimal sampling set by using the high-fidelity model can obtain a relatively accurate result, so that the result can be used as a basis for selecting the optimal production scheme. Based on the above concept, the present invention can achieve the following advantageous effects.
(1) The multi-fidelity simulation optimization method applied to the workshop planning production has high practicability, and on the basis of the traditional multi-fidelity simulation optimization framework, the search speed of the solution space of the low-fidelity model is accelerated by using a genetic algorithm, the efficiency of searching the solution space of the problem by using the low-fidelity simulation model is improved, the decision-making efficiency can be improved, and the decision-making in the discrete manufacturing industry is facilitated to be optimized.
(2) The invention designs the low-fidelity simulation model based on the DBR theory, and the operation result of the high-fidelity simulation model under the same problem has stronger consistency through verification, so the method is a very convenient and effective low-fidelity modeling method.
(3) The invention successfully applies the improved multi-fidelity simulation optimization method to the optimization problem of the discrete production plan, and the method framework is easy to adjust according to the actual production problem of the enterprise, and is flexible and practical.
Drawings
FIG. 1 is a flow diagram of a multi-fidelity simulation optimization method applied to plant planning commissioning;
FIGS. 2 (a) and (b) are schematic diagrams of two crossing processes of GA applied in the multi-fidelity optimization method solution space search process;
FIGS. 3 (a) and (b) are schematic diagrams of two variations of GA applied in the multi-fidelity optimization method solution space search process;
FIG. 4 is a diagram of a plant data report for a practical case of the present invention;
fig. 5 is a schematic diagram of a high fidelity simulation model of the present invention based on the actual case abstraction of fig. 4.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
FIG. 1 is a flow diagram of a multi-fidelity simulation optimization method applied to a shop plan production run, which may be used in a manufacturing environment in more detail. The plant planning commissioning problem is described as follows: the method comprises the following steps that n workpieces need to be processed in a workshop, different workpieces need to be selected for different putting time due to the limitation of a workpiece delivery period and workshop resources, the release time of each workpiece needs to be selected for planning production in the workshop, and scheduling is carried out by using a known scheduling rule to form a complete scheduling scheme. The goal of this problem is generally to minimize the total processing time. The basic concepts of high Fidelity models and low Fidelity models can be referred to Xu J, Zhang S, Huang E, et al, MO 2 TOS, Multi-Fidelity Optimization with orthogonal Transformation and Optimal Sampling [ J ]. Asia Pacific Journal of Operational Research,2016,33(3):1650017.
Preferably, the high fidelity model is modeled using the simulation modeling software SIMIO. Taking a mixed-flow machining workshop for shaft parts as an example, a simulation model data structure relation diagram of the mixed-flow machining workshop for shaft parts is established according to production conditions of a workshop and equipment, and the basic information and the association relation of each object in the workshop are reflected. Therefore, real-time adjustment can be conveniently carried out in practical application according to conditions such as workshop production data, layout change, equipment change and the like, and the simulation model can keep consistent with a workshop real scene.
For ease of understanding, the system objects and properties modeled by the simulation are described in more detail below in a practical case as shown in FIG. 4:
(1) order object
The order object is an order entity in a workshop and is processed and processed as a set of production objects, and the main attributes of the order object are as follows:
TABLE 1 order object Attribute Table
Figure BDA0002433912470000071
(2) Parts product
The part product can be integrated by order objects in a data model, but the conceptual objects and related data still exist, are used for representing the types of processing products and are associated with a process route, and the main attributes are as follows:
TABLE 2 parts product Attribute Table
Figure BDA0002433912470000072
Figure BDA0002433912470000081
(3) Process route
The process route is a processing process route table of the product, and comprises all process routes inside and outside a workshop, and relevant requirements in the process route needing to be sequenced in field scheduling are provided, and the main attributes of the process route table are as follows:
TABLE 3 Process route Attribute Table
Figure BDA0002433912470000082
(4) Production line
The workshop production mode is mixed flow production, but the actual process has many mixed line production conditions, and the product does not strictly correspond to the production line. The main attributes of the production line are:
TABLE 4 production line Attribute Table
Attribute name Data type Description of the invention
Product line name- String I.e. the name of the production line, has uniqueness
Working position apparatus String Belongs to the station apparatus of the production line
Production line equipment String Belonging to the equipment for production line management
(5) Physical distribution resource
Because other side resources in the workshop have fewer production constraints, only a logistics resource object of a station appliance is considered in the data model. The station utensil carries out the transport means and the container that material was transported in the workshop, and its main attribute has:
TABLE 5 Logistics resource Attribute Table
Figure BDA0002433912470000083
Figure BDA0002433912470000091
(6) Processing equipment
The processing equipment object is a processing equipment entity outside the workshop in the workshop, the entity comprises actual objects or concept objects identified by processes such as processing equipment, detection equipment, a detection area, a bench work area, other outsource workshops and the like, and the entity has the following main attributes:
TABLE 6 processing Equipment Attribute Table
Figure BDA0002433912470000092
On the basis of the simulation framework and the data relation table, a high-fidelity digital simulation model can be established for the mixed-flow machining workshop of the shaft parts.
Due to the fact that the current situation of workshop production is very complex, various information statistics are not comprehensive, and information loss is caused when workshop objects are subjected to abstract modeling, the fact that the real scene is required to be completely restored is unrealistic. Therefore, certain abstract and reasonable assumptions are needed to be made in the mixed-flow processing workshop of the shaft parts.
Therefore, the following basic assumptions are made on the mixed-flow machining workshop system simulation model of the shaft parts:
1) under the assumption that equipment and workers can work continuously during working, shutdown caused by various reasons is avoided under normal conditions; if abnormal situations occur, making other scene arrangement consideration;
2) the processing time of all products is assumed to be the same as the standard working hours set by a workshop, and the influence of other factors such as the skill level of workers, the equipment condition and the like on the working hours is not normally considered; special case arrangement consideration is made when necessary;
3) assuming that the defective product rate is zero, a series of influences caused by the defective products are not considered temporarily; in particular cases, other scene arrangement considerations are made for quality fluctuation effects;
4) the external coordination process is supposed to be timely sent to and returned from, or the service level is in a certain distribution range; the capacity and the scheduling of the outsourced units are not considered temporarily;
5) assuming that the capacity of a cache area in a workshop can be changed, the strict limitation of the actual space in the workshop is not considered for the moment; when rigid constraint of space resources is met, other scene arrangement consideration is additionally made.
And carrying out simulation modeling on the workshop based on the basic assumption and the simulation framework. The layout of the simulation model is basically arranged according to a CAD layout provided by a workshop and by combining with the actual situation on site. The logistics path in the model is planned according to the actual situation of a workshop, the workshop is provided with a unique inlet, a main channel is longitudinally formed and communicated to the precision room, a logistics channel is transversely arranged in the precision room, two logistics channels are transversely arranged outside the precision room, and two longitudinal channels are arranged at two ends of the workshop. The high-fidelity simulation model of the workshop is shown in FIG. 5, wherein black solid line segments represent the logistics path of the workshop.
Preferably, the low-fidelity simulation model is generally obtained by simplifying and abstracting the low-fidelity model. The patent provides a low-fidelity simulation model building method based on a constraint theory. Taking the mixed-flow processing workshop of the shaft parts as an example, the bottleneck equipment and the equipment which can become temporary bottleneck in the workshop are reserved by carrying out statistical analysis on the equipment load of the workshop. Other devices with sufficient capacity or idle standby state most of the time can be combined and simplified into a device group with infinite capacity, and fixed processing time is used instead. The low-fidelity simulation model is established according to the establishing method of the high-fidelity simulation model and by combining a bottleneck theory.
Aiming at the problem of workshop planning production, the multi-fidelity simulation optimization process applied to the workshop planning production mainly comprises the following implementation steps:
step 1, establishing a high fidelity model and a low fidelity model of the production problem of an actual production system
Step 1.1, establishing a high-fidelity simulation model which is almost the same as a real scene according to the production operation and process flow characteristics of an actual production system, and setting a scheduling rule and related parameters in the operation process in the model according to the actual production. And (5) recording the operation result obtained by the high-fidelity simulation model operation scheme x as h (x).
And 1.2, simplifying the production operation and the process flow characteristics of the actual production system according to the DBR theory, highlighting the modeling of the bottleneck problem in the system, and considering other non-critical processes and resources to use infinite capacity substitution. The scheduling rules and related parameters in the low fidelity simulation model are set to be the same as in the high fidelity simulation model. The running result obtained by the low-fidelity simulation model running scheme x is marked as l (x).
And 2, performing low-fidelity approximate solution space search on the specific production problems by using the low-fidelity simulation model established in the step 1.2 and combining the search process of the GA.
Step 2.1 setting Total budget number M of Low Fidelity searchmaxAnd setting the evolution algebra as r to be 0 according to parameters such as the GA population quantity P and the like.
Step 2.2 generating a solution set { x) for the initial populationr1，xr2，...，xrpRun using a low fidelity simulation model and get a solution set { l (x)r1)，l(xr2)，...，l(xrp)}。
Step 2.3 sets the evolution algebra to r ═ r + 1. Solution set { l (x)r1)，l(xr2)，...，l(xrp) Choose cross population 1 using elite selection rule, then set solution { l (x)r1)，l(xr2)，...，l(xrp) Sequentially disorganizing, selecting a cross population 2 by using an elite selection rule, using the cross population 1 and the cross population 2, and carrying out genetic evolution to obtain a new scheme solution set { x }r1，xr2，...，xrp}。
Step 2.4, judging whether the low fidelity searching budget is used up, namely judging whether rP is less than MmaxIf the inequality is true, repeat step 2.3, otherwise go to step 2.5.
Step 2.5 collect all the solutions in the evolution process to form a set of low fidelity search solutions { x }1，x2，...，xMGet the corresponding low fidelity search solution space set { l (x)1)，l(x2)，...，l(xM)}。
Step 3, the solution space set { l (x) in the step 2.5 is collected1)，l(x2)，...，l(xM) Using ordinal number conversion strategy to form ordered solution space set.
Step 3.1, ordinal number conversion: searching the low fidelity for a solution space set { l (x)1)，l(x2)，...，l(xM) Sorting according to the size of the solution to form a solution space set { l (x) with the result converted from good to second ordinal numbersOT1)，l(xOT2)，...，l(xOTM)}。
Step 3.2 set of ordinal number converted solution space { l (x)OT1)，l(xOT2)，...，l(xOTM) Divide evenly into K subsets Θj(j ═ 1.. K), then each subset ΘjWhich contains N solutions.
Step 4. the solution space set { l (x) after ordinal number conversion in step 3.1 is setOT1)，l(xOT2)，...，l(xOTM) Sampling using an optimal sampling strategy to form an optimal sampling set.
Step 4.1 setting high fidelity search budget NmaxNumber of initial samples N0And total incremental sample number Δ. And setting the evolution algebra as r to be 0.
Step 4.2 at each subset ΘjIn the random selection
Figure BDA0002433912470000121
Samples were run to obtain results
Step 4.3 if
Figure BDA0002433912470000122
Jump to step 4.5, otherwise increase a budget, according to the formula
Figure BDA0002433912470000123
And
Figure BDA0002433912470000124
calculating to obtain new budget result of each subset
Figure BDA0002433912470000125
The new budget amount for each subset is
Figure BDA0002433912470000126
Figure BDA0002433912470000127
Step 4.4 from the subset ΘjIn the random selection of NrjAnd (4) sampling. And setting the evolution algebra as r + 1.
Step 4.5 obtains each subset ΘjCorresponding optimal sampling subset Nj。
Step 5 run the optimal sample subset N using the high fidelity modeljAnd the best solution is selected.
Step 5.1N in step 4.5j(j 1.. K) are collected to form a high fidelity search optimal sample set
Figure BDA0002433912470000128
Using the high fidelity model and corresponding gauge in step 1.1Then operating with the parameters to obtain a more accurate solution space set
Figure BDA0002433912470000129
Step 5.2 selects the solution with the smallest run time from the solution space set in step 5.1 as the final solution.
Fig. 2 is a schematic diagram of a crossing process of a GA applied to a multi-fidelity optimization method solution space search process. The individuals 1, 2 shown in the figure are solutions in the plant commissioning planning problem. Each cell and the number in the cell represent a workpiece and a selected point in time of production for that workpiece. The intersection process is shown in the figure, two intersections are randomly selected, and the solution segments between the two individual intersections are intersected to form a new solution.
FIG. 3 is a schematic diagram of the variation process of GA applied in the solution space search process of the multi-fidelity optimization method. The individual 1 shown in the figure is a solution in the plant commissioning planning problem. And in the variation process, as shown in the figure, one variation point is randomly selected, and the production time point corresponding to the variation point is randomly modified into other production points to form a new solution.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims (5)

1. A multi-fidelity simulation optimization method applied to workshop planning production is characterized by specifically comprising the following steps:
step 1, establishing a high fidelity simulation model and a low fidelity simulation model of the actual production system production problem, wherein an operation result obtained by an operation scheme X of the high fidelity simulation model is recorded as h (X), and an operation result obtained by the operation scheme X of the low fidelity simulation model is recorded as l (X);
step 2, performing GA search by using a low-fidelity simulation model, thereby performing low-fidelity approximate solution space search on specific production problems to obtain a low-fidelity search solution space set;
step 3, reordering the low-fidelity search solution space set in the step 2 according to the quality of the solution;
step 4, sampling from the reordered solution space set to form an optimal sampling subset;
step 5, operating the optimal sampling subset N by using a high fidelity modeljSelecting the optimal scheme to obtain the optimal production plan;
the step 1 comprises the following substeps:
step 1.1: establishing a high-fidelity simulation model of a real scene according to production operation and process flow characteristics of an actual production system, and setting a scheduling rule and related parameters in an operation process in the high-fidelity simulation model according to production practice;
step 1.2: simplifying production operation and process flow characteristics of an actual production system, reserving modeling of bottleneck problems in the system, and replacing other non-critical processes and resources by using infinite capacity to obtain a low-fidelity simulation model; setting the scheduling rules and related parameters in the low-fidelity simulation model to be the same as those of the high-fidelity simulation model;
the step 2 comprises the following substeps:
step 2.1: setting total budget number M of low fidelity searchmaxInitializing the evolution algebra to be r as 0 according to the GA population quantity P;
step 2.2: generating a solution set of an initial population { xr1，xr2，...，xrp}，xrpRepresenting the p solution obtained by the evolution of the r generation; run and get solution set { l (x) using low fidelity simulation modelr1)，l(xr2)，...，l(xrp)}，l(xrp) Is xrpA corresponding low fidelity simulation solution;
step 2.3: setting an evolution algebra as r ═ r + 1; solution set { l (x)r1)，l(xr2)，...，l(xrp) Choose cross population 1 using elite selection rule, then set solution { l (x)r1)，l(xr2)，...，l(xrp) The sequence is disturbed, cross population 2 is selected by using an elite selection rule, and genetic evolution is carried out by using the cross population 1 and the cross population 2 to obtainGet a new solution set xr＝{xr1，xr2，...，xrp}；
Step 2.4: judging whether the low fidelity search budget is used up, namely judging whether r P < MmaxIf the inequality is true, repeating the step 2.3, otherwise, turning to the step 2.5;
step 2.5: collecting all the schemes in the evolution process to form a low fidelity search scheme set x1，x2，...，xr，...，xMM is the total number of solutions, and obtains the corresponding low-fidelity search solution space set { l (x)1)，l(x2)，...，l(xr)，...，l(xM)}；
The elite selection rule in step 2.3 is as follows: sequentially selecting two unselected individuals in the population, comparing the results of the two individuals, and selecting a better individual until all the individuals in the population are selected;
the genetic evolution method in step 2.3 is as follows: performing cross operation on the two schemes by using a chromosome crossing and mutation thought in a genetic algorithm and a two-point crossing mode to realize global search, and performing mutation operation on the two schemes by using a single-point mutation mode to realize neighborhood search;
step 3 comprises the following substeps:
step 3.1: ordinal number conversion
Searching the low fidelity for a solution space set { l (x)1)，l(x2)，...，l(xM) Sorting according to the size of the solution to form a solution space set { l (x) with the result converted from good to second ordinal numbersOT1)，l(xOT2)，...，l(xOTM)}；
Step 3.2: set of ordinal converted solution space { l (x)OT1)，l(xOT2)，...，l(xOTM) Divide evenly into K subsets ΘjJ 1.. K, then each subset ΘjN solutions, M ═ N × K;
in step 4, the following optimal sampling strategy is adopted for sampling:
step 4.1: setting a high fidelity search budget NmaxBeginning ofNumber of initial samples N0And total incremental sample number Δ; setting an evolution algebra as r to be 0;
step 4.2: at each subset ΘjIn the random selection
Figure FDA0003636552430000031
Obtaining a total K groups of operation results of the samples through the operation of a high-fidelity simulation model;
step 4.3: if it is not
Figure FDA0003636552430000032
Jump to step 4.5, otherwise increase a budget, according to the formula
Figure FDA0003636552430000033
And
Figure FDA0003636552430000034
calculating to obtain new budget result of each subset
Figure FDA0003636552430000035
j, b, l ═ 1.·, K; the new budget amount for each subset is
Figure FDA0003636552430000036
δb，jRepresentation subset ΘbAnd ΘjMean difference between, δb，lRepresentation subset ΘbAnd ΘlMean difference between, σjDenotes ΘjStandard deviation of (a)lDenotes ΘlStandard deviation of (a)bDenotes ΘbStandard deviation of (1), NlRepresentation assignment to ΘlThe number of high fidelity assessment budgets;
step 4.4: from the subset ΘjIn the random selection of NrjSamples of each
Figure FDA0003636552430000037
Increasing NrjA sample is obtained; is arranged intoThe generation number is r ═ r +1, and the step 4.3 is repeated;
step 4.5: obtain each subset ΘjCorresponding optimal sampling subset Nj。
2. The multi-fidelity simulation optimization method applied to the plant planning production in claim 1, wherein the step 5 comprises the following sub-steps:
step 5.1: n in step 4.5jCollected to form a high fidelity search best sample set
Figure FDA0003636552430000038
K, operating using the high-fidelity model and corresponding rules and parameters in step 1.1 to obtain a high-fidelity solution space set
Figure FDA0003636552430000039
Step 5.2: the solution with the smallest run time is selected from the solution space set in step 5.1 as the final solution, i.e. the best production plan.
3. The method of claim 1, wherein capacity utilization transit time of non-bottleneck processes is substituted in the low fidelity simulation model.
4. A computer-readable storage medium, having stored thereon a computer program which, when executed by a processor, implements the method of any one of claims 1 to 3.
5. A multi-fidelity simulation optimization apparatus applied to plant planning production, comprising the computer-readable storage medium of claim 4 and a processor for invoking and processing the computer program stored in the computer-readable storage medium.
CN202010245660.1A 2020-03-31 2020-03-31 Multi-fidelity simulation optimization method and equipment applied to workshop plan production Active CN111445079B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010245660.1A CN111445079B (en) 2020-03-31 2020-03-31 Multi-fidelity simulation optimization method and equipment applied to workshop plan production

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010245660.1A CN111445079B (en) 2020-03-31 2020-03-31 Multi-fidelity simulation optimization method and equipment applied to workshop plan production

Publications (2)

Publication Number Publication Date
CN111445079A CN111445079A (en) 2020-07-24
CN111445079B true CN111445079B (en) 2022-07-05

Family

ID=71649483

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010245660.1A Active CN111445079B (en) 2020-03-31 2020-03-31 Multi-fidelity simulation optimization method and equipment applied to workshop plan production

Country Status (1)

Country Link
CN (1) CN111445079B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112083698B (en) * 2020-09-10 2021-08-20 南昌市言诺科技有限公司 Production line real-time control method and device based on digital twin model

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007041950A (en) * 2005-08-04 2007-02-15 Sharp Corp Production simulation management apparatus
CN106250650A (en) * 2016-08-15 2016-12-21 北京理工大学 The resource allocation and optimization method of model in high flux emulation
CN107767022A (en) * 2017-09-12 2018-03-06 重庆邮电大学 A kind of Dynamic Job-shop Scheduling rule intelligent selecting method of creation data driving
CN107767068A (en) * 2017-11-02 2018-03-06 山东中烟工业有限责任公司 A kind of lean production high-level plan product plan and method

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102007485B (en) * 2008-04-18 2014-06-25 埃克森美孚上游研究公司 Markov decision process-based decision support tool for reservoir development planning
US10896270B2 (en) * 2015-07-08 2021-01-19 The Boeing Company Method for solving multi-fidelity optimization problems
TWI633504B (en) * 2017-11-16 2018-08-21 財團法人工業技術研究院 Tree search-based scheduling method and an apparatus using the same

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007041950A (en) * 2005-08-04 2007-02-15 Sharp Corp Production simulation management apparatus
CN106250650A (en) * 2016-08-15 2016-12-21 北京理工大学 The resource allocation and optimization method of model in high flux emulation
CN107767022A (en) * 2017-09-12 2018-03-06 重庆邮电大学 A kind of Dynamic Job-shop Scheduling rule intelligent selecting method of creation data driving
CN107767068A (en) * 2017-11-02 2018-03-06 山东中烟工业有限责任公司 A kind of lean production high-level plan product plan and method

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
基于SIMIO的装配车间生产线配送物流仿真分析;岳磊等;《机械制造》;20130220;全文 *
基于交互仿真及神经网络的生产单元换线决策专家系统研究;陈进;《中国优秀博硕士学位论文全文数据库(博士) 工程科技Ⅱ辑》;20130215;全文 *
基于遗传算法的MES热轧型钢生产计划调度系统的研究;花季伟等;《计算机应用与软件》;20131015(第10期);全文 *

Also Published As

Publication number Publication date
CN111445079A (en) 2020-07-24

Similar Documents

Publication Publication Date Title
Zhou et al. A multi-objective differential evolution algorithm for parallel batch processing machine scheduling considering electricity consumption cost
CN109946965B (en) Discrete manufacturing workshop scheduling method based on improved multi-target Jaya algorithm
CN107169628B (en) Power distribution network reliability assessment method based on big data mutual information attribute reduction
CN112561194B (en) Integrated scheduling method and system for production and logistics of mixed flow shop
CN115577849A (en) Workshop operation unit multi-target layout optimization method and electronic equipment
CN111445079B (en) Multi-fidelity simulation optimization method and equipment applied to workshop plan production
JP4135311B2 (en) Production planning method and system
Cherri et al. A stochastic programming approach to the cutting stock problem with usable leftovers
CN116957177A (en) Flexible workshop production line planning method, system, equipment and medium
Bernedixen et al. Simulation-based multi-objective bottleneck improvement: towards an automated toolset for industry
Liu et al. Integrated optimization of mixed-model assembly line balancing and buffer allocation based on operation time complexity
Ding et al. A simulation-based optimization method for production-distribution network design
CN110705844A (en) Robust optimization method of job shop scheduling scheme based on non-forced idle time
Gao et al. Flow shop scheduling with variable processing times based on differential shuffled frog leaping algorithm
Piplani et al. Simplification strategies for simulation models of semiconductor facilities
CN114879618A (en) Knowledge-driven workshop intelligent scheduling method
CN110298538B (en) Intelligent workshop combined scheduling control method
CN116468137A (en) Distributed process planning and workshop scheduling integrated optimization method
Zhong et al. Sequencing problem for a hull mixed-model assembly line considering manufacturing complexity
TWI653587B (en) Dispatching method and system based on multiple levels of steady state production rate in working benches
CN110287272A (en) A kind of configurable real-time feature extraction method, apparatus and system
CN110348623A (en) Complex Product Development time prediction and optimization method based on Design Structure Model
CN112734286B (en) Workshop scheduling method based on multi-strategy deep reinforcement learning
Wang et al. Multi-order scheduling optimisation considering product operation and worker allocation in divisional seru
CN107368917A (en) A kind of Power Material inventory optimization system and method based on KNN algorithms

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant